COURT OF APPEAL FOR ONTARIO

CITATION: Barclays Bank PLC v. Devonshire Trust, 2013 ONCA 494

DATE: 20130726

DOCKET: C54400

Goudge, Sharpe and Simmons JJ.A.

BETWEEN

Barclays Bank PLC

Plaintiff (Appellant)

and

Metcalfe & Mansfield Alternative Investments VII Corp., in its capacity as Trustee of Devonshire Trust, The Bank of New York, as Custodian and CIBC Mellon Trust Company, in its capacity as Indenture Trustee

Defendants (Respondent)

Peter Howard, Eliot Kolers and James Wilson, for the appellant

J. Thomas Curry, Monique Jilesen, Brendan Gray and Brian Kolenda, for the respondent

Heard: March 4-7, 2013

On appeal from the judgment of Justice Frank J. C. Newbould of the Superior Court of Justice dated September 7, 2011, with reasons reported at 2011 ONSC 5008, 93 B.L.R. (4th) 205.

By the Court:

I. INTRODUCTION

[1] This case arises from a complex 2006 transaction involving asset-backed commercial paper (“ABCP”) based on two credit default swaps (“CDSs”). The financial arrangements between the parties became economically untenable by 2007. Protracted efforts to salvage the situation ultimately failed. At issue is which party was entitled to terminate the arrangement and on what terms.

II. FACTS

A) The Third-Party ABCP Market in Canada

[2] The parties in this case are in the business of assessing risk in the financial market at a very high level, arranging transactions to allocate that risk, and profiting from arbitrage opportunities between the anticipated and actual risk.

[3] To facilitate our analysis of the issues, we begin by explaining in simple terms the design of the type of transaction that gives rise to this appeal.

[4] ABCP is a form of secured note. It is generally short-term commercial paper, meaning that it is debt that reaches maturity in less than one year, typically 30 to 90 days.

[5] The ABCP at issue in this case was issued by the respondent Metcalfe & Mansfield Alternative Investments VII Corp., in its capacity as Trustee of Devonshire Trust (“Metcalfe”). Devonshire Trust (“Devonshire”) is a special purpose trust, established to acquire income-producing assets funded through the issuance of ABCP. Devonshire’s role in this transaction is typically referred to as that of a “conduit”.

[6] The conduit sells the ABCP to investors and pays the ABCP investors, or noteholders, interest on the notes at a spread over the Canadian Dealer Offered Rate.

[7] The conduit typically acquires assets from an asset provider. Here the asset provider was the appellant Barclays Bank PLC (“Barclays”), a global investment bank, headquartered in London, England. The transaction at issue was undertaken by a Barclays deal team in New York City but involved senior management and traders in other Barclays’ offices worldwide. Barclays was active in the ABCP market internationally, but this transaction was its only Canadian ACBP transaction.

[8] Quanto Financial Corporation (“Quanto”) was the financial services agent, or sponsor, for Devonshire. In the Canadian industry, Quanto falls into the category of what is known as a “third-party”, a term used to describe sponsors other than the major Canadian banks.

[9] The conduit earns a return from the income produced by the assets and uses the income earned to pay interest on the ABCP. The conduit profits from the spread between the return it earns on the underlying asset and the cost of the interest it must pay to the ABCP investors.

[10] Traditionally, relatively tangible assets were used as security to support the ABCP, including receivables such as mortgages, loans, leases, and credit card debts. The ACBP is said to be “asset backed” because the conduit’s obligation to repay the purchaser of the ABCP is supported by the collateral of the conduit’s assets.

[11] In more esoteric arrangements, such as the transaction involved in this case, the assets involved are called “structured financial assets”, including CDSs. A CDS is a form of credit protection contract between two parties whereby one party buys from the other protection against the risk of loss in an investment such as a portfolio of corporate bonds.

[12] In this case, the CDSs were “synthetic”. Barclays, the buyer of protection, did not own the bonds for which it had purchased protection. The CDSs were a sophisticated form of derivative contract based on allocation of risk in two reference portfolios comprised of two lists of corporate bonds.

[13] The reference portfolio in a CDS is valued each day. The asset provider pays a premium to the conduit on an ongoing basis and, in exchange, the conduit agrees to pay the asset provider a certain amount if the credit losses in the reference portfolio reach defined points. The conduit is required to post collateral as security against this eventuality. However, when the transaction is leveraged, the amount of credit protection sold by the conduit is greater than the amount of collateral pledged by the conduit, the credit protection seller. If the credit losses reach certain agreed points, the conduit may be required to post additional collateral.

[14] When a CDS is used as the underlying asset to secure ABCP, the ABCP will generally mature many times within the life of the CDS. To sustain the structure of the transaction, either the holders of the ABCP have to reinvest – “roll” – their notes many times, or new investors must be found to fund the conduit’s payment obligations on the ABCP as it matures. This arrangement works as long as the notes keep rolling, but without that liquidity, the pyramid collapses.

[15] To alleviate the risk flowing from the timing mismatch between the conduit’s obligation to repay its ABCP investors and the cash flow from the longer-term assets securing the notes, conduits typically purchase liquidity protection, similar to a form of insurance contract. Liquidity protection gives the conduit access to funds required to repay the ABCP on maturity if the ABCP holders do not roll their notes and repayment cannot be funded by selling new ABCP. Asset providers are sometimes, but not always, the parties who provide liquidity protection to the conduits.

B) The Barclays-Devonshire ABCP – CDS transaction

[16] In the transaction at issue on this appeal, Barclays was Devonshire’s exclusive asset provider. Barclays was the credit protection buyer and Devonshire was the credit protection seller with respect to the two CDSs acquired by Devonshire from Barclays. As part of this transaction, Barclays was also Devonshire’s liquidity provider.

[17] Devonshire was established as a special purpose trust to acquire and hold income-producing assets financed through the issuance of ABCP. It did not have any existence or purpose outside of the transaction at issue in this case. Devonshire’s sponsor, Quanto, was formed by former National Bank executives. Devonshire retained Quanto as its financial services agent and Metcalfe & Mansfield Capital Corporation (“MMCC”) as its administrative agent. Metcalfe is the named defendant and respondent (in its capacity as Issuer Trustee of Devonshire Trust), but for the purposes of this appeal we will refer to the defendant and respondent as Devonshire, as did the parties.

[18] The classes of notes issued by Devonshire to its ABCP investors were $209 million liquidity-backed notes (Class A), $279 million extendible notes (Class E), and $190 million floating rate notes (Class FRN). Claims by noteholders of all three classes of notes ranked pari passu.

[19] The Devonshire noteholders with the most notes were financial institutions. The small Devonshire noteholders included universities, a municipality and others. When the noteholders purchased the notes from Devonshire, the notes were viewed as relatively risk-free investments. In the course of the events to be described below, Barclays eventually became a noteholder of Devonshire notes as well.

[20] Originally the parties intended this to be the first of many such transactions, but given what happened in the larger market, this ended up being their only transaction.

[21] Devonshire entered into the two CDSs or “swaps” with Barclays in August 2006. The swaps were structured and sold in one transaction. The transaction was governed by several agreements, including a 1992 ISDA Master Agreement (a standard-form contract now published by the International Swaps and Derivatives Association) and a number of other standard-form documents (collectively the “Agreements”). The term of the two swaps was ten years. The ISDA Master Agreement and the other standard-form documents were modified by the parties and tailored to their needs and objectives. In particular, the terms reflected the fact that they provided protection on two customized – “bespoke” – portfolios of corporate bonds.

[22] The CDS transaction was structured to produce income to Devonshire by putting Barclays in the role of “credit protection buyer” and Devonshire as “credit protection seller”. Barclays paid Devonshire monthly premiums in exchange for Devonshire’s commitment to pay Barclays if credit losses in the reference portfolios reached certain pre-determined levels. If the credit losses occurred, the transaction would (subject to a threshold referred to as the “attachment point”, expressed as a percentage of the total size of the portfolio), require a protection payment from Devonshire to Barclays for a portion of those losses.

[23] Liquidity protection was an important factor in Devonshire’s Class A notes being rated as low risk. These short-term notes matured within 30 to 90 days and were either rolled on maturity by the noteholders or cashed in with new notes being issued by Devonshire to other investors. The “Liquidity Facility” signed by the parties required Barclays to provide liquidity to Devonshire when the Class A notes matured if a “Market Disruption Event” (“MDE”) occurred. The Class E (extendible) notes and Class FRN (floating rate) notes did not benefit from liquidity support in the event of an MDE.

[24] The precise definition of an MDE need not be decided now, but essentially it means an event that caused the market for Devonshire’s ABCP to freeze. As we explain below, as a result of an order bifurcating the trial, the trial judge was bound to assume that an MDE occurred in August 2007, triggering Barclays’ obligation to provide Devonshire with liquidity protection.

[25] For its part, Devonshire agreed to pay Barclays if losses exceeded the attachment point, being 15 per cent and 16 per cent of the notional amount of the respective reference portfolios. The risk that losses would exceed the attachment point and require a protection payment by Devonshire was viewed as remote in 2006 because of the high attachment points and the perceived quality of the assets underlying the reference portfolio. The credit protection afforded to Barclays by Devonshire was in respect of the “super senior tranche” of the portfolio, in other words highly-rated debt. If the losses on the CDSs exceeded attachment points, Devonshire also became responsible for losses on an agreed notional portfolio of asset-backed securities. This was an additional form of credit protection that was never engaged.

[26] The swaps in this transaction, called “leveraged super senior credit default swaps”, were highly leveraged. The combined effect of highly-rated debt and highly-leveraged protection meant that the risk assumed by Devonshire under the CDSs was very low but that the extent of Devonshire’s liability was very high if the risk materialized.

[27] Devonshire was required to pre-pay Barclays $600 million as collateral to secure its obligation when entering into the swaps but, because of the leveraged nature of the transaction, Devonshire was exposed to potential credit default losses that were ten times higher – $6 billion. To fund its initial $600 million collateral payment, Devonshire raised money by issuing and selling ABCP.

[28] The terms of the transaction required Barclays to post $600 million with the Custodian (Bank of New York) as collateral in favour of Devonshire to secure Barclays’ obligation to return the collateral owed to Devonshire at the end of the term or upon the termination of the transaction. In other words, it was contemplated that at the end of the term of the transaction, Barclays would repay the $600 million CDS collateral to Devonshire. Devonshire would use that money to repay the holders of the outstanding ABCP notes. Barclays would be entitled to the return of the $600 million collateral held by the Custodian.

[29] Barclays was entitled under the Agreements to call for more collateral under certain specified conditions of increased risk of default in the reference portfolios because the transaction was so highly leveraged.

[30] In a worst case scenario, if the risk of default increased and the CDS market became significantly unfavourable to Devonshire, it could decline to post additional collateral and terminate the transaction. This “stop-loss” option built into the terms of the transaction allowed Devonshire, in such circumstances, to preserve a substantial proportion of its assets for the benefit of its noteholders.

C) “Events of Default” and “Early Termination”

[31] The Agreements specify various events that constitute “Events of Default” that entitle one party or the other to terminate the Agreements prior to the end of the ten-year period. The Agreements provide for the consequences of default and “Early Termination” in various scenarios.

[32] Section 6(a) of the ISDA Master Agreement provides that if an “Event of Default” occurs with respect to one party, the other party is entitled to designate an “Early Termination Date”:

6. Early Termination

(a) Right to Terminate Following Event of Default. If at any time an Event of Default with respect to a party (the “Defaulting Party”) has occurred and is then continuing, the other party (the “Non-defaulting Party”) may, by not more than 20 days notice to the Defaulting Party specifying the relevant Event of Default, designate a day not earlier than the day such notice is effective as an Early Termination Date in respect of all outstanding Transactions.

[33] Devonshire alleges that Barclays breached s. 5(a)(i) of the ISDA Master Agreement by failing to provide liquidity when the market became illiquid, as will be explained below. Section 5(a)(i) provides that a failure to make a required payment is an Event of Default if not “cured” within three days:

5. Events of Default and Termination Events

(a) Events of Default. The occurrence at any time with respect to a party … of any of the following events constitutes an event of default (an “Event of Default”) with respect to such party:─

(i) Failure to Pay or Deliver. Failure by the party to make, when due, any payment under this Agreement or delivery under Section 2(a)(i) or 2(e) required to be made by it if such failure is not remedied on or before the third Local Business Day after notice of such failure is given to the party[.]

[34] In turn, Barclays denies that it was under an obligation to pay the liquidity payments. However, as a result of a bifurcation order discussed below, it must be assumed at this point that Barclays did breach that obligation in August 2007.

[35] For its part, Barclays alleges that Devonshire breached the Agreements by being insolvent, contrary to s. 5(a)(vii)(2) of the ISDA Master Agreement, which provides:

5. Events of Default and Termination Events

(a) Events of Default. The occurrence at any time with respect to a party … of any of the following events constitutes an event of default (an “Event of Default”) with respect to such party:─

(vii) Bankruptcy. The party ... (2) becomes insolvent or is unable to pay its debts or fails or admits in writing its inability generally to pay its debts as they become due[.]

[36] In January 2009, Devonshire notes were in default, although because of a temporary standstill resolution not to take any steps to realize on the notes to facilitate on-going restructuring, the noteholders had not demanded payment. The trial judge, however, found that Devonshire was insolvent in January 2009 and that finding is not challenged on this appeal.

D) Settlement Amount

[37] The Agreements specify a “Settlement Amount” to be determined by the “Non-defaulting Party” upon Early Termination. Depending upon the circumstances, the Settlement Amount can be positive or negative and thus the “Defaulting Party” could be entitled to payment in some situations. Moreover, the Agreements provide that if Barclays is in default, any amount payable to Barclays as a Defaulting Party is subordinated to the claims of the noteholders.

E) The Turmoil in the ABCP Market Beginning in August 2007

[38] The Barclays-Devonshire transaction was caught up in wider problems with the ABCP market in 2007. As it became apparent that many bonds were over-rated, particularly in the United States, and that many ABCP transactions like this one were under-collateralized, investors became unwilling to roll or to purchase ABCP notes. As a result, the conduits did not have the money to continue paying its ABCP noteholders as the short-term notes came due. The timing mismatch between the maturity of the notes and the longer term of the underlying CDSs used to securitize the notes became a fatal problem.

[39] On August 13, 2007, the third-party (i.e. non-Chartered bank-sponsored) ABCP market froze in Canada.

[40] Because of the uncertainty in the marketplace and the lack of liquidity, the likelihood of collateral calls being made on the conduits by the asset providers to provide more collateral increased. Because noteholders were not rolling their notes, liquidity calls were being made by conduits for cash to pay out the noteholders.

F) Devonshire’s Market Disruption Notices and Default Notice

[41] Devonshire sent market disruption notices to Barclays on August 13, 14 and 15, 2007, requesting payments from Barclays under the Liquidity Facility. Devonshire’s position was that given the illiquidity in the ABCP market, an MDE as contemplated in the Liquidity Facility had occurred.

[42] Barclays took the position that an MDE had not occurred in the third-party ABCP market and refused to provide any liquidity payments to Devonshire. On August 14, 2007, Devonshire delivered a default notice to Barclays. There was no dispute, and the trial judge found, that the effect of Devonshire’s default notice under the Liquidity Facility was to give Barclays three days to cure the default.

