Royal Courts of Justice
Strand, London, WC2A 2LL
Before :
MR JUSTICE HAMBLEN
Between :
Cassa di Risparmio della Repubblica di San Marino SpA | Claimant |
- and - | |
Barclays Bank Ltd | Defendant |
George Leggatt QC and Andrew Henshaw (instructed by Nabarro) for the Claimant
Andrew Baker QC and Thomas Raphael (instructed by Linklaters) for the Defendant
Hearing dates:
22, 29 November, 1-3, 6-10, 13-17, 20-21 December 2010, 8-9 February 2011
Judgment
Mr Justice Hamblen:
I. Introduction and Overview
This is a case about the alleged misselling of a complex financial product. It concerns the sale to the Claimant (“CRSM”) by the Defendant (“Barclays”) in 2004/early 2005 of four sets of structured notes (“the Notes”) with a total nominal value of €406 million, and the subsequent restructuring of the Notes in June 2005. CRSM’s case was that it was induced to buy the Notes from Barclays, and to agree to the subsequent restructuring, by misrepresentations which in each case were fraudulent, or which in any event Barclays did not believe or have reasonable grounds to believe were true. CRSM claimed damages for deceit, or alternatively under s.2(1) of the Misrepresentation Act (“the Act”), alternatively for breach of an implied term of the contract, amounting on CRSM’s most recent calculations to some €92 million.
The Notes constructed by Barclays and sold to CRSM had embedded within them credit derivatives known as collateralised debt obligations (or CDOs) which gave exposure to the credit risk of a pool of “reference assets” through a portfolio credit default swap. In this case the reference assets themselves included further CDOs, each of which was in turn referenced to another pool (or “portfolio”) of some 50 credit default swaps – hence the description “CDO squared” or “CDO2”..
The CDO2s had terms of 5 to 7½ years. At the start of this period CRSM paid the principal amount of each Note to Barclays, on which CRSM then received a coupon (generally Euribor + 0.95%). It was CRSM’s intention to hold the Notes until maturity. In these circumstances the key risk for CRSM was the “credit risk” or risk of default of the CDO2s – that is, the risk that CRSM would cease to receive the full coupon and would not get back the full principal amount when the Notes matured. That would occur if a sufficient number and combination of “credit events” (e.g. insolvency or default on a debt) occurred in relation to entities named in the portfolios underlying the CDO2s.
In agreeing to purchase the Notes, CRSM contended that it relied on the credit rating given to the CSOcomponents of the Notes and on representations allegedly made by Barclays as to the credit risk of these instruments.
It was CRSM’s case that Barclays sold the Notes on the basis of an agreed “AAA” rating, which they intended CRSM to rely on and knew that it did rely on as indicating a minimal level of risk, when Barclays also knew and intended that the instruments which they had constructed were in fact far riskier than their credit rating indicated. CRSM contended that whereas a ’AAA’ rating signifies a default risk over a 5 year period which is close to zero, Barclays’ internal Expected Loss (“EL”) financial modelling indicated that at their dates of issue the CDO2s had a probability of default over their lives of around 30% (equivalent to single ’B’ or ’junk’). The result was that the value of the investments at inception was allegedly some €50m less than CRSM paid for them.
CRSM’s case was that this “alchemy” was achieved by the use of a practice which it described as ’credit ratings arbitrage’. It submitted that the intention and result of this practice was (a) to obscure the actual risk of the instrument and (b) to create an instrument whose actual risk of default was far higher than its AAA rating implied.
CRSM contended that the advantage to Barclays of exposing it to much greater credit risk than CRSM appreciated and the AAA rating implied of the CDO2s was that it enabled Barclays to book large profits from the transactions – which CRSM estimated (although Barclays denied) were almost equivalent to CRSM’s €50m loss on the CDO2s at the point of sale. This was because the levels of profit for Barclays closely corresponded with what CRSM contended was the levels of risk to which it was exposed by the Notes. CRSM’s case is that Barclays constructed and sold to CRSM financial products for which, on their own calculations, a coupon commensurate with the risk would have been around 5% above Euribor, while paying CRSM only the far lower coupon appropriate to a AAA-rated investment, and profited from the difference, and that this was achieved as a result of the misrepresentations relied upon.
There is a second stage to the relevant events. In early 2005 CRSM started to become concerned that some of the companies whose names had been included in the portfolios underlying the CDO2s were experiencing financial difficulties and that the credit risk of the instruments might therefore be higher than CRSM had previously thought. Barclays responded to CRSM’s concerns by recommending to CRSM a restructuring of the CDO2s. This took the form of substitutions of some of the entities referenced in the portfolios and other structural changes to the CDO2s, which Barclays allegedly represented would be achieved without profit to Barclays and by reference to stated criteria which conveyed the impression that the restructuring would be beneficial to CRSM in risk terms.
It was CRSM’s case that, in fact, the effect of the restructuring was, and was known and intended by Barclays at the time to be, the opposite of what was represented, and that it significantly increased the risk to which CRSM was exposed – thereby reducing the value of CRSM’s investments - and enabling Barclays to extract further profits.
Of the three CDO2s embedded in the Notes, Barclays subsequently agreed to repurchase two from CRSM. The third CDO2 has now lost its entire value.