G) The Montreal Accord

[43] A meeting of the major players in the third-party ABCP market was held in Montreal on August 15 and into the early hours of August 16, 2007. It was organized in large part by the Caisse de dépôt et placement du Québec (the “Caisse”), a very large investor in ABCP, and by National Bank, a large dealer of ABCP. It was attended by ABCP noteholders, dealers, and asset and liquidity providers. Barclays attended the meeting. The conduits were not represented.

[44] The goal of this meeting was to get the asset providers to agree on a moratorium against any collateral calls being made for more security, and to have the conduits agree to a moratorium from making liquidity calls for funds to pay noteholders who were not rolling their notes. In the words of trial judge, the purpose of the meeting was to prevent a “blow-up of the market and to have everyone put their weapons down and take a pause”: at para. 26.

[45] The “Montreal Accord” was reached on August 16, 2007, before the opening of the markets. It contained an interim agreement (the “Standstill Agreement”) precluding calls by the conduits for liquidity payments and calls by the asset providers for collateral to be posted by the conduits (the “Standstill”) for an initial period of 60 days (the “Standstill Period”). The Montreal Accord also contained a proposal with a framework of principles to be used in restructuring each of the conduits. It was later extended to March 14, 2008.

[46] Barclays, as asset and liquidity provider to Devonshire and no other conduit, was a signatory to the Montreal Accord. Other major noteholders of Devonshire who signed the Montreal Accord were the Caisse, National Bank and Desjardins Group (“Desjardins”). Not all of the 22 conduits in the third-party ABCP market were initial signatories. However, on October 15, 2007 Devonshire and all other affected conduits signed the Accord as well.

[47] The Montreal Accord contained an explicit reference to good faith as the parties undertook to “work together in good faith with the other participants in the discussions to bring about the timely implementation of these arrangements”.

[48] Following the Montreal Accord, the “Pan-Canadian Third Party Asset-Backed Commercial Paper Investors Committee” was formed by investors of ABCP notes to negotiate for investors in the restructuring of the ABCP market (the “Investors Committee”). Purdy Crawford, Q.C. was appointed its chairman. The Investors Committee and its advisors led the negotiations on behalf of the conduits, including Devonshire.

[49] A “Framework Agreement” was made on December 23, 2007, covering 20 of the trusts, but not Devonshire. This was an agreement in principle as to how those conduits were to be restructured and it eventually led to a restructuring under the Companies’ Creditors Arrangement Act, R.S.C. 1985, c. C-36 (the “CCAA”). Barclays was not prepared to make the concessions required by the Framework Agreement and refused to sign it.

[50] The signatories to the Framework Agreement ultimately came to a negotiated resolution. A CCAA filing took place in March 2008 covering the restructuring of the 20 conduits that were parties to the Framework Agreement. The CCAA plan was later approved by Campbell J. and then by this court in August 2008: ATB Financial v. Metcalfe & Mansfield Alternative Investments II Corp., 2008 ONCA 587, 92 O.R. (3d) 513. After that, because of dramatic market changes that took place in the fall of 2008 (such as the collapse of the investment bank Lehman Brothers), the plan was twice renegotiated in December 2008 at the insistence of the Investors Committee. This larger restructuring closed in January 2009.

H) Devonshire’s Suspension Notice and its Extension by Barclays

[51] On August 16, 2007, the same day the Montreal Accord was reached, Devonshire delivered a suspension notice to Barclays (the “Suspension Notice”), in light of the Accord and the anticipated negotiations to restructure Devonshire. Devonshire thereby suspended its default notice (without prejudice) and agreed not to take any further steps to enforce its rights under that notice until the end of the Standstill Period. There is no dispute that at this point, one of Barclays’ three days in which to make timely liquidity payments had passed. Barclays wanted Devonshire to simply rescind the default notice, but Devonshire refused unless further assurances were provided, which Barclays did not provide.

[52] The attempts to restructure Devonshire were carried on outside the provisions of the Montreal Accord following Barclays’ withdrawal in December 2007 from the negotiations leading to the Framework Agreement. Most of the discussions to restructure Devonshire were held without the direct participation of Devonshire. Barclays negotiated directly with the major noteholders of Devonshire, in particular the Caisse.

[53] To facilitate these negotiations, the Standstill Period was extended by agreement between Barclays and Devonshire for fixed periods of time until February 22, 2008, then daily until March 14, 2008, then for a fixed period until April 16, 2008, and thereafter on a daily basis until January 12, 2009.

[54] The daily extensions of the Standstill Period were made by identical e-mail messages from Barclays to a principal of Quanto, Devonshire’s sponsor, stating:

There are still a number of issues being worked out regarding the proposed restructuring of Devonshire Trust. Accordingly, for the sake of good order we are confirming that, as between Barclays and Devonshire, the Montreal Accord standstills and the related suspension of default notices have and will continue through [the next business day] to allow for these negotiations to continue. If anyone takes a different position, please let us know ASAP.

I) Termination of the Transaction: January 13, 2009

[55] The events of January 2009 lie at the heart of this appeal. We review those events in detail in the Analysis portion of these reasons. Essentially, what occurred was as follows.

[56] The negotiations to restructure Devonshire collapsed in January 2009 when Barclays delivered an ultimatum to the Caisse insisting that the Caisse accept a term sheet for Devonshire’s proposed restructuring that Barclays had made in April 2008 and that the Caisse had not accepted. Barclays combined the ultimatum with an intricate series of steps that it claims cured its default as liquidity provider, entitling it to terminate the swaps on the ground of Devonshire’s insolvency and to claim $1.2 billion as a termination payment under the Agreements.

[57] On Friday, January 9, 2009, Barclays sent the usual e-mail to Devonshire to ask for an extension to the close of business on January 12, 2009 without disclosing to Devonshire the ultimatum it had delivered to the Caisse. On January 13, Barclays wired funds to Devonshire’s bank, in the amount of the requested liquidity payments and interest. Within minutes of the transfer and two hours before the funds were actually credited to Devonshire’s account, Barclays delivered a Notice of Early Termination, citing s. 5(a)(vii) of the ISDA Master Agreement, which provides that bankruptcy of a party is an Event of Default. No particulars were provided in the Notice. After making the transfer, Barclays asserted a security interest over the funds and any other funds held by the bank on Devonshire’s account. Later that same day, Devonshire delivered its own Notice of Early Termination.

[58] Barclays asserts that its January 13, 2009 payment cured any default under the Liquidity Facility and put it in a position to terminate the Agreements on the ground that Devonshire was insolvent. Devonshire asserts that the payment did not have that effect and that as Barclays was in default, Devonshire’s Notice of Early Termination is valid.

J) This Action

[59] Barclays commenced this action the day it purported to cure its default, seeking, inter alia, declarations that it had properly designated January 13, 2009 as an Early Termination Date and was entitled to the return of the $600 million collateral held by the Custodian. It also asserted a security interest over Devonshire’s assets to secure the Settlement Amount.

[60] Devonshire counterclaimed for, inter alia, declarations that an Event of Default had occurred with respect to Barclays and that Devonshire’s designation of Early Termination was valid. It also sought judgment in the amount of approximately $725 million and a declaration that Barclays’ claim for the Settlement Amount was subordinated to the repayment of the noteholders.

K) The Bifurcation Order

[61] Prior to the trial, a bifurcation order was made by C. L. Campbell J. on consent, in which it was agreed that a number of issues would be bifurcated and deemed to be decided for the purposes of the first trial. The bifurcated issues were: (i) whether there was a MDE in August 2007; (ii) whether Devonshire’s market disruption notices and notice of default were valid; (iii) whether Barclays was in default under the notices sent by Devonshire up to August 16, 2007; and (iv) whether Devonshire was precluded from asserting the occurrence of an MDE. It was agreed that for the purposes of the first trial, now before us on appeal, that these issues would be determined in Devonshire’s favour without prejudice to Barclays’ position that its payment on January 13, 2009 cured any default.

[62] Phase one of the trial was expedited.

III. THE TRIAL JUDGMENT

[63] The crucial issues at the 51-day trial were focussed on the events that took place in January 2009. The record was extensive and included the testimony of a number of fact witnesses and several experts, and thousands of pages of documents. The trial judge had the benefit of lengthy oral and written submissions.

[64] He provided detailed reasons, making findings of fact largely adverse to Barclays. He also found that the evidence of the Barclays fact witnesses, while helpful in some respects, was self-serving, unreliable and incredible in several other respects. He did not have the same reservations with respect to the Devonshire witnesses and generally preferred their evidence.

A) The Trial Judge’s Findings: Summary

[65] To introduce the issues that arise on this appeal and to provide a road map for the analysis that follows, we provide at this point a brief summary of the trial judge’s principal findings and conclusions. We will provide a more detailed review of the evidence and the trial judge’s findings on an issue-by-issue basis in the analysis portion of these reasons.

B) Misrepresentation

[66] The trial judge found that Barclays had misrepresented to Devonshire the state of its negotiations with the noteholders when it requested and obtained extensions of the Standstill on January 8 and 9, 2009. The trial judge found that Devonshire had agreed to those extensions in reliance on Barclays’ representation that the negotiations were proceeding towards a possible resolution. That representation was untrue. Barclays had decided to terminate the swaps and it knew that its ultimatum to the Caisse would be refused. He held that the misrepresentation was not only negligent, but fraudulent.

[67] The trial judge found that as the extensions of the Standstill had been agreed to on the basis of misrepresentation, the extensions should be set aside. That meant that the two days that remained when the Standstill Period began for Barclays to make timely liquidity payments had expired by January 13, 2009. Barclays was therefore in default on that date. However, the trial judge also found that while the three-day cure period had expired, Barclays was still entitled to cure its default until such time as Devonshire terminated the swaps.

C) Timing of Barclays’ Cure Payment

[68] On January 13, 2009, Barclays made a payment which it argued cured its default under the Liquidity Facility, thereby entitling it to serve its Notice of Early Termination that day. The trial judge dismissed this argument because he found that the payment was not effected by the time Barclays delivered its Notice of Early Termination. Barclays therefore had failed to make timely payment when it delivered that notice and as a result did not have the right to terminate the ISDA Master Agreement.

D) Breach of Duty of Good Faith

[69] Barclays purported to cure its default by making the liquidity payments on the morning of January 13, 2009. The trial judge found that in doing so Barclays breached its duty of good faith which arose from the wording and operation of the Suspension Notice and the agreements to extend the Standstill. The breach of good faith was the design and execution of Barclays’ strategy to terminate the swaps, including its ultimatum to the Caisse. The payment Barclays made was not a genuine effort to cure its liquidity default but a thinly-veiled move to terminate the swaps on terms favourable to Barclays. Accordingly, because it acted in bad faith, Barclays could not rely on the payment to cure its default. The trial judge found that this also meant Barclays did not have the right to terminate the ISDA Master Agreement.

E) Benefit from Own Wrong

[70] Devonshire also argued that Barclays’ failure to make the liquidity payments was the cause of its insolvency and, therefore, that Barclays should be prevented by its own wrongdoing from relying on Devonshire’s insolvency as an Event of Default and a ground for termination. The trial judge found that while a term should be implied to prevent Barclays from relying on the insolvency of Devonshire, this could assist Devonshire only with respect to the Class A notes but not the Class E and Class FRN notes, as the liquidity obligation did not extend to them. Moreover, the trial judge concluded that it was Devonshire’s decision not to pay the interest on the Class E and Class FRN notes. Thus, with respect to the latter two classes, the trial judge found that Barclays was not prevented from relying on Devonshire’s insolvency on the basis of the principle that a party cannot benefit from its own wrongdoing.

F) Waiver and Election

[71] The trial judge rejected Devonshire’s argument that by insisting that it had no obligation to make the liquidity payments demanded of it by Devonshire on August 13, 14 and 15, 2007, Barclays had waived its right to later cure its default.

[72] However, the trial judge found that in continuing to make protection payments to Devonshire during the Standstill Period when it alleged Devonshire was insolvent, Barclays had elected to affirm the contract and lost the right to rely on Devonshire’s insolvency as a ground of default. The trial judge concluded that this was another reason why Barclays did not have the right to deliver its Notice of Early Termination.

G) Repudiation

[73] Before dealing with repudiation, the trial judge briefly summarized his conclusions concerning why Barclays was not entitled to rely on its Notice of Early Termination: (i) it had elected not to rely on Devonshire’s insolvency; (ii) it failed to make a timely payment of its liquidity obligation before delivering its Notice of Early Termination; and (iii) it could not rely on its conditional payment of its liquidity obligation because it had breached its good faith obligations.

[74] Following this summary, the trial judge went on to find that Barclays’ delivery of its Notice of Early Termination together with related steps constituted a repudiation of the parties’ Agreements. He stated that a party’s intention not to be bound by a contract “may be evinced by a refusal to perform, even though the party refusing mistakenly thinks he is exercising a contractual right”: at para. 318.

[75] The trial judge found that because Barclays had elected to waive Devonshire’s insolvency, Devonshire must be treated as a Non-defaulting Party and entitled to deliver its Notice of Early Termination on January 13, 2009. Devonshire was entitled to terminate the CDSs on the ground that Barclays was in default under the Liquidity Facility. The trial judge went on to find, however, that if Devonshire were not able to deliver its Notice of Early Termination as a Non-defaulting Party, by doing so, it accepted Barclays’ repudiation of the contract and thus brought the contract to an end.

H) Devonshire’s Notice of Early Termination

[76] The trial judge found that Devonshire’s Notice of Early Termination delivered on the afternoon of January 13, 2009 was valid and effective.

[77] Although the trial judge concluded that Devonshire was insolvent as of January 13, 2009 and that, under s. 6(a) of the ISDA Master Agreement, only a Non-defaulting Party could deliver a valid Notice of Early Termination, he found that Barclays could not rely on Devonshire’s insolvency as a bar to Devonshire’s right to deliver a Notice of Early Termination. This was because he concluded that Barclays had elected to waive reliance on Devonshire’s insolvency.

I) Settlement Amount

[78] The trial judge found that Devonshire was entitled to the return of its $600 million in collateral, minus the liquidity call payment and interest it had received, and other “Unpaid Amounts”, as defined in the Agreements, for a total of $532.7 million.

[79] The trial judge went on to consider what Barclays’ “Loss” as defined in the Agreements would have been if it had been the Non-defaulting Party and Devonshire was the Defaulting Party. The parties asked the trial judge to include this finding in his judgment, presumably because it has significance for the second part of the bifurcated trial, depending on the outcome of this appeal.

[80] The trial judge rejected the Barclays’ expert’s valuation of its Loss at $1.2 billion. He found that the model used by the expert, while used on a day-to-day basis to estimate the value of the swaps, was not the appropriate model in this situation and that it improperly excluded certain key features of the Agreements. He preferred the model used by the Devonshire expert, a cash flow model, which he concluded resulted in a valuation of the Loss at a mere $12,000, reduced to $0 because of mitigation. In reaching this conclusion, he disagreed with the Devonshire expert that Barclays should receive a risk premium of $264 million, the difference between the real-world estimate of loss and the market implied estimate of loss.

J) Limited-recourse, Priorities and Subordination

[81] Finally, the trial judge found that the issue of which assets Barclays had recourse to (other than the collateral) was moot, and that Barclays’ interests were subordinated to those of the other noteholders in light of the Intercreditor Agreement between Barclays, Devonshire, and CIBC Mellon Trust Company.

IV. ISSUES

[82] There are four broad questions raised by this appeal:

i. Was the trial judge correct in finding that Barclays’ Notice of Early Termination was invalid?

ii. Was the trial judge correct in finding that Devonshire’s Notice of Early Termination was valid?

iii. Was the trial judge correct in his determination of the Settlement Amount?

iv. Was the trial judge correct in his determination of the issues of priorities, subordination and limited-recourse?

[83] On the first question, Barclays attacks the trial judge’s findings as to:

i. Misrepresentation;

ii. Bad faith;

iii. Timing of the cure payment; and

iv. Election not to rely on Devonshire’s insolvency.

[84] By way of cross-appeal on this question, Devonshire attacks the following findings of the trial judge:

i. In delivering its Notice of Early Termination, Barclays was not barred by the principle that it should not be permitted to benefit from its own wrong.

ii. Barclays had a right to make the liquidity payments after the cure period expired, provided Devonshire had not terminated the contract.

iii. Barclays had not waived its right to cure the default by failing to make the liquidity payments demanded by Devonshire in August 2007.

[85] On the second question, Barclays attacks the trial judge’s finding that even if Devonshire was not able to deliver a Notice of Early Termination under the Agreements, by doing so it accepted Barclays’ repudiation and thereby brought the contract to an end. Barclays argues that as the Agreements spelled out the consequences of terminating the swaps, the parties contracted out of the common law of repudiation. Barclays also argues that, in any event, by serving its Notice of Early Termination and relying on the terms that spelled out the consequences of default, Devonshire affirmed the Agreements rather than accepted Barclays’ repudiation.

[86] On the third question, Barclays’ position is that the trial judge erred in rejecting Barclays’ estimate of its own Loss and in fashioning his own unprecedented valuation methodology.

[87] On the fourth question, Barclays argues that the trial judge erred in his interpretation of the priority provisions in the Agreements and contends that it has priority over the noteholders. Further, it argues that if the Agreements were terminated by an Event of Default affecting Barclays as asset provider, subordination to the noteholders does not apply to fees due to Barclays in its other capacities: liquidity provider; holder of Series A notes; or recipient of the Unpaid Amounts portion of any Early Termination payment. Barclays also submits that the trial judge erred in finding that Barclays did not have recourse to assets of Devonshire beyond the $600 million in collateral.

[88] As will be explained below, given our findings, it is not necessary to address all of these issues.

V. ANALYSIS

A) Issue 1. Is Barclays’ Notice of Early Termination valid?

[89] We begin our analysis of this issue by noting two important points that are not in dispute and that considerably narrow the scope of this appeal. First, as we have already pointed out, as a result of the bifurcation order, it is an assumed fact that (i) there was an MDE in August 2007, (ii) Devonshire’s market disruption notices and notice of default were valid, and (iii) Barclays was in default under the notices sent by Devonshire up to August 16, 2007 when the Standstill Agreement was entered into.

[90] Second, findings of fact by the trial judge attract deference on appeal and are only reviewable by this court if they reveal palpable and overriding error: Waxman v. Waxman (2004), 2004 CanLII 39040 (ON CA), 186 O.A.C. 201 (C.A.), at para. 291. This standard applies to all factual findings, whether based on credibility assessments, the weighing of competing evidence, expert evidence, or the drawing of inferences from primary facts: see Waxman, at paras. 359-60. Findings of fact grounded in credibility assessments are particularly difficult to disturb on appeal: see Housen v. Nikolaisen, 2002 SCC 33, 2 S.C.R. 235, at para. 24. The trial judge was free to accept some, none, or all of the evidence of the Barclays witnesses, and he gave adequate reasons for rejecting much of it.

[91] As will become clear, it is our view that Barclays’ attack on the trial judgment essentially relates to findings of fact that attract significant deference before this court.

(1) Misrepresentation

[92] The trial judge found that Barclays’ e-mail requests for extensions of the Standstill sent on January 8 and 9, 2009 amounted to a fraudulent misrepresentation and that Devonshire’s acceptance of the extensions should be rescinded. Barclays attacks the trial judge’s finding of misrepresentation essentially on three grounds.

[93] First, Barclays submits that the trial judge erred in his interpretation of the ultimatum delivered to the Caisse that triggered the end of negotiations to restructure Devonshire. Barclays contends that the ultimatum was a legitimate negotiating strategy and that it was not open to the trial judge to find that Barclays did anything misleading when it asked for an extension of the Standstill on a “business as usual” basis even though it had already sent the ultimatum knowing that it would be refused.

[94] Second, Barclays submits that the trial judge erred by focusing on Devonshire rather than on the Caisse. The restructuring negotiations were largely being conducted between Barclays and the Caisse with no involvement from Devonshire. The Caisse knew the state of the negotiations and was not misled by the ultimatum and, even if it had known the state of negotiations, Devonshire would have acted as directed by the Caisse.

[95] Third, Barclays argues that even if there was a misrepresentation, the trial judge erred in law and in fact in finding that Devonshire relied to its detriment in agreeing to the extension of the Standstill.

[96] For the following reasons, we reject these arguments.

(a) The trial judge’s finding of misrepresentation

[97] We begin our analysis of these issues with a review of the evidence and the finding that Barclays was guilty of fraudulent misrepresentation.

[98] The trial judge found that from early April 2008, Barclays was considering terminating the Devonshire swaps if it could not come to an agreement with the largest Devonshire noteholders.

[99] A meeting took place in New York City on April 8, 2008 between Barclays, the Caisse, National Bank and the Investors Committee to discuss the Devonshire restructuring.

[100] On April 9, 2008, Barclays sent a revised term sheet to the Caisse, National Bank, Desjardins and Citibank, the four largest Devonshire noteholders. It contained terms which were not discussed or agreed to at the April 8 meeting. Desjardins and Citibank declined to participate, and later that month, Barclays acquired their notes at a small fraction of their face value. In October 2008, Barclays and National Bank signed a framework agreement under which Barclays also acquired the Devonshire notes held by National Bank.

[101] After the agreement was made between Barclays and National Bank, the Caisse, which still held Devonshire notes with a face value of $385 million, remained the only large noteholder which had not come to an agreement. The only other noteholders with which an agreement had not been reached were the small noteholders, which together held about $75 million in notes.

[102] Barclays and the Caisse continued negotiations in the second half of 2008 as the ABCP market continued to deteriorate. Lehman Brothers and other financial institutions failed in the fall of 2008 and the world markets became more illiquid. Debt previously considered investment grade was now trading as junk.

[103] In early November 2008, the Caisse began asking for revised terms to reflect market changes, along the lines of those obtained by the Investors Committee in the CCAA restructuring. Barclays was not happy with the requested changes, but did make some concessions during the ongoing negotiations. Then, in December 2008, Barclays took the position that it had already reached an agreement with the Caisse in April 2008. The Caisse denied the existence of any agreement and demanded even more favourable terms, since the market was moving in its favour.

[104] On January 8, 2009, a Barclays executive sent an e-mail to his counterparts at the Caisse, enclosing a term sheet that reflected what Barclays claimed the parties had agreed to in April 2008. It did not reflect the concessions that Barclays had subsequently made in November, much less the additional demands the Caisse had made in December 2008. This e-mail requested that a meeting be called for the Caisse board of directors to “ratify our previous agreement on the restructuring of the Caisse’s holdings”. The e-mail concluded by stating that Barclays was looking forward to receiving a copy of the term sheet, signed back, no later than the end of the business day on January 12, 2009.

[105] A senior officer of Barclays admitted on cross-examination that the deadline Barclays gave the Caisse in this e-mail could fairly be called “an ultimatum.” Barclays knew that it was unlikely that the Caisse executives could deal with the e-mail by January 12, 2009 because they were in Toronto, working on the closing of the CCAA restructuring deal involving the other conduits. The trial judge found that the timing was not coincidental. Not only did Barclays know that the Montreal Caisse executives were preoccupied that week, but Barclays did not want to risk being exposed to the Caisse after the CCAA restructuring, when the Caisse would have incentive to walk away from the trades.

[106] A Barclays representative spoke to a Caisse representative by telephone on Friday, January 9, 2009, to say that January 12 was a real deadline. The Caisse said it would respond to Barclays’ e-mail that afternoon. However, there was no further communication between Barclays and the Caisse before Barclays sent its final daily extension of the Standstill Period with Devonshire at the end of the day on January 9.

[107] The trial judge found that Barclays never expected the Caisse to agree to the terms of the ultimatum and that Barclays’ real intention in delivering the ultimatum was to terminate the swaps.

[108] Barclays sent the last of the e-mails extending the Standstill to Quanto (effectively to Devonshire) on January 8 and 9, 2009, the second to be effective through the close of business on Monday, January 12, 2009. It repeated the same language that it had used for months, stating that “[t]here are still a number of issues being worked out regarding the proposed restructuring of Devonshire Trust” and asking for a further extension “to allow for these negotiations to continue.” The trial judge found that, as these same statements had been made on a daily basis for over eight months, they “gave the impression that it was business as usual so far as negotiations were concerned”: at para. 114. The trial judge found clearly: “It was not” (emphasis added).

[109] The trial judge found that Barclays did not believe that negotiations were continuing when it sent the e-mail requesting an extension. Rather, Barclays had already decided upon a carefully choreographed set of steps purporting to cure its default in meeting Devonshire’s call for liquidity support but in fact designed to extricate Barclays from the transaction and to maximize recovery of a Settlement Amount under the terms of the Agreements. The trial judge found that the e-mails extending the Standstill were part of that strategy.

[110] Barclays did not tell Devonshire about the ultimatum to the Caisse. The trial judge found that Barclays did not want Devonshire to know the state of its negotiations with the Caisse because Barclays wanted the strategic advantage of being able to terminate the swaps first.

[111] Senior representatives of Barclays testified at trial that the ultimatum e-mail to the Caisse was an attempt to continue negotiations that they hoped would result in some acceptable restructuring of Devonshire. The trial judge explicitly rejected that evidence and found that Barclays knew that the Caisse’s position was that there had not been an agreement in April 2008 and that the April 9 term sheet drafted by Barclays contained items not discussed in the meeting in New York City the day before and never agreed to afterward. He also found that the April 9 term sheet, sent by Barclays on January 8, 2009 as part of its ultimatum e-mail, was materially worse to the Caisse than the proposal Barclays had made on December 18, 2008, and that it would have been completely against the Caisse’s economic interests to agree to the terms Barclays proposed in the January 8 e-mail.

[112] The trial judge found that no one at Barclays expected the Caisse to agree to these terms. A Caisse witness testified that he was insulted by the proposed terms and had no intention of recommending them to his Board of Directors. The trial judge concluded that Barclays was “posturing” and “positioning itself to exit the Devonshire trades” and to deliberately end negotiations with the Caisse: at para. 120. In the words of one Barclays witness, if the Caisse did not agree to the ultimatum e-mail, Barclays intended to “blow up the box”, meaning to terminate the Devonshire swaps.

[113] The trial judge further held that the misrepresentation in the e-mails had been amplified by the failure of Barclays’ executives to respond to direct inquiries by counsel for the Indenture Trustee on December 30, 2008, as to the status of Barclays’ negotiations on the restructuring of Devonshire. This, he found, “compounded the misleading nature of the extensions e-mails”: at para. 178.

[114] While it was not strictly necessary for him do so, the trial judge went on to find that Barclays’ misrepresentations amounted to fraudulent misrepresentations. He found that the members of the Barclays team responsible for the e-mails knew full well that statements in the extension e-mails were false.

(b) Did the trial judge err in his interpretation of the ultimatum delivered to the Caisse?

[115] We see no basis for holding that the trial judge erred in his interpretation of the ultimatum. No doubt, there are circumstances in which the delivery of an ultimatum can be a legitimate strategy aimed at moving the negotiations forward. But whether this ultimatum was consistent with an intention to continue negotiations or rather a ploy to bring negotiations to an end was a matter of fact, not law, and the assessment of Barclays’ knowledge and intentions in delivering the ultimatum was very much a factual issue for the trial judge to decide.

[116] The trial judge supported his findings regarding the ultimatum with a thorough review of the evolution of these negotiations. In our view, that evidence, the essentials of which we have just reviewed, is clearly capable of supporting the trial judge’s finding that Barclays asked Devonshire to extend the Standstill Period to facilitate further negotiations even though it had already made the decision to terminate those negotiations by delivering an ultimatum to the Caisse that Barclays knew would be refused. The misrepresentation so found was, in the trial judge’s view, part and parcel of Barclays’ carefully conceived strategy to terminate the swaps to its own advantage before Devonshire was able to do so.

[117] As the trial judge put it, “[s]ometimes actions speak louder than words”: at para. 150. Barclays took several steps that were simply inconsistent with the prospect or expectation of further negotiations to restructure Devonshire. Those steps reveal that the ultimatum was part of a scheme designed by Barclays to “blow up” the Devonshire swaps. On January 8, 2009, Barclays began to gather relevant information on profit and loss that it would need once the trade was terminated. On January 9, Barclays arranged for the liquidity payments to be made. On January 8 and 10, it made trades, in the words of a Barclays trader “in anticipation of [the] Devonshire unwind” and to deal with the “potential Devonshire impact next week”. Barclays also contacted government authorities and initiated a planned public relations response to protect its franchise from the fallout of terminating the Devonshire transaction.

[118] We conclude that there is no merit to Barclays’ submission that the trial judge erred in his interpretation of the ultimatum delivered to the Caisse that triggered the end of negotiations to restructure Devonshire. It was open to him on the evidence to reject Barclays’ contention that the ultimatum was a legitimate negotiating strategy and to find that the e-mails to Devonshire requesting extensions of the Standstill were inaccurate and misleading.

(c) Did the trial judge err by focusing his misrepresentation analysis on Devonshire rather than on the Caisse?

[119] Barclays argues that the trial judge erred by focussing his analysis of the effect of the misrepresentation on Devonshire rather than the Caisse. Barclays submits that Devonshire was not involved in the restructuring negotiations and that Devonshire would have taken direction from the Caisse as the noteholder with the most significant stake in the restructuring negotiations.

[120] We disagree. Simply put, Barclays could not and did not ignore the structure and essential terms of the Agreements. The Agreements clearly required Barclays to deal with Devonshire (through its sponsor Quanto) with respect to extensions of the Standstill and all issues concerning Events of Default and Early Termination. Barclays did precisely that and, indeed, it was a vital component of Barclays’ strategy to deal separately with the Caisse and with Devonshire.

[121] Devonshire and the Caisse were distinct parties with distinct roles under the Agreements and in the restructuring negotiations. While their interests may have overlapped on many or even most issues, they did not stand in each other’s shoes with respect to the provisions in the Agreements relating to Events of Default and Early Termination. Communication with the Caisse, which was not a party to the Agreements, could not, in law, be a substitute for communication with Devonshire for those purposes.

(d) Did the trial judge err in finding that Devonshire had reasonably relied to its detriment on Barclays’ misrepresentation?

[122] Barclays argues that the trial judge erred in finding that Devonshire relied on the extension e-mails to its detriment and suggests that Devonshire was not worse off than if it had known about the ultimatum. According to Barclays, if Devonshire had known that the negotiations to restructure were effectively at an end, it would not and could not have done anything differently.