Barclays denied liability on all of the claims and strongly rejected any suggestion of fraud or dishonest dealing on its part. It contended that all the claims are unfounded on the facts, for multiple reasons. It further submitted that the claims (other than for fraud) are defeated, as a matter of contract, by the terms of the purchase and restructuring transactions agreed between the parties; that the claims are defeated, because the extensive pre-contract disclaimers remove any factual foundation for the claims; and that there is no basis for the alleged implied term, not least because it is inconsistent with the terms and structure of the contracts.
Barclays submitted that CRSM’s “arbitrage” case is built on comparing the incomparable. An AAA credit rating is an expert opinion from a ratings agency that a debt instrument is of highest credit quality involving the lowest expectation of credit default. It is an expert opinion as to the actual prospects of a debt instrument as issued performing in full over time, that is highly relevant for all investors and sufficient for many (for making investment decisions). By contrast internal EL modelling such as that carried out by Barclays is not concerned with risk but rather with deriving a price or value based on credit default swap market spreads to be used by banks to mark their books to market, to assist them in hedging and to calculate notional profit at the time of trading.
Barclays further submitted that it is inherent in putting together a CSO2 structure, that the inner CSO portfolios will be populated by reference names with high credit spreads relative to the coupon being paid on any AAA-rated CSO2 tranche. Repackaging portfolios of higher-yielding, lower-rated credits into a lower-yielding, higher-rated structure is the whole purpose of the exercise. Further, the practice of favouring higher-spread names in populating the inner CSO reference portfolios, so long as the CSO2 tranche still attains any target rating(s), was practised not just by Barclays but by other structurers in the field.It was the basis of the CDO business.
II. The Factual History
The Parties
CRSM is the oldest of the twelve banks established in the Republic of San Marino. It was founded in 1882 and has headquarters and 16 branches in the Republic of San Marino. CRSM engages primarily in private banking, retail banking, sales and trading of securities, and asset management. As of 31 December 2007, it had total assets of €3.7 billion and net income of €7.2 million.
CRSM directly or indirectly controls 14 subsidiaries. At the time the Notes were purchased it had a 70% shareholding in Delta S.p.A., which is the holding company for two Italian consumer finance companies, Carifin S.p.A. (“Carifin”) and PlusValore S.p.A. (“PlusValore”) (together referred to as “the Delta companies”). Delta itself and the Delta companies are based in Bologna, Italy.
The main people involved in the transactions on CRSM’s side were Mr Sandro Sapignoli, Treasurer of CRSM (who left CRSM in September 2004); Mr Wilmo Montanari, CFO and Head of the Uffico Finanziaro(finance department) of CRSM until October 2004 (who left CRSM in October 2004); Mr Maurizio Morolli, Consultant in the Uffico Finanziaroof CRSM from June 1999 to March 2005 and Senior Executive in the Uffico Finanziarofrom March 2005 to date, Mr Guiseppe Buoncompagni, Member of the Uffico Finanziarofrom March 2005 to December 2005 (who left CRSM February 2009) and Mr Fantini, the CEO of CRSM at the material time. I heard evidence from all of these people save for Mr Sapignoli and Mr Fantini. I also heard expert evidence from Mr Nasr, a specialist in fixed income and derivative instruments.
Barclays is a UK-based financial services group with a large international presence in Europe, the USA, Africa and Asia. It is engaged primarily in banking, investment banking and investment management. Barclays is a member of the London Stock Exchange and regulated by the Financial Services Authority. In terms of market capitalisation, Barclays is one of the largest financial services companies in the world. For the year ended 31 December 2007, the Barclays group reported pre-tax profits of £7.1 billion.
The transactions at issue in this case were undertaken by Barclays Capital, a trading name for the investment banking division of Barclays. Barclays Capital has offices in 26 countries, and provides financing and risk management services to large corporate, government and institutional clients.
The main people involved in the transaction on Barclays side were Mr Matteo Ferrario, Director August 2003-2006; Head of Italian Distribution 2006 to date; Mr Antonio Agresta, the Structurer of the Notes; Mr Antoine Bechet, Member of the Credit Derivatives department, and Mr Thibault Scaramanga, Member of the Research team 2004-5. There would also have been input from both the sales and compliance departments into the documentary presentations made and input from the legal department into the legal documentation. I heard evidence from all these individuals except Mr Bechet. I also heard expert evidence from Dr Ellis, an expert in securities and derivatives markets, and Ms Duhon, a market professional with 16 years of experience in the financial derivatives market.
Characteristics of the Notes
The series of Notes sold by Barclays to CRSM is summarised in the following table:
Note | Trade date | Settlement Date | Principal | CDO² | PlusValore CDS | Carifin CDS |
Carnivale | 8.7.04 | 28.7.04 | €200m | €100m Umbria II | €63.4m | €36.6m |
Cernobbio | 28.10.04 | 22.11.04 | €100m | €60m Como | €22m | €18m |
Bellagio | 21.12.04 | 20.1.05 | €54m | €40m Como II | €11m | €3m |
Tremezzo | 17.1.05 | 21.2.05 | €52m | €30m Como II | €13m | €9m |
Menaggio | 25.2.05 | 8.4.05 | €44m | €20m Apollo (CCO) | €13m | €9m |
TOTAL | €450m | €250m | €118.4m | €81.6m |
Leaving aside the Menaggio Note, which did not include a CDO2 component and does not form part of the claim, the Notes incorporated:
Credit default swaps referencing each of the two Delta companies (Carifin and PlusValore); and
CDO²s referencing corporate debt and asset backed securities issued by companies unconnected with CRSM.