[123] Barclays contends that even if it had wanted to, Devonshire would not have been able to do anything that made a difference to the outcome. Barclays’ position is that Devonshire would have done as the Caisse directed, and the Caisse was fully aware of the status of the negotiations. There was evidence that the Caisse thought that Barclays was acting in good faith, but no evidence that the Caisse told Devonshire that it intended to respond to the ultimatum before the January 12 deadline. If the Caisse had told Devonshire that, Devonshire would not have objected to the extension of the Standstill. Moreover, even if Devonshire had objected to the extension, it could not have terminated the swaps.

[124] We are not persuaded that the trial judge erred in finding that Devonshire had reasonably relied on the misrepresentation to its detriment, for the following reasons.

[125] One of the Devonshire witnesses, found to be reliable by the trial judge, testified that the only reason Devonshire did not terminate the swaps in 2007 was Barclays’ agreement to negotiate with the investors in good faith arising out of the Montreal Accord. The trial judge accepted evidence that if Devonshire had known the negotiations were effectively at an end in January 2009, then “the only option would have been to protect the assets of Devonshire and take steps to terminate the swaps”: at para. 163.

[126] The trial judge found that it was “quite evident” that this was why Barclays had not disclosed the true state of affairs to Devonshire: Barclays did not want Devonshire to terminate the trades before Barclays did: at para. 175. Indeed, a Barclays witness admitted as much at trial.

[127] We see no basis to interfere with the trial judge’s finding of detrimental reliance. It is well-grounded in the evidence and it reveals no error of law. It was, in our view, essentially a factual finding that was amply supported in the evidence and is entitled to deference.

(e) Misrepresentation: Conclusion

[128] We are not persuaded that the trial judge erred in finding that the Barclays’ e-mails of January 8 and 9, 2009 requesting extensions of the Standstill contained material misrepresentations as to the state of the negotiations with the Caisse, that Devonshire relied upon those misrepresentations to its detriment in agreeing to the extensions, and that it follows that the extensions were properly rescinded.

(2) Breach of Duty of Good Faith

[129] The trial judge found that Barclays was not entitled to rely on its purported cure of its default on the Liquidity Facility because it had breached its good faith obligations in January 2009. Barclays submits the trial judge’s finding that it had breached its duty of good faith, if such an obligation even existed, constitutes an error warranting appellate intervention.

[130] The breach of good faith issue is closely linked with the issue of misrepresentation.

[131] The trial judge’s findings as to good faith clearly do not rest on the imposition of a general or stand-alone duty to bargain in good faith. It is well-established that a party to a commercial agreement is entitled to enforce the agreement to its own advantage according to its terms, including rights of termination: Agribrands Purina Canada Inc. v. Kasamekas, 2011 ONCA 460, 106 O.R. (3d) 427, at paras. 50-51.

[132] However, there are two features of the relationship between Barclays and Devonshire that support the duty of good faith found by the trial judge.

[133] First, as we have already noted, as a signatory to the Montreal Accord, Barclays expressly agreed to engage in good faith negotiations to restructure the Devonshire notes. While the parties were outside the framework of the Montreal Accord by January 2009, the trial judge did not err in concluding that just as the effort to negotiate a restructuring continued, so too did the obligation to do so in good faith. Contrary to Barclays’ submission, the Montreal Accord was not merely an “agreement to agree”. It was a framework designed to allow the parties to deal with complex restructuring arrangements in turbulent times that imposed contractual good faith obligation on Barclays. The trial judge made no error in finding that both the Suspension Notice, which was explicitly conditioned on Barclays’ compliance with its obligations under the Montreal Accord, and the agreements to extend the Standstill, imported Barclays’ duty to Devonshire to carry out the negotiations with the Devonshire noteholders in good faith. As the trial judge pointed out, it would not have made commercial sense to Devonshire to agree to the extension of the Standstill if Barclays was not dealing with the noteholders in good faith.

[134] Second, as the trial judge noted, this court has endorsed as an established principle that a duty of good faith arises when necessary to ensure that the parties do not act in a way that defeats the objects of the very contract the parties have entered. As O’Connor A.C.J.O. explained in Transamerica Life Canada Inc. v. ING Canada Inc. (2003), 2003 CanLII 9923 (ON CA), 68 O.R. (3d) 457, at para. 53:

[C]ourts have implied a duty of good faith with a view to securing the performance and enforcement of the contract made by the parties, or as it is sometimes put, to ensure that parties do not act in a way that eviscerates or defeats the objectives of the agreement that they have entered into…. [Citations omitted.]

Similarly, in Nareerux Import Co. Ltd. v. Canadian Imperial Bank of Commerce, 2009 ONCA 764, 97 O.R. (3d) 481, at para. 69, Blair J.A. wrote:

[T]he jurisprudence establishes that there is an implied contractual duty of good faith not to act in a way that defeats or eviscerates the very purpose and objective of the agreement…. [Citations omitted.]

[135] Barclays was entitled to act in its own self-interest. In doing so, however, it had to be honest and candid with Devonshire and had to act in a way that would not defeat or eviscerate the very purpose of the Agreements. Barclays did neither.

[136] Barclays breached its duty of good faith in two ways. First, Barclays’ misrepresentation induced Devonshire to agree to the extension of the Standstill and thereby precluded Devonshire from taking the steps it would otherwise have taken to enforce its rights. The trial judge found that “Devonshire was entitled to the true facts from Barclays, which it did not receive”: at para. 309. While Barclays was not required to disclose its litigation strategy in advance, the trial judge found that when making statements of fact to Devonshire, it was obligated to ensure that those facts were not misleading, either directly or by omission.

[137] Second, Barclays’ purported cure payment and the manner in which Barclays purported to terminate the Agreements defeated and eviscerated the very purpose of the agreement. The crucial facts relating to the liquidity payments support that conclusion.

[138] Just prior to 9:00 a.m. on January 13, 2009, Barclays wired $71 million plus interest to Devonshire’s bank, representing the amount of the liquidity payments it had refused to pay in August 2007. The payment was made in a manner that would take Devonshire completely by surprise. Although Barclays arranged to make the payment on January 9, 2009, Barclays’ staff had been instructed, contrary to the usual practice, not to tell Devonshire that Barclays was planning to wire funds.

[139] Within minutes of transferring the funds to Devonshire’s bank and before the funds had been credited to Devonshire’s account, Barclays delivered notices to Devonshire stating it had made arrangements for the liquidity payments and purporting to designate an Early Termination Date under the ISDA Master Agreement and to name Devonshire as the Defaulting Party. The liquidity payment notice asserted that the payment was made without admission that Devonshire’s Event of Default notice of August 2007 was proper.

[140] Before the funds had been credited to Devonshire’s account and approximately 30 minutes after the transfer, Barclays advised Devonshire’s bank that it had security over the funds it had just transferred.

[141] Within hours of making the payment, Barclays issued and served its statement of claim commencing this action, and claiming, inter alia, injunctive relief and payment in relation to the collateral and the return of the liquidity funds it had just transferred.

[142] The funds were not actually credited to Devonshire's bank account until around 11:00 a.m. that same morning. On the same day, at 2:22 p.m., Devonshire delivered a Notice of Early Termination to Barclays, itself purporting to designate an Early Termination Date under the ISDA Master Agreement and to name Barclays as the Defaulting Party because of Barclays’ failure to pay the liquidity calls made by Devonshire. Devonshire subsequently filed a statement of defence and counterclaim in this action, alleging that Barclays, not Devonshire, had breached the Agreements.

[143] The trial judge found that as the payment purporting to provide liquidity was immediately followed by the termination of the swaps and litigation, Barclays’ payment could not possibly achieve the purpose for which it was designed.

[144] The liquidity payments were not made for the stated purpose of curing Barclays’ liquidity default, but rather in furtherance of Barclays’ strategy to grab the collateral. The trial judge found that “[t]he payment was not a good faith exercise with a view to securing the performance and enforcement of the contract made by the parties, but rather one that defeated the objectives of the agreement”: at para. 313.

[145] In our view, the evidence fully supports that finding. The purpose of the Liquidity Facility was to provide Devonshire with the funds it required to pay the noteholders in the event of a market disruption. The manner in which Barclays made the payment is entirely inconsistent with that purpose. The payment was combined with a refusal to admit responsibility to make the payment, an assertion of security over the payment, a demand in the form of a lawsuit for their immediate return, and a notice terminating the Agreements on the basis of the very insolvency the payment was supposed to cure. To accept Barclays’ payment as a cure in these circumstances would gut a fundamental part of the Agreements of its meaning and purpose. Barclays’ payment was simply meaningless as a cure for its default and served as part of Barclays’ strategy to set itself up to terminate the swaps. Recognizing the liquidity payments as curing Barclays’ default would undermine the very purpose for which the right to cure was granted.

[146] As with the misrepresentation findings, the trial judge’s findings as to Barclays’ intentions and the strategy it developed to terminate the Devonshire swaps on the most favourable terms possible are essentially factual in nature. There was ample evidence to support those findings which were only made after a careful and detailed review of the evidence. Those findings bring the case within a well-recognized legal principle that justified the trial judge’s conclusion that Barclays’ Notice of Early Termination was fatally infected by a breach of duty of good faith. We see no basis for appellate intervention.

(3) Benefit from Own Wrong

[147] As an alternative basis for upholding the trial judge’s finding that Barclays’ Notice of Early Termination was ineffective, Devonshire argues that Barclays should not be able to rely on Devonshire’s insolvency to terminate the swaps since it would allow Barclays to benefit from its own wrong. Insofar as the Class A notes are concerned, the trial judge agreed and found that Barclays could not terminate the swaps on the ground that Devonshire was insolvent when Barclays caused that insolvency by failing to make the liquidity payments, since it would allow Barclays to take advantage of its own wrongdoing. He restricted his finding, however, to the Class A notes because Barclays did not have a liquidity obligation with respect to the Class E and Class FRN notes.

[148] As we have already noted, it must be assumed that Barclays was in breach of its obligation to make the liquidity payments from August 2007 as a consequence of the bifurcation order.

[149] It is a widely-recognized principle that a party is precluded from taking advantage of and benefitting from a state of affairs produced by its own wrong. As this court put it in Southcott Estates Inc. v. Toronto Catholic District School Board, 2010 ONCA 310, 104 O.R. (3d) 784, at para. 13:

It is a well-established principle of contract law that a party cannot use its own breach or default in satisfying a condition precedent as a basis for being relieved of its contractual obligations…. [Citations omitted.]

As the House of Lords explained in Alghussein Establishment v. Eton College, [1991] 1 W.L.R. 587 H.L. (Eng.), at p. 594:

[N]o man can take advantage of his own wrong…. A party who seeks to obtain a benefit under a continuing contract on account of his breach is just as much taking advantage of his own wrong as is a party who relies on his breach to avoid a contract and thereby escape his obligations.

[150] Barclays’ default put Devonshire in the position that it could not pay the noteholders. That, in turn, caused Devonshire to take steps to protect the interests of its noteholders by agreeing to the Standstill and later the Montreal Accord. Insofar as the Class A notes are concerned, the trial judge held that because Barclays was assumed to be in default in failing to make the liquidity payments, Barclays would impermissibly benefit by its own breach or wrong by relying on Devonshire’s insolvency as an Event of Default.

[151] We see no reason to interfere with that finding.

[152] We reject Barclays’ characterization of its own actions as simply achieving its legitimate self-interest as contemplated by the Agreements. We agree with the trial judge’s finding that it was appropriate to imply a term prohibiting Barclays from relying upon its own failure to respond to Devonshire’s market disruption notice in order to give business efficacy to the contract. This finding, in our view, bears close relation to the matter of good faith that we have just canvassed. Both findings rest on the need to interpret and apply the Agreements in a manner that ensures they are carried out in manner that corresponds to their terms and objectives.

[153] We do not agree with Barclays’ submission that because the Standstill Agreement postponed Devonshire’s liquidity claim, Barclays’ failure to make the liquidity payments cannot be regarded as a “wrong”. In our view, that submission ignores the fact that the Standstill Agreement was explicitly without prejudice to Devonshire’s position that Barclays was in default.

[154] The trial judge restricted his finding to the Class A notes having a value of $209 million, as it was only in relation to those notes that Barclays’ liquidity obligation extended. He found that as Barclays did not have a liquidity obligation with respect to the Class E and Class FRN notes, it was Devonshire’s failure to pay, not Barclays’ liquidity default, which led to those notes not being paid and Devonshire’s insolvency as a consequence.

[155] Devonshire cross-appeals that finding and argues that the trial judge erred by refusing to hold that Barclays was precluded from relying on the insolvency that flowed from Barclays’ failure to live up to its liquidity obligation.

[156] Barclays knew that the Class A notes (which were protected by the liquidity line and started to come due first) were ranked pari passu with the Class E and Class FRN notes. Interest became payable on one group of Class FRN notes on August 16, 2007. Although Devonshire had the funds to make the interest payment on those Class FRN notes on August 16, it did not pay the interest because it decided that it could not treat one class of noteholders differently from the others. The Class E notes that came to maturity were extended, and when the interest became payable on them, it was not paid for the same reason.

[157] We agree with Devonshire’s submission that the failure to pay interest on the Class E and Class FRN notes is properly seen as part of a series of inter-connected events closely related to Barclays’ failure to pay its liquidity obligations. Because the notes were ranked pari passu, Devonshire’s Issuer Trustee could not pay one class of notes while not paying another. Devonshire’s default in paying interest on the Class E and Class FRN notes was a direct consequence of Barclays’ failure to make the liquidity payments.

[158] The trial judge rejected Devonshire’s argument that the principle that a party should not be entitled to benefit from its own wrong should operate as a bar to Barclays relying on Devonshire’s insolvency with respect to the Class E and Class FRN notes because he found that it was Devonshire’s decision not to pay the interest on the Class E and Class FRN notes. He wrote, at para. 246:

The liquidity obligation did not apply to those notes. Also, it was a decision by Devonshire, not Barclays, which led to those notes initially not being paid because they ranked pari passu with the Class A notes. [Emphasis added.]

[159] Respectfully, in our view, this finding ignores the obligation on Devonshire to honour the pari passu ranking. As the trial judge noted, this ranking was established by the Series A Supplemental Indenture, which was the subject of negotiations with Barclays. In these circumstances, Barclays must have known that its failure to make the liquidity payments would result in Devonshire withholding interest payments on all classes of notes.

[160] In our view, Barclays’ failure to make the liquidity payments was a material contributing cause of Devonshire’s insolvency in relation to all three classes of notes. Accordingly, Barclays is barred by its own wrong from relying on Devonshire’s insolvency as a basis for terminating the transaction.

[161] Given our conclusion with respect to misrepresentation and bad faith, this is simply an additional basis for holding that Barclays’ Notice of Early Termination was invalid.

(4) Timing of the Cure Payment

[162] The trial judge provided detailed reasons for finding that Barclays’ Notice of Early Termination was invalid because it was delivered before the funds were actually credited to Devonshire’s account. Barclays submits that he erred in so holding.

[163] In our view, it is not necessary for this court to decide this issue given our conclusion with respect to Barclays’ breach of good faith. We have concluded that given the manner and timing of Barclays’ purported cure payment, it could not possibly achieve the purpose for which it was designed. That conclusion would not be altered were we to find that Devonshire should be deemed to have received funds shortly after 9:00 a.m. before those funds were credited to its account at about 11:00 a.m. Indeed, Barclays’ submission on this point – that Devonshire should be deemed to have received the payment before the funds had found their way into Devonshire’s account – demonstrates the highly artificial nature of its overall position on the “cure” payment.

(5) Waiver and Election

(a) Did Barclays waive its right to remedy its default?

[164] At trial, Devonshire argued that Barclays had waived its right to remedy its default by failing to meet Devonshire’s demands to provide liquidity on August 13, 14 and 15, 2007. Devonshire’s position was that Barclays, by its failure, waived the right to cure that default by making the payment on January 13, 2009. Devonshire says that this waiver is another reason why Barclays was in default on January 13, 2009 and had no right to deliver a Notice of Early Termination.

[165] The trial judge rejected this argument. He concluded that Barclays had no intention to make the liquidity payments demanded by Devonshire in August 2007 because Barclays did not believe that an MDE had occurred. Consequently, Barclays did not consider itself obliged to make the liquidity payments. However, the trial judge concluded that the evidence could not sustain the conclusion that Barclays’ intention would have remained the same even if its position regarding the occurrence of an MDE were wrong. The trial judge went on to conclude that while Barclays communicated its intention not to pay in August 2007, it did not communicate to Devonshire that it waived its right to remedy its default later on.

[166] Devonshire renews its argument in this court. In our opinion, it is simply disposed of. Both conclusions of the trial judge referred to are findings of fact for which there was ample evidence. There is no basis for us to interfere with them. We agree with the trial judge that these findings create an insuperable obstacle for Devonshire’s argument. Barclays did not have the unequivocal intention to relinquish its right to meet Devonshire’s liquidity demands in the future. Nor did it communicate such an intention to Devonshire. These two elements of waiver were missing: see Saskatchewan River Bungalows Ltd. v. Maritime Life Assurance Co., 1994 CanLII 100 (SCC), [1994] 2 S.C.R. 490, at pp. 499-500.

(b) Did Barclays elect to abandon its right to rely on the insolvency of Devonshire?

[167] At trial, Barclays’ position was that Devonshire was insolvent on January 13, 2009 and that, under the ISDA Master Agreement, the insolvency constituted an Event of Default, which entitled Barclays to deliver a Notice of Early Termination on that day.

[168] Barclays did not rely on any Event of Default prior to that date. However, as the trial judge described it, Barclays’ case at trial was that Devonshire’s insolvency commenced in August 2007 and continued throughout the Standstill Period.

[169] At the outset of the trial, the trial judge ruled that because of the scope of the bifurcation order, Barclays could not lead evidence on the issue of Devonshire’s insolvency prior to August 16, 2007, the date of the Suspension Notice. However, the evidence left no doubt that Devonshire was insolvent on January 13, 2009. In addition, throughout the Standstill Period, Barclays continued to pay premiums to Devonshire for credit protection. Barclays did so despite saying Devonshire was insolvent during the Standstill Period and despite the provision in the ISDA Master Agreement that its obligation to make these payments was subject to the condition precedent that no Event of Default or Potential Event of Default (such as insolvency) had occurred with respect to Devonshire.

[170] In these circumstances, the trial judge concluded that, by continuing to make these payments during the Standstill Period, Barclays elected not to exercise its right to refuse to make these payments and thereby elected to keep the swap contracts in effect. He held that Barclays was thus estopped from relying on Devonshire’s insolvency as grounds for terminating the swap contracts. The trial judge concluded that Barclays therefore did not have the right on January 13, 2009 to terminate the swap contracts on the basis of Devonshire’s insolvency on that date.

[171] In addition, the trial judge held that the non-waiver provisions in the ISDA Master Agreement were not strong enough to preclude his conclusion that Barclays had elected to waive the right to rely on Devonshire’s insolvency as an Event of Default. These provisions include s. 9(b), which requires that a waiver be in writing and executed by the parties, and s. 9(f), which provides that a failure or delay in exercising any right under the ISDA Master Agreement will not be presumed to operate as a waiver.

[172] Barclays attacks that conclusion in this court. Its principal argument is that, while the trial judge correctly articulated the legal principles of election, he erred in their application. Barclays argues that trial judge committed a legal error when he found that, because Barclays had paid premiums to Devonshire during the Standstill Period, Barclays had elected not to terminate the swap contracts on the basis of Devonshire’s insolvency after the Standstill Period came to an end. Barclays says that there is no inconsistency between making payments during the Standstill Period and eventually relying on an Event of Default after the conclusion of that Period. Barclays also argues that the trial judge erred in failing to apply the non-waiver provisions in the ISDA Master Agreement. Finally, Barclays argues that Devonshire failed to raise the election issue in its pleadings and should therefore not have been permitted to advance it at trial. Devonshire joins issue with Barclays on all of these arguments.

[173] Our analysis begins with the relevant legal principle, on which both parties agree. They cite and rely on this court’s decision in Charter Building Company v. 1540957 Ontario Inc. (Mademoiselle Women’s Fitness & Day Spa), 2011 ONCA 487, 107 O.R. (3d) 133, where Epstein J.A. put it this way, at para. 19:

Election at common law takes place where a party is faced with a choice between two inconsistent courses of action that affect another party’s rights or obligations, and knowing that the two courses of action are inconsistent and that he or she has the right to choose between them, makes an unequivocal choice and communicates that choice to the other party. The doctrine provides that the party making the election is afterwards precluded from resorting to the course of action that he has rejected. The election is effective at the point of communication on the basis that the parties to an ongoing relationship are entitled to know where they stand …. [Citation omitted.].

[174] Two provisions of the ISDA Master Agreement are also germane to this issue. Section 5(a)(vii) describes events that constitute an Event of Default, one of which is the insolvency of a party. The relevant provision, repeated for ease of reference, reads as follows:

5. Events of Default and Termination Events

(a) Events of Default. The occurrence at any time with respect to a party … of any of the following events constitutes an event of default (an “Event of Default”) with respect to such party:─

(vii) Bankruptcy. The party … (2) becomes insolvent or is unable to pay its debts or fails or admits in writing its inability generally to pay its debts as they become due….

[175] Section 2(a) sets out provisions regarding the requirement to pay money. The relevant part of the section is this:

2. Obligations

(a) General Conditions.

(i) Each party will make each payment or delivery specified in each Confirmation to be made by it….

(iii)Each obligation of each party under Section 2(a)(i) is subject to (1) the condition precedent that no Event of Default or Potential Event of Default with respect to the other party has occurred and is continuing….

[176] The consequence of this provision is, as the trial judge found, that Barclays’ obligation to make payments to Devonshire for credit protection was subject to the condition precedent that there was no insolvency Event of Default on the part of Devonshire.

[177] We agree with the trial judge that during the Standstill Period, Barclays could have refused to make the premium payments on the basis of Devonshire’s insolvency. By making the payments, however, Barclays chose a course of action inconsistent with its option not to pay. Each payment constituted evidence of an election by Barclays to waive its right not to pay due to Devonshire’s insolvency at that point in time.

[178] However, we disagree with the trial judge that Barclays’ payments during the Standstill Period constituted its election to waive its right to rely on Devonshire’s insolvency forever, including after the conclusion of the Standstill Period. At each point in time, Barclays’ choice was between, on the one hand, making the credit protection payments and, on the other, relying on Devonshire’s insolvency at that point and refusing to pay.

[179] It was not inconsistent for Barclays to make those payments during the Standstill Period and yet, when the Standstill Period ended, to rely on Devonshire’s insolvency then. As a matter of law, Barclays retained that right. Barclays could not have used Devonshire’s insolvency after the Standstill Period ended to excuse its payment obligations during the Standstill Period. By making the credit protection premium payments during the Standstill Period, Barclays did waive its right not to do so because of Devonshire’s insolvency at the time those payments were made. However, making the payments during the Standstill Period was not inconsistent with the right Barclays later acquired, to act on the Event of Default and issue a Notice of Early Termination on the basis of Devonshire’s insolvency after the Standstill Period ended.

[180] We therefore conclude that the trial judge erred in finding that Barclays had elected to waive its right to rely on Devonshire’s insolvency on January 13, 2009, after the Standstill Period had ended, as the Event of Default to justify the Notice of Early Termination it delivered that day.

[181] In light of this conclusion, it is unnecessary to address Barclays’ arguments concerning the non-waiver provisions in the ISDA Master Agreement and Devonshire’s failure to plead the election issue.

(6) Barclays’ Notice of Early Termination: Conclusions

[182] The issues of misrepresentation, bad faith and benefiting from one’s own wrong are closely linked on the facts of this case. On the trial judge’s findings, the January 8 and 9 extensions of the Standstill were rescinded because of Barclays’ fraudulent misrepresentation arising from the ultimatum. Whether Barclays’ purported cure payment of January 13 was timely or not, it was invalid because of Barclays’ breach of its duty of good faith. The breach of that duty, in turn, was closely tied to the misrepresentation.

[183] As we have explained, it is our view that the findings as to misrepresentation and bad faith are unassailable on this appeal. While we conclude that Devonshire gets no help from its arguments that Barclays waived its right to cure its failure to make the liquidity payments demanded in August and that Barclays elected to waive its right to rely on Devonshire’s insolvency as an Event of Default, the findings we view as unassailable render Barclays’ Notice of Early Termination invalid.

[184] Simply put, Barclays was not a Non-defaulting Party within the meaning of s. 6(a) of the ISDA Master Agreement with respect to Devonshire’s insolvency. Barclays’ failure to make the liquidity payments was directly related to Devonshire’s insolvency, the very Event of Default that Barclays relied on. Indeed, Barclays’ entire strategy in purporting to “cure” its default prior to delivering the Notice of Early Termination amounts to an admission that if it did not make the promised liquidity payments it could not rely on Devonshire’s insolvency as an Event of Default.

[185] This conclusion is related to and fortified by the principle that a party cannot benefit from its own wrong. Barclays’ failure to make the liquidity payments was an Event of Default and a significant cause of Devonshire’s insolvency. From both the perspective of the common law and the language of s. 6(a), that precluded Barclays from delivering a valid Notice of Early Termination.

[186] We conclude, accordingly, that the trial judge did not err in finding that Barclays' Notice of Early Termination was invalid.

B) Issue 2. Is Devonshire's Notice of Early Termination valid?

(1) The Validity of Devonshire’s Notice of Early Termination

[187] We begin our discussion in this section by noting once again that, as a result of the bifurcation order, the following facts are assumed: (i) there was an MDE in August 2007; (ii) Devonshire’s market disruption notices and notice of default were valid; and (iii) Barclays was in default under the notices sent by Devonshire up to August 16, 2007 when the Standstill Agreement was entered into.

[188] The trial judge found that Devonshire’s Notice of Early Termination with respect to Barclays’ failure to make the liquidity payments was valid. Although he concluded that Devonshire was insolvent and that, under s. 6(a) of the ISDA Master Agreement, only a Non-defaulting Party can deliver a valid Notice of Early Termination, the trial judge found that Barclays had elected to waive reliance on Devonshire’s insolvency. He therefore concluded that, for the purposes of s. 6(a) of the ISDA Master Agreement, Devonshire was a Non-defaulting Party and entitled to deliver a valid Notice of Early Termination.

[189] On appeal, we have rejected the trial judge’s finding that Barclays elected to waive reliance on Devonshire’s insolvency. Barclays adopts the trial judge’s implicit conclusion that, under s. 6(a) of the ISDA Master Agreement, only a Non-defaulting Party can deliver a valid Notice of Early Termination. As the trial judge held that Devonshire was insolvent on January 13, 2009, Barclays submits that Devonshire’s Notice of Early Termination is invalid.

[190] We do not accept Barclays’ submission or the trial judge’s apparent assumption that, but for waiver, Devonshire’s insolvency would have precluded Devonshire from delivering a valid Notice of Early Termination. In our view, that submission cannot be sustained in the face of the wording of the relevant portions of s. 6(a) of the ISDA Master Agreement, repeated for ease of reference:

6. Early Termination

(a) Right to Terminate Following Event of Default. If at any time an Event of Default with respect to a party (the “Defaulting Party”) has occurred and is then continuing, the other party (the “Non-defaulting Party”) may, by not more than 20 days notice to the Defaulting Party specifying the relevant Event of Default, designate a day not earlier than the day such notice is effective as an Early Termination Date in respect of all outstanding Transactions.

[191] A party becomes a “Defaulting Party” when an Event of Default occurs in respect to that party, but an Event of Default does not automatically entail a termination. A Non-defaulting Party may deliver a Notice of Early Termination, but need not. If no valid Notice of Early Termination is delivered, by their terms, the Agreements are not terminated.

[192] As Firth puts it: “Where an Event of Default takes place, the party with respect to which it has occurred is referred to as the ‘Defaulting Party’ and the other as the ‘Non-defaulting Party’”: Simon Firth, Derivatives Law and Practice, loose-leaf (London: Sweet & Maxwell, 2012), at para. 11.044. The Event of Default is a condition precedent which gives the Non-defaulting Party the option to serve a Notice of Early Termination.

[193] As we read the words of s. 6(a), the terms Defaulting Party and Non-defaulting Party refer to a specific Event of Default. At issue on this appeal are two Events of Default: Barclays’ failure to make the liquidity payments and Devonshire’s insolvency. We are not persuaded that Devonshire’s insolvency made it ineligible to serve a valid Notice of Early Termination under s. 6(a) of the ISDA Master Agreement based on Barclays’ failure to make the liquidity payments.

[194] As a practical matter, there can be more than one Event of Default at any given time so that each party can be both a Defaulting Party and a Non-defaulting Party with respect to different Events of Default. If they are both non-defaulting parties with respect to different purported Events of Default, each party has the right to serve a Notice of Early Termination on the other. Firth refers to “simultaneous default” and provides the following illustration: “If neither party performed on the due date, both would be in breach of contract. Either could then give notice of the failure to pay or deliver to the other and, if the default is not cured by the end of the grace period, close out the outstanding transactions”: at para. 11.049.

[195] In our view, it follows that there is nothing in the Agreements that precludes a party in default under some other obligation from delivering a valid Notice of Early Termination if it is a Non-defaulting Party with respect to the specified Event of Default. Devonshire is a Non-defaulting Party with respect to Barclays’ failure to make the liquidity payments. It follows that under the ISDA Master Agreement, even if Devonshire’s insolvency made it a Defaulting Party with respect to the Event of Default of insolvency, its insolvency did not preclude it from delivering an effective Notice of Early Termination to Barclays under s. 6(a) of the ISDA Master Agreement with respect to Barclays’ failure to make the liquidity payments.

[196] Indeed, as we pointed out in our analysis of the validity of Barclays’ Notice of Early Termination, Barclays’ failure to make the liquidity payments was a material contributing cause of Devonshire’s insolvency in relation to all three classes of notes. It follows, in our view, that both as a matter of strict interpretation of the language of s. 6(a) and as a matter of interpreting the Agreements to give their terms commercial efficacy consistent with established common law principles, Devonshire’s insolvency did not preclude it from delivering a valid Notice of Early Termination.

(2) Repudiation

[199] The trial judge concluded his discussion of whether Devonshire’s Notice of Early Termination of January 13, 2009 was valid by addressing the issue of repudiation. He found that even if Devonshire was not able to deliver a Notice of Early Termination under s. 6(a) of the ISDA Master Agreement, the doctrine of repudiation brought the contract to an end on that day. Barclays’ conduct, including delivering its own Notice and immediately commencing litigation following delivery, constituted a firm refusal to perform its obligations under the contract. Such conduct evinced Barclays’ clear intention not to be bound by the contract. Devonshire, by its Notice of Early Termination, indicated that it too regarded the contract as at an end and thus accepted Barclays’ repudiation.