Credit Default Swaps
A credit default swap (or “CDS”) is a contract under which one party (“the protection seller”) agrees to provide protection to another party (“the protection buyer”) against the risk of a default or other “credit event” suffered in connection with a bond or other obligation (the “reference asset”). The reference asset has generally been issued by a third party (“the reference entity”) which is usually not even aware of the existence of the CDS.
In the simplest form of CDS, the protection buyer makes fixed periodic payments to the protection seller (usually called the “CDS spread” or “CDS premium”) on the notional amount of the CDS throughout the life of the CDS. The CDS spread is a reflection of the perceived credit risks assumed by the protection seller under the CDS.
CDS spread is a particular instance of credit spread (or, simply, “spread”) which represents the additional yield that an investor can earn from an investment with more credit risk. It is typically measured as the difference between the yield on the relevant investment and the so-called “risk free” interest rate, i.e. the yield that can be obtained from investing in a financial instrument which is treated as having a risk of default which is as close to zero as it is possible to get, such as the interest rate on US Treasury Bills or the Euribor rate.
The contractual terms of a CDS are usually set out in a “confirmation”, which incorporates by reference one or more sets of standard terms such as those published by ISDA (the International Swap and Derivatives Association). However, the terms can be, and in the present case were, customised to a significant degree.
The CDS confirmation will list the "credit events" (various categories of adverse development affecting the reference asset) which will trigger an obligation on the protection seller under the CDS. Typically credit events include (i) a bankruptcy filing (or similar insolvency process) by the reference entity, (ii) the reference entity’s failure to make a payment of interest or principal on a debt obligation, and sometimes (iii) a restructuring of existing debt resulting from the reference entity’s distressed condition.
The occurrence of a credit event gives the protection buyer the right to "trigger" its remedies under the CDS contract, which can involve either "physical settlement" or "cash settlement". Under physical settlement, the protection buyer delivers a loan or bond of the reference entity to the protection seller, whose notional amount is equal to the notional amount of the CDS; the protection seller, in return, pays the protection buyer the face value for the loan or bond. Under cash settlement, the current market value of a representative debt instrument of the reference entity is determined as a percentage of the par value of the instrument, typically by polling leading dealers of debt instruments; this is referred to as the “recovery rate”. The protection seller then pays the protection buyer an amount in cash equal to the notional amount of the CDS minus the notional amount of the CDS multiplied by the recovery rate.
CDSs have been widely used by banks and other entities to manage their exposures to corporate, bank and sovereign entities, as well as structured finance and project finance exposures; they are used also to achieve a reduction in applicable capital requirements for bank regulatory purposes.
The CDS outlined above is “unfunded”, since the protection seller makes no upfront payment to the protection buyer; a payment from the protection seller becomes due only if a credit event occurs. This means that the protection buyer is exposed to the credit risk of the protection seller, and will find the benefits of his bargain eroded if the protection seller is itself unable to meet its obligations at the time of settlement.
To eliminate the risk of non-performance by the protection seller, the CDS contract may require the protection seller to make a payment to the protection buyer at the outset equal to the notional amount of the CDS, on which interest accrues until either this cash is returned to the protection seller on the termination date, or is used following a credit event to satisfy the protection seller’s obligation under the physical or cash settlement alternative. This type of CDS arrangement is known as a “funded” CDS.
A credit linked note ("CLN") combines the characteristics of a funded CDS into a note, with the protection buyer simply issuing to the protection seller a note whose payment terms are expressly linked to the reference asset. The protection seller (investor) pays to the issuer / protection buyer at inception the principal amount of the note. During the life of the CLN, the issuer makes periodic payments (coupons) to the investor. The note is repaid in full at maturity so long as no credit event occurs concerning the reference entity. If a credit event occurs, however, the issuer retains the amount paid at inception by the investor, in exchange for delivery of the reference asset (if physical settlement applies); if cash settlement applies, the issuer will repay to the investor an amount calculated using the recovery rate (explained as above) and will keep the difference between this amount and the sum paid at inception by the investor.
CDOs and CDO2s
A “CDO” (collateralised debt obligation) is a security issued by a company (usually set up specially for this purpose) which either owns or assumes obligations in relation to a pool of financial instruments, generally loans or bonds (the “reference portfolio”). The reference portfolio of a CDO may contain dozens or hundreds (or even thousands) of instruments.
The CDO issuer issues securities to investors representing different tranches of risk that are transferred to the investors.
By way of example, tranches might consist of:
a tranche representing the first 5% of portfolio losses (often called the "First Loss Piece" or the "equity tranche");
a tranche representing the next 10% of portfolio losses (often called the "mezzanine tranche"), and
a tranche representing the remaining 85% of portfolio losses (often called the "senior tranche").
There may be a greater or smaller number of tranches, and the thickness of each tranche varies from deal to deal.
All tranches other than the equity tranche are said to have the benefit of "subordination", since they do not absorb losses until one or more other tranches which are subordinate to them have been exhausted first. The equity tranche is subordinate to the mezzanine tranche, with the mezzanine tranche (in the example given above) said to have "5% subordination"; and the equity and mezzanine tranches are subordinate to the senior tranche, which is said to have "15% subordination".