[200] Barclays contests the trial judge’s conclusion. It argues that, by its actions, it sought to affirm the contract, rather than to resile from it. In addition, Barclays argues that Devonshire did not accept the repudiation, but instead attempted to keep the contract alive. Finally, Barclays says that, in any event, the express terms of the ISDA Master Agreement exclude the common law doctrine of repudiation.

[201] Devonshire, on the other hand, says that Barclays’ conduct made clear that the bank would no longer perform. Barclays’ mistaken belief that it was enforcing its rights under the contract does not negate the finding of repudiation. Moreover, Devonshire’s reliance on the termination provisions of the contract in its own Notice of Early Termination is not inconsistent with its acceptance of Barclays’ repudiation. Although the repudiation and its acceptance bring the contract to an end, these termination provisions survive. Finally, Devonshire says that the ISDA Master Agreement clearly contemplates common law remedies. It does not exclude them.

[202] We agree with Devonshire’s position. We see no basis to disturb the trial judge’s conclusion.

[203] In finding that Barclays’ conduct constituted a repudiation of the contract, the trial judge applied the proper legal analysis. He assessed whether that conduct, viewed objectively, evinced an intention not to be bound by the contract. He correctly set aside as immaterial Barclays’ mistaken belief that it was exercising its contractual right. We agree with the trial judge that Barclays’ actions on January 13, 2009 constituted a clear refusal to perform and that this conduct demonstrated Barclays’ intention no longer to be bound by the contract. The record fully supports these findings and they deserve deference in this court. We would not interfere with them.

[204] We would apply the same analysis to the trial judge’s conclusion that Devonshire accepted Barclays’ repudiation, and would not interfere with it. There was ample evidence to sustain the finding that, by its conduct, Devonshire clearly and unequivocally signalled that it accepted the repudiation. Devonshire did not assert that it rejected the repudiation nor did it elect to have the contract continue. Devonshire’s reliance on the termination provisions of the contract is not inconsistent with that conclusion. While the accepted repudiation discharged future obligations, it did not eliminate the provisions that deal with the breach or termination of the contract. They survive, and it is those provisions that Devonshire sought to rely on when it accepted the repudiation.

[205] Finally, we disagree with Barclays that the contract ousts the common law of repudiation. Plain language would be required to do so: see Modern Engineering (Bristol) Ltd. v. Gilbert Ash (Northern), [1974] A.C. 689 H.L. (Eng.), at pp. 716-17. Not only is this contract devoid of any such language, but we may draw an inference the other way. Section 9(d) of the ISDA Master Agreement provides that the rights given in the contract do not exclude any rights provided by law.

[206] In summary, we conclude that the trial judge was correct that, even if Devonshire was not able to deliver a valid Notice of Early Termination, the common law principle of repudiation resulted in the termination of the contract, thereby discharging the parties from their future obligations under it.

C) Issue 3. Did the trial judge err in his determination of Barclays’ Settlement Amount?

(1) Introduction

[207] As we have said, the trial judge found that Barclays’ Notice of Early Termination was invalid; that Devonshire’s Notice of Early Termination was valid; and that, in these circumstances, under the Intercreditor Agreement, any amount payable to Barclays on Early Termination was subordinated to the amount owing to Devonshire noteholders for principal and interest.

[208] Because of these findings, it was not strictly necessary that the trial judge determine the amount, if any, that would have been payable to Barclays under its Early Termination Notice had it been the Non-defaulting Party.

[209] Nonetheless, because determination of the issue could be important for the second part of the bifurcated trial, the trial judge chose to address it.

[210] Under the terms of the ISDA Master Agreement, the parties had the option of choosing various methods of determining the amount payable on Early Termination of the CDSs. They chose a method called Second Method and Market Quotation.

[211] Under the Second Method and Market Quotation formula, the amount payable on Early Termination is the sum of the Settlement Amount (as determined by the Non-defaulting Party) plus net Unpaid Amounts (the difference between Unpaid Amounts owing to the Non-defaulting Party and Unpaid Amounts owing to the Defaulting Party). As the net Unpaid Amounts were not disputed, in the event Barclays was the Non-defaulting Party, the amount payable on Early Termination turned on Barclays’ determination of the Settlement Amount.

[212] Where the Settlement Amount plus net Unpaid Amounts is a positive figure, the Defaulting Party pays that amount to the Non-defaulting Party; if the Settlement Amount plus net Unpaid Amounts is a negative figure, the Non-defaulting Party pays the Defaulting Party that figure.

[213] Under the ISDA Master Agreement, the Settlement Amount is the Market Quotation amount for the terminated transaction, so long as a Market Quotation amount can be determined and produces a commercially reasonable result. To determine a Market Quotation amount, a Non-defaulting Party must obtain at least three market quotations of the amount that would have to be paid essentially to replace the terminated transaction.

[214] Where a Market Quotation amount cannot be obtained or would not produce a commercially reasonable result, the Settlement Amount is deemed to be the Non-defaulting Party’s Loss in relation to a terminated transaction.

[215] Under the ISDA Master Agreement, Loss is defined to include the amount a party “reasonably determines in good faith to be its total losses and costs … including any loss of bargain”.

[216] In this case, Barclays was unable to obtain any firm market quotations for the terminated CDSs. Accordingly, using a model it had used to value the CDSs on a daily basis prior to August 2007, Barclays calculated its Loss at $1.2 billion.

[217] The trial judge rejected Barclays’ calculation. He found that Barclays’ model did not value the CDSs taking account of all their relevant features. In addition, he concluded that Barclays had not established that $1.2 billion was a commercially reasonable figure or that it reflected a value the market would have placed on the CDSs.

[218] Instead, the trial judge accepted Devonshire’s expert’s opinion that a discounted cash flow method of valuation was reasonable in the circumstances. Relying on this model, the trial judge valued Barclays’ Loss at $12,000 but reduced it to $0 based on mitigation. In arriving at this value, the trial judge declined to adopt Devonshire’s expert’s view that the $12,000 figure should be increased by a $264 million risk premium. Nor did he adopt her view that additional liquidity payments Barclays would have had to make but for the Standstill Agreement should be deducted from the calculation.

[219] In its appeal of this aspect of the trial judge’s decision, Barclays raises multiple issues. In our view, these issues boil down to three basic questions:

i. Did the trial judge err in rejecting Barclays’ determination of the Settlement Amount?

ii. Did the trial judge err in valuing Barclays’ Loss at $12,000?

iii. Did the trial judge err in his treatment of mitigation?

(2) Background

(a) Contractual provisions relevant to determining Barclays’ Settlement Amount

[220] Section 6(e) of the ISDA Master Agreement sets out alternate formulae contracting parties may choose for determining the amounts to be paid on Early Termination of a transaction. For termination involving an Event of Default, s. 6(e) gives parties the option of choosing a payment method (either the “First Method” or the “Second Method”) and also a payment measure (either “Market Quotation” or “Loss”).

[221] When they entered into the CDS transaction, Barclays and Devonshire selected Second Method and Market Quotation as the formula that would apply for determining the payment to be made on Early Termination.

[222] Section 6(e)(i)(3) of the ISDA Master Agreement sets out the Second Method and Market Quotation formula for an Early Termination payment, being the sum of the Settlement Amount and the net Unpaid Amounts:

6. Early Termination

(e) Payments on Early Termination.…

(i) Events of Default. If the Early Termination Date results from an Event of Default … (3) Second Method and Market Quotation. If the Second Method and Market Quotation apply, an amount will be payable equal to (A) the sum of the Settlement Amount (determined by the Non-defaulting Party) in respect of the Terminated Transactions and [the net Unpaid Amounts]. If that amount is a positive number, the Defaulting Party will pay it to the Non-defaulting Party; if it is a negative number, the Non-defaulting Party will pay the absolute value of that amount to the Defaulting Party.

[223] As we have said, the net Unpaid Amounts are not in dispute. The quantum of the Early Termination payment in the event Barclays is the Non-defaulting party therefore turns on the calculation of the Settlement Amount.

[224] “Settlement Amount”, “Market Quotation” and “Loss” are defined in s. 14 of the ISDA Master Agreement. The definition of Settlement Amount is important because it makes it clear that where a Market Quotation amount cannot be determined or would not produce a commercially reasonable result, the Settlement Amount is the amount of the Non-defaulting party’s Loss in relation to that transaction:

“Settlement Amount” means, with respect to a party and any Early Termination Date, the sum of:—

(a) the Termination Currency Equivalent of the Market Quotations (whether positive or negative) for each Terminated Transaction or group of Terminated Transactions for which a Market Quotation is determined; and

(b) such party’s Loss (whether positive or negative and without reference to any Unpaid Amounts) for each Terminated Transaction or group of Terminated Transactions for which a Market Quotation cannot be determined or would not (in the reasonable belief of the party making the determination) produce a commercially reasonable result.

[225] The definition of Market Quotation is also important because it makes clear Market Quotation is a replacement value concept and that it will apply only where at least three market quotations can be obtained. The definition stipulates that the Non-defaulting Party is to request quotations from third-party market participants for an amount to be paid to or by such party for a transaction that would have the effect of preserving the economic equivalent of any payment or delivery that, but for the Early Termination of the CDSs, would have been required after the date of termination. If fewer than three quotations are provided, it will be deemed that the Market Quotation amount cannot be determined. The definition of Market Quotation, found in s. 14 of the ISDA Master Agreement, reads in part as follows:

“Market Quotation” means, with respect to one or more Terminated Transactions and a party making the determination, an amount determined on the basis of quotations from Reference Market-makers. Each quotation will be for an amount, if any, that would be paid to such party (expressed as a negative number) or by such party (expressed as a positive number) in consideration of an agreement between such party … and the quoting Reference Market-maker to enter into a transaction (the “Replacement Transaction”) that would have the effect of preserving for such party the economic equivalent of any payment or delivery (whether the underlying obligation was absolute or contingent and assuming the satisfaction of each applicable condition precedent) by the parties under s. 2(a)(i) in respect of such Terminated Transaction or group of Terminated Transactions that would, but for the occurrence of the relevant Early Termination Date, have been required after that date. For this purpose, Unpaid Amounts in respect of the Terminated Transaction or group of Terminated Transactions are to be excluded but, without limitation, any payment or delivery that would, but for the relevant Early Termination Date, have been required (assuming satisfaction of each applicable condition precedent) after that Early Termination Date is to be included.… If fewer than three quotations are provided, it will be deemed that the Market Quotation in respect of such Terminated Transaction or group of Terminated Transactions cannot be determined. [Emphasis added.]

[226] As noted above, where Market Quotation cannot be determined, the Settlement Amount is the amount of the Non-defaulting Party’s Loss. The definition makes it clear that a Non-defaulting Party’s Loss in connection with a terminated transaction includes any loss of bargain and that a party may, but need not, determine its Loss by reference to market quotations. The relevant portions of the definition, also in s. 14 of the ISDA Master Agreement, reads as follows:

“Loss” means, with respect to this Agreement or one or more Terminated Transactions, as the case may be, and a party, the Termination Currency Equivalent of an amount that party reasonably determines in good faith to be its total losses and costs (or gain, in which case expressed as a negative number) in connection with this Agreement or that Terminated Transaction or group of Terminated Transactions, as the case may be, including any loss of bargain, cost of funding or, at the election of such party but without duplication, loss or cost incurred as a result of its terminating, liquidating, obtaining or re-establishing any hedge or related trading position (or any gain resulting from any of them). Loss includes losses and costs (or gains) in respect of any payment or delivery required to have been made (assuming satisfaction of each applicable condition precedent) on or before the relevant Early Termination Date and not made.… A party will determine its Loss as of the relevant Early Termination Date or, if that is not reasonably practicable, as of the earliest date thereafter as is reasonably practicable. A party may (but need not) determine its Loss by reference to quotations of relevant rates or prices from one or more leading dealers in the relevant markets. [Emphasis added.]

(b) Barclays’ evidence relating to the Settlement Amount

[227] Mr. James Lee of Barclays was responsible for calculating Barclays’ Loss as of the Early Termination Date, January 13, 2009. Barclays claimed $1.2 billion as a Loss, including $1.02 billion representing a model-based estimate of the mid-market price of the CDSs. “Mid-market price” refers to the theoretical price at which a credit protection buyer would offer to purchase the CDSs and “offer price” refers to the price at which a credit protection seller would offer to replace the CDSs. The remaining Loss consisted of hypothetical hedging costs and the value of credit protection related to an asset-backed securities portion of the transaction that would provide additional protection if the attachment points were reached.

[228] According to the trial judge, Mr. Lee testified that, in order to determine the $1.02 billion figure, Barclays proxied the Market Quotation process, meaning that Barclays attempted to determine the replacement cost of the CDSs, assuming that Barclays stood in the place of one of the dealers that it had asked to bid.

[229] Barclays calculated the $1.02 billion figure using a proprietary model it had developed to determine what it calls a mark-to-market value of the CDSs on a daily basis during the life of the CDSs. The Barclays’ model is referred to as a Gaussian copula model because it uses assumptions about credit default spreads and applies correlations (copula) between corporate swaps using a Gaussian (normal) distribution. According to the trial judge, Mr. Lee testified that the mark-to-market value derived from the model is a representation of the market value of a derivative at a point in time using mid-market values.

[230] Barclays called two expert witnesses at trial to support Mr. Lee’s calculation of Loss: Dr. John Hull, a professor of derivatives and risk management at the University of Toronto’s Joseph L. Rotman School of Management, and Miles Draycott, a consultant and former derivatives trader.

[231] According to the trial judge, Professor Hull testified that assets can be valued in two broad ways: first, by using the valuations of related assets; or, second, by estimating expected cash flows and discounting them to the present. Professor Hull indicated that market practice is normally to use the first approach and, in particular, a Gaussian copula model. Mr. Draycott testified that the only way to value the CDSs in question here is to use a mark-to-market model.

(c) Devonshire’s evidence relating to the Settlement Amount

[232] Ms. Leslie Rahl testified as an expert witness for Devonshire concerning the quantum of the Settlement Amount. Her primary position was that the CDSs should be valued as of August 2007 rather than January 13, 2009. If August 2007 was rejected as the valuation date, November 2007 was an appropriate alternative. In the absence of the Standstill Agreement, November 2007 was when Barclays would have made additional collateral calls and, presumably, the point at which the CDSs would have terminated. If January 13, 2009 was chosen as the valuation date, she opined that the CDSs should be valued using projected real-world losses and by adding a normalized risk premium.

[233] As the trial judge selected January 13, 2009 as the valuation date and as that finding has not been appealed, we will restrict our discussion of Ms. Rahl’s evidence to her valuation evidence using that date and to her critique of Barclays’ valuation.

[234] Ms. Rahl calculated the Settlement Amount in three steps. First, she estimated the credit protection payments Barclays would have received from Devonshire over the remaining life of the CDSs by calculating the projected losses in the two reference portfolios during that period and discounting that amount to the present. This step yielded expected losses of $12,347. Second, she added a normalized risk premium in the amount of $264 million. Finally, she deducted the liquidity payments she believed Barclays would have been required to make had the Standstill Agreement not been in place.

[235] In her report, Ms. Rahl explained that, when valuing CDSs, it is common and permitted to revert to a Loss calculation when poor liquidity makes it impossible to find a valid market price. She explained that a Loss calculation involves, among other things, finding a value of “the bargain” and that Barclays’ mark-to-model method is intended to find the mid-market theoretical valuation. She confirmed that a Loss calculation also requires finding the costs of re-hedging the “bargain” in the market – and that such hedging costs are similar to moving from a mid-market valuation to a bid-side market valuation.

[236] Although Ms. Rahl acknowledged that using a mark-to-model method (her terminology for Barclays’ mark-to-market method) for valuation had become standard practice, she opined that Barclays’ Loss calculation was unreasonable for two primary reasons. First, it ignored elements of the transaction that reduced the value of the bargain. Second, it relied exclusively on a mark-to-model method of valuation at a time when other valuation methods were being used and yielding radically different results.

[237] According to Ms. Rahl, the following features of the CDSs were customized features, not forming part of a standard ISDA transaction, that ought to have been considered in determining the value of the CDSs: (i) the liquidity feature requiring the swap buyer to provide liquidity in some circumstances; (ii) the “stop-loss” provision effectively requiring the credit protection provider to either provide additional collateral or terminate the transaction if the 50 per cent collateral trigger value was reached; and (iii) the “limited-recourse provision”, which gave no recourse to Barclays to Devonshire’s assets once the collateral was exhausted.

[238] Concerning the appropriate valuation method, Ms. Rahl was of the view that a number of events leading up to January 2009 had changed the economics of the CDSs and created what she viewed as temporarily aberrant risk premiums for CDSs in general: (i) the 17-month Standstill Agreement, an unprecedented factor in addressing issues surrounding the calculation of Early Termination payments under an ISDA Master Agreement; (ii) the 2008 financial crisis, which led to extreme illiquidity in the market; and (iii) the temporary, but extraordinary, illiquidity in the market around January 2009.

[239] Ms. Rahl explained that market spreads for CDSs imply a loss that is generally much higher than the real-world loss market participants actually expect. A risk premium reflects the difference between the real-world estimate of loss and the market implied estimate of loss. Because the reference portfolios in the Devonshire CDSs consisted of super senior tranches that carried little risk of default (credit risk), ordinarily, the market implied risk premium would make up virtually the entire market value of the CDSs.

[240] According to Ms. Rahl, “[c]onsensus opinion about the appropriateness of various valuation techniques was being debated and reassessed in 2008 and 2009”. As of April 2009, the American Financial Accounting Standards Board (“FASB”) altered its previous preference for a uniform method of finding a market valuation even in a liquid market. Ms. Rahl stated: “In view of a lack of robust market pricing and very high risk premiums, FASB permitted corporations to use alternative methods, such as mark-to-model or discounted cash flow analysis, in valuing certain financial instruments”.

[241] In light of the aberrant market conditions in January 2009, in Ms. Rahl’s view, it was appropriate to consider other recognized valuation methods apart from a mark-to-model valuation. Unlike a mark-to-model valuation in which credit risk and risk premium are combined, a cash flow-based valuation separates out the credit risk and risk premium features of a mark-to-model valuation. In her view, “[t]he potential dispersion between these approaches was very large in January 2009 and the subject of much debate.”

[242] In calculating Barclays’ Loss, Ms. Rahl stated: “[W]e do not dispute that there should be a risk premium.” However, she added what she referred to as a “normalized” risk premium to her discounted cash flow credit risk figure because, in her opinion, the unusual events of late 2008 and early 2009 resulted in extraordinarily high and temporary risk premiums. She arrived at a normalized risk premium of $264 million by averaging risk premiums in the CDSs markets for August and November 2007, which she calculated using a Gaussian copula model. She also looked at the risk premium in the CDSs market in April 2010 calculated in the same way and determined it was at the same level as her average for August and November 2007.

(3) The Trial Judge’s Reasons

[243] After reviewing Barclays’ evidence concerning its calculation of Loss, the trial judge stated that the “key difference between the experts is whether, in the circumstances of this case, a mark to model or cash flow-based valuation is justifiable as of January 13, 2009”: at para. 393.

[244] The trial judge rejected the mark-to-model method as being appropriate in this case for three main reasons.

[245] First, he rejected Barclays’ argument that Loss must be calculated by using a proxy for Market Quotation because various English decisions have held that the Market Quotation measure and the Loss measure are intended to lead broadly to the same result.

[246] In this regard, the trial judge referred specifically to one of the English cases, Peregrine Fixed Income Ltd. v. Robinson Department Store Public Co. Ltd., [2000] C.L.C. 1328 (Comm.), at para. 30, in which the court commented that the Market Quotation measure and the Loss measure are intended to lead to broadly the same result. He noted that, in that case, the court concluded that the Market Quotation method led to a commercially unreasonable result and that it was agreed that the figure for Loss was the present value of the future stream of payments that would not have to be made by the Non-defaulting Party. The trial judge concluded that, at least in that case, the method of valuing Loss was “not at all an attempt to use some model to proxy the Market Quotation method of valuing the gain”: at para. 400.

[247] The trial judge also noted that the definitions of Market Quotation and Loss are quite different. In particular, unlike the definition of Market Quotation, the definition of Loss does not prescribe any method to calculate it other than that it must be reasonable. The trial judge observed that for the Market Quotation method to work there must be a liquid market in the particular swap product. Where there is no functioning market, he questioned the logic of attempting to value Loss based on an artificially constructed market.

[248] Second, the trial judge found that Barclays’ model failed to value the CDSs based on all the features laid out in the Agreements, including the collateral triggers, the limited-recourse provision, and the stop-loss feature.

[249] Third, the trial judge reasoned that, without a Market Quotation determination of the Settlement Amount, Barclays would have to establish that the market would be prepared to replace Devonshire in the CDSs on the terms proposed by Barclays as representing its Loss. If Barclays could not establish that, then Loss should be determined on some other basis. The trial judge concluded that Barclays could not establish that the market would replace the CDSs on the terms Barclays proposed for several reasons:

• he accepted Ms. Rahl’s evidence that the market for products such as the Devonshire CDSs was highly illiquid as of January 13, 2009 because of longstanding turmoil in the ABCP market and the 2008 financial crisis;
• he rejected Mr. Draycott’s evidence that there was a market for the Devonshire CDSs in January 2009 and that there were observable market inputs capable of supporting the model;
• taking account of the fact that the Caisse, the largest investor in the Canadian ABCP market, would no longer accept the terms of the original Devonshire CDSs, the trial judge found it unlikely that any other investor would have agreed to those terms; and
• he accepted Ms. Rahl’s opinion that Barclays would not have entered into a replacement transaction on the terms it proposed. While Barclays maintained it would do so provided the collateral triggers were paid on closing, the trial judge was not satisfied any evidence existed to support the likelihood that such a transaction would occur.

[250] In the result, the trial judge concluded that the Barclays valuation was not commercially reasonable. He then turned to Ms. Rahl’s cash flow-based valuation.

[251] The trial judge accepted Ms. Rahl’s evidence that the discounted value of expected losses in the underlying bond portfolios over the life of the CDSs was $12,347. However, he rejected her evidence that a normalized risk premium should be added to this figure to determine Barclays’ Loss. He wrote, at paras. 431-32:

I have some difficulty with this theory. If a reasonable forecast on a cash flow basis of what Barclays has lost by the termination of the swaps is the key, I do not understand why the loss is not the present value of the expected loss of $12,347. Dr. Hull agrees that the alternative cash flow method of valuing an asset is to estimate the cash flow and then discount that cash flow at an appropriate discount rate. He said nothing of adding some risk premium. Ms. Rahl herself said in her report that one could argue that the cash flow projection is the one. She went on to say, however, that to be conservative, she would add a normalized risk premium to the cash flow.

However, the theory of what Barclays’ actual loss is would not lead one to add a risk premium to an expected cash flow loss, based on a mark to model basis, which Ms. Rahl says is not an appropriate way to value in January 2009, let alone a premium of $264 million on $12,247. I do not understand the conceptual basis for doing so. In my view, the loss to Barclays is the present value of $12,347, which I will call $12,000 as it is not known when the expected losses of the underlying portfolio would exceed the attachment point of 16 and 15% on the two swaps.

[252] Concerning mitigation, the trial judge found that Barclays’ Loss should be reduced by the value of any recovery it might receive on Devonshire notes it had purchased from other investors for nominal consideration after the Early Termination Date.

[253] The trial judge agreed with Barclays that, to the extent that the ISDA Master Agreement contemplates payments by a Non-defaulting Party to a Defaulting Party, the common law position that only the party that breaches a contract is liable to pay has been modified by the contract. Nonetheless the trial judge found that the modification does not “necessarily mean that common law contractual principles are entirely abrogated”: at para. 442.

[254] In this regard, the trial judge noted that, in 2008, Barclays purchased approximately $220 million in face value of Devonshire notes for nominal consideration from Citibank, Desjardins and National Bank.

[255] The trial judge held, at para. 462:

[I]t seems to me that if Barclays were entitled to payment from Devonshire collateral on a settlement on a Devonshire default for its Loss, and were able as well to recoup payment on the notes it acquired in Devonshire because Devonshire had more collateral than the settlement amount to be paid to Barclays, it would amount to Barclays obtaining double recovery for its loss. To this extent, if Barclays’ Loss is less than the available collateral plus Devonshire’s cash, there should be a deduction from Barclays’ Loss calculation of the amount it will receive from its Devonshire notes so that the Loss payable to Barclays’ is net of the amount it will receive on its Devonshire notes. [Emphasis added.]

[256] Accordingly, the trial judge reduced Barclays’ Loss to $0, as he was satisfied that Barclays would receive full compensation for its Loss from its recovery on the Devonshire notes it had purchased for nominal consideration.

(4) Analysis

(a) Did the trial judge err in rejecting Barclays’ valuation of its Loss?

[257] Barclays argues that the trial judge made multiple errors in rejecting Barclays’ valuation of its Loss.

[258] First, Barclays submits that the trial judge erred in holding that Loss is not a proxy for Market Quotation. Barclays argues that the trial judge erred in rejecting the interpretation of the English courts that Loss and Market Quotation are broadly intended to achieve the same result: see Anthracite Rated Investments (Jersey) Ltd. v. Lehman Brothers Finance S.A., [2011] EWHC 1822 (Comm.), at para. 116; Australia & New Zealand Banking Group Ltd. v. Société Générale, [2000] C.L.C. 833 (C.A.), at para. 15; Peregrine Fixed Income Ltd.,at para. 30; Britannia Bulk Plc v. Pioneer Navigation Ltd., [2011] EWHC 692 (Comm.), at paras. 44-45; and Pioneer Freight Futures Company Limited v. TMT Asia Limited, [2011] EWHC 778 (Comm.), at para. 98.

[259] For example in Anthracite, at para. 116, after referring to the foregoing authorities, the court stated the following:

Those authorities establish the following broad propositions:

(1) Loss and Market Quotation are, although different formulae, aimed at achieving broadly the same result, so that outcomes derived from one may be usefully tested by way of cross-check by reference to the other: see per Mance LJ in the Australia case at paragraph 2, 15 and 22. This derives from a concession in that case, but has subsequently been reaffirmed after adversarial argument in the Peregrine case at paragraph 30, in the Britannia Bulk case at paragraphs 44 to 46 and 51 and in the Pioneer case at paragraphs 98 and 105. It is one of those sensible concessions which has hardened into hornbook law.

[260] Barclays contends that the Market Quotation payment measure is plainly a replacement cost calculation and that English jurisprudence has correctly determined that the Loss payment measure is also a replacement cost calculation.

[261] Barclays argues that the trial judge erred by failing to accept Barclays’ mark-to-market model as an appropriate method of valuing Loss. This is because mark-to-market models are routinely used to value bespoke CDSs, since bespoke CDSs do not trade regularly and thus must always be valued in the context of an illiquid market. The trial judge also erred in treating the inability to obtain three firm quotations as evidence of illiquidity in the market. Barclays submits that the trial judge’s references in his reasons to specific terms of the ISDA Master Agreement as demonstrating that Market Quotation and Loss are not the same reflect a basic misunderstanding of the concept of Loss as a replacement cost approach.

[262] Second, Barclays argues that the trial judge erred in holding that Barclays’ calculation was flawed because it failed to value the CDSs based on all the features laid out in the Agreements, including the collateral triggers, the limited-recourse provision, and the stop-loss feature.

[263] Barclays contends that the trial judge ignored the requirement in the ISDA Master Agreement and the suggestion in the English case law that Loss should be calculated assuming that the transaction would have proceeded to a conclusion and that the parties would have complied with all payment obligations. The English authorities have referred to this method of valuation as valuing “clean”: see e.g. Anthracite, at para. 116. Expressed another way, Barclays argues that the trial judge erred by valuing the transaction “dirty” rather than clean, as disapproved of in Anthracite, at para. 129, “i.e. by reference to the real world following default, rather than the hypothetical world called for by the authorities on Section 6 of the [ISDA] Master Agreement”.

[264] Finally, Barclays submits that the trial judge erred in holding that, absent a Market Quotation determination, Barclays must show on a balance of probabilities that the market in fact would have been prepared to enter into a replacement transaction on the terms asserted by Barclays in its Loss claim, failing which Loss must be determined on some other basis. Once again, Barclays submits that this conclusion violates the principle that Loss must be valued “clean” rather than “dirty”, that is, by assuming that all necessary conditions for fulfilment of the contract will be satisfied.

[265] In support of its second and third arguments, Barclays relies, for example, on the following statement in Anthracite, at para. 116, which continues the quotation set out earlier in these reasons:

(2) The identification of the non-defaulting party’s loss of bargain arising from the termination of the Derivative Transaction requires a “clean” rather than “dirty” market valuation of the lost transaction. This means that the loss of bargain must be valued on an assumption that, but for termination, the transaction would have proceeded to a conclusion, and that all conditions to its full performance by both sides would have been satisfied, however improbable that assumption may be in the real world: see in the Australia case at paragraphs 5, 22 to 27 and 30-31, the Britannia Bulk case at paragraphs 11 to 14 and 34-35, and in the Pioneer case at paragraphs 112 to 117.

[266] We acknowledge that in some parts of his reasons, the trial judge may have disregarded the apparently well-established practice of using Gaussian copula models to value bespoke CDSs in an illiquid market. We also acknowledge that the trial judge did not take the broad pronouncements of the English authorities concerning valuing Loss “clean” rather than “dirty” into account. Those cases suggest that a Non-defaulting Party’s “loss of bargain must be valued on an assumption that, but for termination, the transaction would have proceeded to a conclusion, and that all conditions to its full performance by both sides would have been satisfied”: Anthracite, at para. 116.

[267] However, even assuming that the trial judge erred in either or both respects, we are not persuaded that he erred in rejecting the Loss valuation produced by Barclays’ model.

[268] As the trial judge observed, the ISDA Master Agreement requires a Non-Defaulting Party’s calculation of Loss to be reasonable and made “in good faith”. It must also produce a commercially reasonable result: see e.g. Firth, at para. 11.140.

[269] In this case, as Ms. Rahl noted in her report, the 17-month Standstill pending the effort at restructuring created a unique and unprecedented challenge to valuing the Non-defaulting Party’s Loss.