If a loss occurs in relation to one of the assets in the reference portfolio, it is borne first by the equity tranche, which is written down by the amount of this loss. Subsequent losses are cumulated with previous losses, eroding the equity tranche until it is written down to zero. If and when this happens, further losses are borne by the next tranche, and then by the one above that, and so on, with each tranche written down by the amount of losses it incurs. For the mezzanine tranche in the example above, it is common to speak of 5% as the “attachment point” and the 15% as the “detachment point”: i.e. this tranche begins absorbing losses once they reach 5% cumulatively (the attachment point), and is wiped out completely when losses reach 15% cumulatively (the detachment point).
The coupons paid to the investors on the various tranches differ and reflect the perceived probability of loss on the relevant tranche. The equity tranche assumes the greatest amount of risk and thus earns the highest spread – often hundreds or thousands of basis points annually (a "basis point" is 1/100 of 1% per annum) – while the most senior tranche (the 15% - 100% tranche in the example) assumes the least risk (despite its large size) and typically pays a more modest coupon.
CDOs are a form of structured credit instrument, the word "structured" referring principally to the tranching of risk via the subordination of investors to one another; this is in contrast to more traditional forms of pooled investment vehicle, such as mutual funds and investment trusts, in which investors typically share risks and returns on a pari-passu basis. Tranching introduces an element which makes pricing and risk assessment more difficult.
A "CDO squared" (or "CDO²") is a CDO which includes one or more tranches of one or more other CDOs in its reference portfolio. With such an instrument, the instrument itself is sometimes described as the "outer CDO", and the CDO tranches included in its reference portfolio are called the "inner CDOs". CDO tranches are less transparent than sovereign, bank or corporate debt, which makes it harder to obtain price information on them.
Many instruments referred to as “CDO squared”, including the ones with which this case is concerned, in fact have a mixed or hybrid portfolio of reference assets, some of which consist of inner CDO tranches and the rest of which consist of traditional structured finance instruments, here referred to as "ABS" (meaning asset-backed securities). These generally consist of securities backed by loans (e.g. mortgage loans, credit card loans, auto loans, or other consumer loans).
The term "cash CDO" is used to describe a CDO whose issuer actually owns the underlying portfolio of assets, either directly or via participation interests. Tranches of a cash CDO for which no investor can be found at a reasonable price are often retained by the arranger. The term "synthetic CDO" is used to describe a CDO whose issuer does not actually own the reference assets; instead, it is exposed to the risk of credit events on the reference assets via one or more credit default swaps and then buys protection against the risk of these events from investors, generally using the tranching approach described earlier. In this case it is possible to structure the transaction such that only one tranche exists. This allows the issuer to create a structure specifically tailored to the investor.
Each of the inner CDOs and the outer CDOs underlying each of the Notes sold to CRSM was synthetic and it would appear that no other tranches were sold to or placed directly with any third party.
A "synthetic CDO" may also be referred to as a “CSO” (Collateralised Synthetic Obligation). All the CDOs with which this case is concerned were synthetic, and they are described sometimes as “CDOs” and sometimes as “CSOs”, these terms being for present purposes effectively interchangeable.
Assessment of the risks and value of CDOs, and even more so of CDO²s, is difficult. Default risk at the outer CDO level depends, amongst other factors, on the levels of subordination of both the outer CDO and the inner CDOs included in its reference pool. It also depends on the probability and correlation of defaults within and between the reference assets of the inner CDOs. Estimating these correlations can be even more difficult than for standard CDOs. Further, since there is a limited universe of actively traded credit default swaps, most CDO²s have a structure where two or more of the inner CDOs reference a number of the same entities. The degree of this “overlap” influences the risk profile of a CDO². In particular, overlap in the inner CDOs, especially if it occurs on riskier reference entities, generally increases the expected losses of the outer CDO.
Credit Ratings
Credit ratings agencies (“CRAs”) are private organisations that specialise in the assessment of the credit risk of debt securities. The three largest CRAs are Standard & Poor’s, Moody’s and Fitch. They analyse quantitative and qualitative aspects of a borrower’s credit condition to determine the likelihood that the borrower will be able to service its debt obligations in accordance with their terms, and assign a credit rating on that basis. Most credit ratings are specific to a particular security rather than to the borrower as a whole, and a single borrower may have different securities rated differently depending on the level of legal seniority of the particular security and the availability and quality of collateral or guarantees supporting it. Generally, CRAs take a view on the entity’s ability to meet both interest and principal payments, although more limited credit ratings are sometimes requested.
The table below sets out the rating hierarchy for the three main CRAs, with AAA or Aaa being the highest possible rating and indicating the lowest risk of default, and C (or Moody’s Ca) being the lowest possible rating for a performing security. The distinction between investment grade and non-investment grade is not a legal one, but has traditionally been used to distinguish securities with negligible default risk from those with a material risk of default. Many institutional investors limit their investments to investment grade securities, or put strict caps on the amount of non-investment grade securities they are willing to purchase.
S&P | Fitch | Moodys | |
"Investment Grade" | AAA | AAA | Aaa |
AA+ | AA+ | Aa1 | |
AA | AA | Aa2 | |
AA- | AA- | Aa3 | |
A+ | A+ | A1 | |
A | A | A2 | |
A- | A- | A3 | |
BBB+ | BBB+ | Baa1 | |
BBB | BBB | Baa2 | |
BBB- | BBB- | Baa3 | |
"Non-investment grade" | BB+ | BB+ | Ba1 |
BB | BB | Ba2 | |
BB- | BB- | Ba3 | |
B+ | B+ | B1 | |
B | B | B2 | |
B- | B- | B3 | |
CCC+ | CCC | Caa1 | |
CCC | Caa2 | ||
CCC- | Caa3 | ||
CC | CC | Ca | |
C | C | ||
D | D | C |
Fees for credit ratings are normally paid by the issuer, but in the case of structured finance securities for which no issuer (in the traditional meaning) exists, fees are normally paid by the structurer/arranger of the transaction, in this case Barclays.
Sale of the Notes
The Carnivale Note
The genesis of the sale of the Notes was an approach made to Barclays by CRSM and Delta for financing of about €500 million for the Delta companies’ consumer lending business. CRSM was unable to provide further direct funding for that business because of risk concentration limits prescribed by the Central Bank of San Marino.
At meetings on 11-12 December 2003, CRSM asked Barclays to help with asset-backed secured lending to the Delta companies. It was envisaged that this would involve a €400+ million securitisation of the Delta companies’ consumer loans to be arranged by Barclays in 2004. As an initial measure, CRSM wished Barclays to make a 5-year €50 million loan to PlusValore, which Barclays would repackage into a Credit Linked Note and sell to CRSM to obtain protection against the risk of default.
In February 2004, Barclays did a separate trade with CRSM under which Barclays lent €12m to two clients of CRSM’s, Rafal and Falco, and CRSM purchased from an SPV a CLN, structured by Barclays, called Raffaello, which referenced two CDSs referenced to the Rafal/Falco loans plus a AAA-rated CSO with a nominal value of €8m. The effect of the CDS contained in the CLN was to give Barclays 100% credit protection over the Rafal/Falco loans, and this was combined with CRSM’s purchase of the CSO asset to create a single asset purchased by CRSM, paying coupon to CRSM. As a preliminary to the transaction, Barclays required CRSM to sign (on every page) a purchase letter containing a number of disclaimers and protective provisions. These were similar to those relied upon by Barclays in relation to the transactions in this case.
A securitisation remained CRSM’s preferred route for the lending as a whole and the provisional plan was for a revolving securitisation to be entered during the second half of 2004. The fallback plan was for Barclays’ lending to be supported by CLNs bought by CRSM.
Barclays concluded fairly quickly that a securitisation was not feasible, at least in the short term, and the fallback plan was adopted. Barclays proposed to lend to the Delta companies taking advantage of the new Article 2447-decies of the Italian Civil Code, which permitted a company to borrow money to finance a particular transaction or business undertaking (in this case, the Delta companies’ lending to consumers) and to repay the loan exclusively with the revenues generated by that transaction (in this case the consumer loans) which cannot be used to pay other creditors. Under Article 2447-decies a lender is protected from the claims of other creditors, but its recourse is limited to the segregated assets. The Delta Loans, as in due course made, were 20% over-collateralised; in other words, the security that Barclays held in the form of the segregated assets (consumer loans) was 20% greater in amount than the principal amount of the loans.
It was also agreed at a relatively early stage that the CLNs that CRSM would purchase would contain CDSs only for 50% of the loan value (thus giving Barclays only 50% credit protection over the loans), but would also contain an AAA note (that CRSM would invest in, by purchasing the CLN) with nominal value equal to the 50% of the Delta credit risk that Barclays would take. It was Barclays’ evidence that an important part of the deal, from CRSM’s perspective, was that the resultant CLNs would pay a “good”, i.e. relatively high, coupon.
On 24 May 2004 Mr Ferrario sent to CRSM a summary “Descrizione della transazione”, describing a proposed note to be issued by Arlo II Limited linked to three reference assets:
a CDS referencing PlusValore with notional amount €50 million;
a CDS referencing Carifin with notional amount €50 million; and
a CSO, at this stage called “Grafton 2004”, of notional amount €100 million referencing a portfolio of ABS securities with AAA rating and CSOs with A rating; thus the CSO was to be a CDO².
The summary described the proposed CSO as having a rating of “[AAA]”.
This was followed on 28 May 2004 by a fact sheet dated February 2004 for a CDO named “Atlas” referring to tranches rated AAA, AA and A. The fact sheet concluded with various qualifications and disclaimers.
A further version of this fact sheet, albeit still bearing the date February 2004, was sent to CRSM on 1 June 2004, describing the rating as AAA. This version included an “Analysis of Various Scenarios” beginning:
“Atlas is a structure exposed to a highly diversified portfolio which is able to tolerate various defaults before they are able to damage to and erode the capital or coupon.
All the assets within the Atlas portfolio have a level of subordination that guarantees protection against a significant number of defaults.”
On 7 June 2004 Mr Ferrario sent CRSM another version of the Atlas fact sheet containing similar statements to the 1 June version, and also a “Zebra” fact sheet. This was dated June 2004 and was similar to the Atlas fact sheet. It again described the rating of the proposed CDO as AAA. The coupon was stated to be 3 month Euribor + 1.0%. The “Analysis of Various Scenarios” began with equivalent statements to those in the Atlas fact sheet quoted above. It also included the following further statements set out in bold type which were referenced to a graph:
“The AAA tranche of Zebra is particularly robust with respect to defaults. It is possible to compare the default rate of Zebra’s AAA tranche in the different scenarios against the 5 year historical default rates of a portfolio with an average rating corresponding to that of the corporate exposures of the CSO (Baa). Assuming in the worst case that only corporate exposures are subject to default, the Zebra AAA tranche can withstand a default rate significantly higher than the worst default rate that has occurred in any 5 year period between 1940 to 1991 (the five years between 1986-1991). In scenario 1 “Worst Case”, the AAA tranche can withstand a default rate of 11.1% per CSO. This cushion is double the worst historical default rate that occurred during the five years 1986-1991. In scenario 2 “Intermediate Case”, the AAA tranche can withstand a default rate of 27.7% per CSO, 5 times higher than the worst historical default rate.”