[270] According to Ms. Rahl, the ISDA Master Agreement presumes the CDSs would be promptly unwound upon an Event of Default at spreads consistent with the market conditions that precipitated the default. Applying quite different market circumstances 17 months later, in circumstances where liquidity calls, collateral calls and stop-loss features were suspended in the intervening period, departs from the original intent of the ISDA Master Agreement. Particularly given the events that occurred during the intervening 17-month period, including the 2008 financial crisis, the extreme illiquidity in the market, and the temporarily aberrant risk premiums that existed in January 2009, the operation of the Standstill Agreement rendered Barclays’ model-based valuation of Loss commercially unreasonable.

[271] In normal circumstances, the special features of the swaps (including the liquidity feature, the stop-loss feature, and the limited-recourse provision) effectively capped the real value of the CDSs to Barclays on termination at the amount of the posted collateral in place at any particular time. In effect, Barclays’ use of its model-based valuation was an attempt to shift the entire consequences of failing to achieve a restructuring during the Standstill on to Devonshire.

[272] Moreover, as the trial judge found, as of January 2009, there was simply no willing buyer and no willing seller for the CDSs at any price approximating that produced by Barclays’ model. The significant inflation in risk premiums during the Standstill Period had caused the model-based value of the CDSs to increase beyond any realistic estimate of their underlying real worth, taking account of their special features.

[273] None of the English authorities relied on by Barclays address the proper approach to valuation in circumstances involving the Standstill Agreement. Having regard to the unique circumstances created by the Standstill Agreement, in our view, the trial judge did not err in expressing the following conclusions, at paras. 422-23 of his reasons:

In my view, Barclays has not established that the model that it used to value the Devonshire swaps with its conditions as they existed from the time the swaps were agreed in 2006. What a different model might have calculated is of course not before me.

Nor am I satisfied that Barclays has established on a balance of probabilities that its claimed loss of $1.2 billion is a value that the market in fact would have placed on the Devonshire swaps and that its loss calculation is commercially reasonable. While its model indicated that its Devonshire swaps were in the money, and that their value was $1.2 billion, the evidence does not support such a real value. It is an artificial construct. Barclays has not established that the swaps had the replacement value it claims they had at the time it decided to terminate the Devonshire swaps in January, 2009.

[274] Barclays argues that the trial judge dismissed the effect of the Standstill Agreement as being a consideration relevant to valuation when he rejected Ms. Rahl’s opinion that Barclays’ Loss should be valued as of a date other than January 13, 2009. As Devonshire did not appeal this finding, Barclays submits that the effect of the Standstill Agreement is not a relevant consideration on appeal. We do not accept that submission.

[275] In concluding that January 13, 2009 was the contractually required date for valuation, the trial judge stated, at para. 374:

While there is a reasonableness requirement involved in establishing a loss under an ISDA Master Agreement, I do not think that it is permissible on that ground to change the valuation date to a date other than as prescribed. The definition of Loss in the ISDA Master Agreement provides that the party will determine its Loss as of the relevant Early Termination Date or, if that is not reasonably practicable, as of the earliest date thereafter as is reasonably practicable. If the loss calculation is determined to be commercially unreasonable, that may require a different calculation of loss, but it would not permit the valuation date to be changed.

[276] In our view, although the trial judge was satisfied that the valuation date could not be changed because of the Standstill Agreement, he did not conclude that the Standstill Agreement was irrelevant.

[277] Accordingly, we would not give effect to Barclays’ arguments that the trial judge erred in rejecting its calculation of its own Loss.

(b) Did the trial judge err in valuing Barclays’ Loss at $12,000?

[278] Barclays asserts that the trial judge made three main errors in valuing its Loss at $12,000.

[279] First, Barclays relies on its submission that the Market Quotation payment measure is plainly a replacement cost calculation and that English jurisprudence has correctly determined that the Loss payment measure is also a replacement cost calculation. Barclays submits that the trial judge erred in his attempt to distinguish the English authorities and that his calculation of Loss results in a figure that in no way approximates the replacement cost of the CDSs on the Early Termination Date.

[280] Second, Barclays submits that the trial judge demonstrated a fundamental misunderstanding of the loss of bargain component of Loss when he rejected Ms. Rahl’s opinion that a risk premium should be added to her discounted cash flow figure.

[281] Third, Barclays disputes Ms. Rahl’s assertion that the risk premium to be added should be “normalized”. According to Barclays, Ms. Rahl’s opinion that a normalized risk premium should be added to her discounted cash flow figure is nothing more than an attempt to change the valuation date from January 13, 2009 (the Early Termination Date) to some other date.

[282] We agree that the trial judge demonstrated a misunderstanding of the loss of bargain component of Loss when he rejected Ms. Rahl’s opinion that a normalized risk premium should be added to her discounted cash flow figure.

[283] According to Ms. Rahl, her discounted cash flow figure represented the present value of the projected losses that would occur in the two reference portfolios from the Early Termination Date to maturity. However, that figure did not take account of the market implied estimate of projected losses, or “risk premium” that is part of the cost of purchasing credit protection. Put another way, Loss includes the Non-defaulting Party’s loss of bargain for the CDSs. Accordingly, by rejecting the normalized risk premium component of Ms. Rahl’s opinion, the trial judge valued the likely loss to be suffered in the underlying portfolios; he did not value the loss of bargain in relation to the CDSs.

[284] As of January 13, 2009, Barclays was paying Devonshire $3.6 million per year to obtain credit default protection on the two reference portfolios. That was the price agreed upon for credit default protection in 2006 when the CDSs were established — and that was at a time that preceded the events that led to a dramatic increase in the price of credit default protection. Considered in this context, the $12,000 figure arrived at by the trial judge clearly ignores Barclays’ loss of bargain in losing the benefit of the CDSs. It also ignores the evidence of all the experts that a risk premium constitutes a significant component of the price of a CDS.

[285] In our view, Ms. Rahl did not err in adding a normalized risk premium to her discounted cash flow figure. For the reasons we have already explained, Barclays’ use of its model-based valuation as of January 13, 2009 produced a commercially unreasonable result. That said, it was undisputed among the experts at trial that use of a Gaussian copula model to value bespoke CDSs had become standard practice.

[286] As we have said, Ms. Rahl arrived at a normalized risk premium of $264 million by averaging risk premiums in the CDSs markets for August and November 2007, which were calculated using a Gaussian copula model. She also looked at the risk premium in the CDS market in April 2010 (calculated in the same way) and determined it was at the same level as her average for August and November 2007.

[287] In her report, Ms. Rahl provided the following explanation for adding a normalized risk premium to her discounted cash flow figure:

An estimate of loss that starts with projected real-world cash flows but that is nonetheless consistent with market practice in normal times should also, in our opinion, include a “normalized” risk premium. The events of late 2008 and early 2009 resulted in extraordinarily high and temporary risk premiums, as we have seen previously in this report. Adding losses projected from cash-flows to a “normalized” risk premium results in a loss estimate that is consistent with a mark-to-model or mark-to-market approach in normal times, but that also discounts the extraordinarily and temporarily high risk premiums of the financial crisis. Projected real-world losses before and after the crisis are typically very small or deminimus for AAA-rated investments. The swap value in question in normal times is therefore likely comprised entirely of “normalized” risk premium as at all times the actual and expected real-world loss projections are deminimus. Adding such a “normalized” swap value to the real-world losses that are projected during the midst of a crisis results in a loss estimate that is not inflated by temporary and aberrant illiquidity and risk premiums.

[288] In all the circumstances, we agree that this reflects a reasonable approach to valuation. Moreover, in our view, had the trial judge properly understood the need to recognize the loss of bargain component relating to the credit protection contract aspect of the CDSs, he would undoubtedly have accepted Ms. Rahl’s evidence in this regard.

[289] Accordingly, we accept Barclays’ submission that the trial judge erred in calculating Barclays’ Loss as $12,000 and we substitute a figure of $264 million.

(c) Did the trial judge err in his treatment of mitigation?

[290] As we have said, the trial judge held that Barclays’ Loss should be reduced by the value of any recovery it receives on the $220 million face value of Devonshire notes it purchased from other investors in 2008 for nominal consideration.

[291] Barclays contends that, in reaching this conclusion, the trial judge made multiple errors. First, he erred in concluding that mitigation principles apply to a Loss calculation. Second, he erred in failing to recognize that Barclays’ entitlements under the CDSs and as a noteholder are separate and distinct. Third, he erred in finding that steps taken by Barclays prior to the termination date should be recognized as mitigation. Fourth, he erred in failing to recognize that, if Barclays’ recovery on the CDSs is to be reduced by the amount of its recovery on its notes, more collateral will be available for distribution to the noteholders (including Barclays), which will further reduce Barclays’ recovery on the CDSs, which will again make more collateral available for distribution, and lead to an illogical ad infinitum calculation.

[292] We do not accept these submissions.

[293] On our review of his reasons, the trial judge carefully considered the definition of Loss and other relevant contractual provisions together with the relevant case law and commentary dealing with the question of the applicability of principles of mitigation. While he agreed with Barclays that common law principles have been modified to the extent that the ISDA Master Agreement contemplates payments by a Non-defaulting Party to a Defaulting Party, he concluded that that “does not necessarily mean that common law contractual principles are entirely abrogated”: at para. 442.

[294] We agree with that conclusion. Moreover, like the trial judge, we consider that certain aspects of the Agreements point to a duty to mitigate, or at least a duty to take reasonable steps to avoid losses and costs:

• Loss is defined to include “an amount that a party reasonably determines in good faith to be its total losses and costs … including any loss of bargain” (emphasis added) – as the trial judge noted, this language may import common law principles because compensating a contracting party for the loss of bargain on the termination of contract is precisely what the common law measure of damages is meant to achieve; and
• the party determining Loss is required to act reasonably and in good faith – this is consistent with a contracting party’s duty to mitigate, and, as Firth points out, at para. 11.148, likely requires the determining party to reflect costs or losses that have been avoided in the Loss calculation.

[295] We see no error in the trial judge’s conclusion that the parties did not expressly contract out of the necessity of mitigation. Further, we are not satisfied that the case law relied on by Barclays does anything more than make it clear that the ISDA Master Agreement departs from common law principles to the extent that it contemplates payments by a Non-defaulting Party to a Defaulting Party.

[296] As for Barclays’ arguments that the trial judge failed to recognize that its entitlement under the CDSs and as a noteholder are separate and distinct and that it purchased the notes prior to the Early Termination Date, we agree with the trial judge’s assessment. He concluded that “if Barclays were entitled to payment from Devonshire collateral … on a Devonshire default for its Loss, and were able as well to recoup payment on the notes it acquired in Devonshire because Devonshire had more collateral than the Settlement Amount to be paid to Barclays, it would amount to Barclays obtaining double recovery for its Loss”: at para. 462.

[297] In this regard, we note the trial judge’s observation that Mr. Lee of Barclays described the purchase of these notes in different ways, including as “buying back the risk”, being “a hedge for our original $6 billion trade” and being “equivalent to a sell of protection”: at para. 460.

[298] Although Barclays maintains that its hedging strategies have nothing to do with the replacement value of the CDSs and that its distinct capacities negate double recovery, the circumstances of this case raise the irresistible inference that Barclays was able to purchase the notes for nominal consideration because the notes were perceived to be less valuable as a result of Barclays’ accumulating Loss. In these circumstances, neither the timing of the purchase, nor the difference in capacities, alters the fact that Barclays’ Loss will be offset by its recovery on the notes.

[299] We see no merit in Barclays’ argument that potential difficulties in calculating the amount to be deducted on account of mitigation should somehow affect the issue of mitigation. If the parties are unable to agree on an appropriate amount to be deducted from Barclays’ Loss of $264 million on account of mitigation, and if the issue becomes relevant, we refer it to the trial judge or the judge conducting the second phase of the bifurcated trial, if it takes place, for determination.

D) Issue 4: Did the trial judge err in his determination of the issues of priorities, subordination and non-recourse?

(1) Priorities and Subordination

[300] Paragraph 1(c) of the formal judgment provides that any Settlement Amount payable to Barclays from Devonshire upon termination of the ISDA Master Agreement is subordinated to the prior payment of certain amounts, including amounts payable to the Devonshire noteholders for principal and interest.

[301] Paragraph 1(c) is premised on the trial judge’s findings that: Barclays’ Notice of Early Termination was invalid; Devonshire’s Notice of Early Termination was valid; and, in these circumstances, under the terms of s. 2.2(c) of the Intercreditor Agreement, any Settlement Amount payable to Barclays is subordinated to the prior payment of amounts owing to the Devonshire noteholders.

[302] The relevant portion of s. 2.2(c) of the Amended and Restated Intercreditor Agreement reads as follows:

The parties hereto acknowledge and agree as follows:

(c) so long as any Series A Indenture Obligations are outstanding, if an Early Termination Date is designated under the Devonshire Financial Contract by reason of an Event of Default with respect to the Bank … any Settlement Amount payable to the Bank pursuant to Section 6(e) of the Devonshire Financial Contract … shall be subordinated to the prior payment of the amounts specified in paragraphs (a), (b), (c), (d) [which includes the Devonshire noteholders] and (e) of Section 3.1 of the Series a Supplemental Indenture….

[303] Barclays does not dispute that, if its Notice of Early Termination is invalid, and, if Devonshire designated a valid Early Termination Date under a valid Notice of Early Termination, s. 2.2 of the Intercreditor Agreement subordinates any Settlement Amount owing to it to the prior payment of the Devonshire noteholders. We agree.

[304] As we have reached the same conclusions as the trial judge concerning the validity of the parties’ respective Notices of Early Termination (albeit for somewhat different reasons), it is unnecessary to address the additional arguments raised by the parties concerning priorities and subordination in the event we had reached conclusions that differed from the trial judge’s conclusions.

(2) Limited-recourse

[305] At trial, Barclays argued that if any Settlement Amount due to it from Devonshire exceeded the initial payments of $600 million made by Devonshire, under the terms of the parties’ Agreements, it was entitled to be paid any such excess from the “residual assets” of Devonshire. Although it was not strictly necessary that he address it, the trial judge rejected this argument.

[306] Barclays renewed this argument on appeal. In the light of our conclusion concerning the Settlement Amount, it is unnecessary that we address it.

VI. CONCLUSION

[307] In conclusion, we dismiss the appeal for the following reasons:

• Barclays’ Notice of Early Termination with respect to Devonshire’s insolvency was invalid because of Barclays’ misrepresentation and bad faith, and because Barclays cannot be permitted to benefit from its own wrong.
• Devonshire’s Notice of Early Termination with respect to Barclays’ failure to make the liquidity payments was valid.
• In any event, even if Devonshire’s Notice of Early Termination was invalid, Barclays repudiated the Agreements and Devonshire accepted the repudiation.
• The trial judge erred in his determination of Barclays’ Settlement Amount and we substitute a figure of $264 million for Barclays’ Loss. If the parties are unable to agree on the amount that should be deducted from the $264 million on account of mitigation, we refer this issue back to the trial judge or to the judge conducting the second phase of the bifurcated trial, if it takes place, for determination.
• Given our conclusions on the first three questions, it is not necessary for us to decide the issues of priorities, subordination, or whether Barclays had limited recourse to Devonshire’s assets.

VII. COSTS

[308] If the parties are unable to agree as to the costs of the appeal, they may make brief submissions in writing of not more than 10 pages within 45 days of the release of this judgment.

Released: July 26, 2013 (“S.T.G.”)

“S.T. Goudge J.A.”

“Robert J. Sharpe J.A.”

“Janet Simmons J.A.”