On 2 July 2004 Mr Ferrario sent an email to Mr Montanari of CRSM stating that “[f]ollowing our discussion, we understand that [CRSM] is giving Barclays a mandate to structure and execute the transaction described in the main points of the attached fact sheets”.
The first attachment to Mr Ferrario’s email was a Transaction Outline which described the Note as repackaging (i) a 3-year €63.65m credit default swap referencing PlusValore, (ii) a 3-year €36.35m credit default swap referencing Carifin, and (iii) a 5-year €100m CSO Note. The CSO Note was described as rated “[AAA]” and as paying 6 month Euribor + 100bps.
The second attachment to Mr Ferrario’s email of 2 July 2004 was a fact sheet dated July 2004 and now entitled “Umbria II: ABS CDO”. This included the same statements as are quoted in respect of “Zebra” above except that references to “Zebra” had now been changed to “Umbria II”.
The fact sheet contained certain “Qualifications and Disclaimers” upon which Barclays place reliance. On the top of each page of the fact sheet, it stated “Preliminary Information Only”.
Mr Ferrario requested that both the Transaction Outline and the fact sheet be signed by CRSM on every page, which was done by Mr Montanari on 2 and 5 July 2004 respectively.
On 6 July 2004 Mr Ferrario requested internally an updated P&L split, and chased for it on 8 July saying “Guys I need to see the economics in writing befor[e] I trade”. These figures were provided later on 8 July 2004 and indicated gross profit of 23.8% and net profit of 13.7%.
On 8 July 2004 Mr Ferrario sent to CRSM a term sheet for the transaction and requested that Dr Fantini sign a “representation letter” (the purchase contract) and an indemnity letter. CRSM faxed the signed documents back to Barclays the same day and the purchase contract was thereby concluded. The contract included certain contractual terms and disclaimers which Barclays contended precluded the non-fraud claims (“the precluding terms”).
CRSM’s case was that Mr Montanari made clear during the discussions during this period CRSM’s requirements that:
the Notes should be a safe investment with an AAA rating;
the yield on the Notes should be adequate in the light of market conditions; and
the underlying assets in the CDO had to be stable and unexceptionable.
Mr Montanari’s witness statement evidence was that he recollected that Mr Ferrario consistently stressed the AAA rating of the CDO and assured him that, since the CDO² security underlying the Carnivale Note was AAA rated, it was an investment with a low risk of default; that it was a good investment; that it was virtually impossible for the Note to default; that it was a low-risk CDO full of AAA securities; and that it yielded more than a government bond only because it was less liquid.
Mr Montanari’s oral evidence did not support the making of any material oral representations by Mr Ferrario. In cross examination Mr Montanari confirmed that various points alleged to have been represented to him were simply matters of his own understanding, rather than matters actually stated to him by Mr Ferrario. In particular, Mr Montanari stated that he understood that the AAA rating meant that there was a low risk of default, as far as he knew, and that this was not something that Mr Ferrario had ever told Mr Montanari; Mr Montanari said that Mr Ferrario “only said there were some AAA notes, and in my mind they had a minimal risk of default”; in re-examination, he was asked whether, when he had meetings with Mr Ferrario at which the ARLO notes were discussed, Mr Ferrario had said anything to him about the credit risk or risk of default of the ARLO notes and his answer was “No, we never discussed it because AAA to us was synonymous of that tranquillo, safe and transparent that we were looking for.”
I also find that Mr Ferrario did not say that the Note yielded more than a government bond only because it was less liquid. As Mr Montanari acknowledged, it would make no sense to make such a comparison and I find that Mr Ferrario did not do so, although he did stress the illiquidity of the Note, as Mr Montanari accepted.
On the other hand Mr Montanari did make it clear to Mr Ferrario that CRSM required the Note to have a AAA rating. During the course of the discussions Mr Ferrario also made a comment to the effect that in 10 years in the business, he had have never seen a AAA tranche of a CDO go bust. Mr Montanari was looking for an investment which was stable and unexceptionable (“tranquillo” and “ineccipibile”) and he made this clear to Mr Sapignoli. There was no satisfactory evidence that this was communicated to Mr Ferrario, although Mr Ferrario knew that CRSM wanted an AAA rated instrument. I also find that Mr Montanari did say that CRSM were looking for a “good” return on their investment, and he acknowledged in evidence that 100bp above Euribor was a high yield for an AAA rated security at the time.
Many of the relevant discussions during this period were between Mr Ferrario and Mr Sapignoli and, since there was no evidence from Mr Sapignoli, CRSM was not in a position to challenge much of the evidence Mr Ferrario gave. I accept the substance of Mr Ferrario’s evidence in relation to the discussions at this time. In particular in two separate meetings, one with Messrs Sapignoli and Montanari, the other with Messrs Fantini and Montanari, Mr Ferrario went through the fact sheet scenario analyses illustrating the circumstances in which CRSM would suffer loss, up to a total loss. He explained that the proposed structures were highly leveraged and that their mark to market (“MTM”) value might fluctuate significantly, in particular might fall very far below par. Mr Ferrario was told that CRSM was not troubled by MTM fluctuation, provided there was a good chance of par redemption at maturity. Mr Ferrario indicated that the structure should redeem at par unless a loss-causing scenario of multiple defaults such as illustrated in the fact sheets occurred. Mr Ferrario also explained to Messrs Sapignoli and Montanari in the first meeting that CRSM had to take its own view on how likely that was to occur. His witness statement evidence, which was essentially unchallenged on this issue, described how “I went very carefully through the scenario analysis with CRSM, showing it that if a certain number of names (i.e. reference entities) defaulted, given recovery assumptions, the Notes would lose their entire value. I explained to Mr Montanari and Mr Sapignoli that they would have to use their own judgment to assess how many names were likely to default”. Also in that first meeting, Mr Ferrario went through Barclays’ qualifications and disclaimers with Messrs Sapignoli and Montanari.
Mr Montanari also gave evidence that, from when the proposal for the CDO was first put forward by Barclays, it was his intention to sell interests in the Note to CRSM’s own retail customers based on his understanding that it was a safe and suitable investment, and that he made Mr Ferrario aware of this (something which Mr Ferrario denied). I shall address this issue further below.
Reverting to the documents, Barclays then entered into loan agreements with Carifin and PlusValore dated 9 July 2004 for €200 million in aggregate.
Between 7 July 2004 and 15 July 2004, there were exchanges between Barclays and Standard and Poors (“S&P”) in which Barclays sought, and obtained, a AAA rating for Umbria II. This required, in certain respects, making amendments to Umbria II to satisfy S&P’s criteria for a AAA rating.
On 15 July 2004, the Umbria II constituting documentation was executed. The Umbria II constituting documentation was sent to CRSM for their signature. CRSM was required to sign the CDS confirmation for the Portfolio CDS that was the reference transaction for the Umbria II note, and the Constituting Instrument for the note, because it (CRSM) was the Replacement Selector, with certain entitlements to change CSO2 reference name populations during the life of the note. The relevant documentation was signed on 19 or possibly 26 July 2004.
Between 23 July 2004 and 28 July 2004, the Offering Documentation for Carnivale (including Supplement and Programme Memorandum) was sent to CRSM as it developed in draft, with the final versions, which incorporated the changes required by S&P, being sent on 28 July 2004. The trade was then completed on 28 July 2004; the Carnivale Notes were issued to Barclays with a nominal value of €200m and sold to CRSM. As finalised, they paid a coupon of 3.91% for an initial fixed interest period, then Euribor + 1.00% for the remainder of the life of the Notes.
On 13 August 2004 CRSM requested authorisation from its regulator, the Central Bank of San Marino, for a placing of the Carnivale Note. It attached a summary referring to the credit risk of the CDO as “AAA” and drawing from the scenario analyses contained in the Umbria II fact sheet referred to above. CRSM met the representatives of the Central Bank the following day, and authorisation was granted on 2 September 2004.
Mr Micheloni’s evidence was that in August 2004, he heard Mr Sapignoli discussing with Mr Ferrario a prospectus for Carnivale which CRSM had to submit to the Central Bank of San Marino. Mr Ferrario’s evidence was that whilst he may have given Mr Sapignoli certain details he was not aware he was participating in the drafting of a prospectus. I shall address this issue further below.
The Cernobbio and Later Notes
In September/October 2004 there was discussion of a further transaction. The idea that was developed was for the Delta companies to borrow €500m, divided into a series of tranches, with CRSM purchasing a Note corresponding to each tranche. Drawdowns under the loan agreements would match the purchase of the corresponding Notes. The total values of these later Notes was to be only €250m, referencing €100m of CDSs on the Delta companies and €150m of CSO2 or CCO notes. Thus, Barclays would have 20% credit protection via the CDSs and would sell CSO2 (or CCO) notes with a nominal value of 37.5% of Barclays’ Delta credit risk.
At a meeting in September/October 2004, Mr Ferrario and CRSM discussed whether the Carnivale structure would broadly be repeated or whether alternatives would be considered. One matter which was considered at this time was whether to vary the third Reference Asset from CSO2 notes to a CCO (Collateralised Commodity Obligation, a note referenced to commodity indices). Mr Montanari and Mr Sapignoli (until his departure from CRSM) favoured diversification, but Mr Morolli was against the idea, preferring that CRSM stick to CDOs and his view prevailed.
On 20 October 2004 Mr Ferrario sent by email to Mr Micheloni of CRSM two fact sheets relating to the next proposed transaction, along with three “coupon payment alternatives” and proposed letters to be signed by CRSM.
The first fact sheet described the proposed transaction as a whole, and was broadly similar in content to the Carnivale summary referred to above except that it envisaged CRSM purchasing a series of four Notes, the first with a notional amount of €100m followed by three further Notes with a notional amount of €50m each. These Notes would again comprise credit default swaps referencing loans to Carifin and PlusValore together with CSO Note with an expected rating of “[AAA] by [S&P]”.
The second fact sheet attached to Mr Ferrario’s email, which he referred to in the email as the “fact sheet della AAA” described the “Bramante: ABS CDO”. This fact sheet was in generally similar terms to the fact sheet for Umbria II, except that the duration of the CDO was stated to be 7.5 years (rather than 5 years) and the coupon was slightly lower, being 6 month Euribor + 0.95%. Like the fact sheet for Umbria II, it referred to the tranches offered, and the rating of the structure, as “AAA” and contained an “Analysis of Various Scenarios” which began with the statements:
“Bramante is a structure exposed to a highly diversified portfolio that can tolerate various defaults before the capital and coupon are damaged and eroded.
All the assets which make up the Bramante portfolio possess a level of subordination that guarantees protection against a significant number of defaults.”
Both fact sheets contained qualifications and disclaimers.
On 26 October 2004 Mr Ferrario sent to Mr Micheloni of CRSM details of the proposed portfolio for the new CDO², and explained how the new transaction had been improved by adding to the portfolio many reference entities which were not included in the previous (Umbria II) CDO, thereby limiting the overlap between the two transactions. The email also attached a revised version of the Bramante fact sheet based on the characteristics of the new portfolio. The fact sheet contained the same statements as the previous version, save that the default rates which the AAA tranche was said to be able to sustain would appear to have reduced. It stated:
“The Bramante AAA tranche is particularly robust in relation to defaults. The default rate of the AAA tranche can be compared in the various scenarios with the historical 7 year default rates in a portfolio with an average rating equivalent to the corporate exposures of the CSOs (Baa).
The AAA tranche can sustain a significantly higher default rate than the worst default rate that has occurred over a 7-year period between 1970 and 1997 (7 years 1986-1993). In Scenario 1 “Worst Case”, the AAA tranche can sustain a default rate of 11.6% per CSO. This cushion is 1.54 times higher than the worst 7-year historical default rate from 1981 to 1991. In Scenario 2 “Intermediate Case”, the AAA tranche can sustain a default rate of 12.15% per CSO, almost two times higher than the worst historical default rate.”
Later the same day (26 October 2004) Mr Ferrario sent the draft term sheet to Mr Micheloni, requesting signatures to the fact sheets and term sheet by Thursday (28 October 2004) in order to sign the contract with Carifin and PlusValore the following day. The email referred to the fact sheet for the “AAA” as having been “discussed last week”.
Mr Micheloni’s witness statement evidence was that in the course of their discussions during this period Mr Ferrario explained to him that CDO2s were volatile and illiquid, and that because of this the MTM value of the Notes could substantially decrease; however, provided CRSM held the Notes to maturity as it intended to do they should pay out 100% at that stage. Mr Ferrario’s evidence was that he did not say or imply that the CDO notes “should” pay out 100% at maturity in any sense other than that they would pay out unless there were sufficient defaults to cause a loss, and it was for CRSM to consider how the relevant entities might perform. I accept Mr Ferrario’s evidence on this issue and indeed Mr Micheloni accepted in cross examination that if Mr Ferrario had referred to a payment of 100% at maturity, he appreciated that a scenario of payment of 100% at maturity was unless defaults happened to destroy the protection in the structure.
Mr Micheloni sent an email the following day, 27 October 2004, asking a number of questions about the fact sheet. He was then put in touch with Mr Agresta who (according to an email to Mr Ferrario) “explained everything to him on the phone”.
On the morning of 28 October 2004, Mr Ferrario sent an email to Mr Micheloni referring to this discussion and attaching an updated version of the portfolio. Mr Ferrario stated:
“I hereby confirm that the portfolio of the new transaction is on paper of better quality than the one of the former transaction, and this has been made possible thanks to the longer maturity of the transaction.”
Mr Ferrario also attached a further revised fact sheet, this time entitled “Como: ABS CDO” but otherwise in identical terms to the “Bramante” fact sheet sent on 26 October 2004 referred to above.
On the morning of 28 October 2004, following a further discussion with Mr Montanari about the terms of the transaction, Mr Ferrario sent the two fact sheets to CRSM with a request that they should be signed on each page. Mr Ferrario also said that the final term sheet would follow later that day.
It appears that, before closing the deal, Mr Ferrario again wanted to know what profit Barclays would be making. Thus, at 12.07 on 28 October 2004 Mr Agresta sent an email to Mr Ferrario setting out the estimated profit from the sale of the Como CDO. This showed an estimated gross profit of 24.1% and net profit of 15%.
The signed fact sheets were duly faxed back by CRSM to Barclays at 14.48. Also at that time, the “term sheet” referred to in Mr Ferrario’s earlier email was sent to Mr Micheloni for signature on Mr Ferrario’s instructions. The attached document in fact comprised a purchase letter to confirm CRSM’s agreement to purchase from Barclays further Notes in four tranches, referred to respectively as Cernobbio, Bellagio, Tremezzo and Menaggio, together with the term sheet for the Cernobbio Note. The purchase letter was duly signed by CRSM. It contained the same precluding terms as the contract for the Carnivale Note.
On 29 October 2004 Barclays entered into further loan agreements with each of Carifin and PlusValore (for €500 million in aggregate), which were drawn down in stages as CRSM purchased successively the Cernobbio, Bellagio, Tremezzo and Menaggio Notes.
The Como I notes were issued on 19 November 2004.
S&P rated the Como I notes AAA on 22 November 2004 (after exchanges between Barclays and S&P in which S&P had required or caused various changes to be made to the Notes to secure the AAA rating, including a small increase in the subordination).
The Cernobbio Notes were issued with a nominal value of €100m, and the sale to CRSM completed, on that day. Barclays provided to CRSM the final Cernobbio Note and Como note documentation (including Programme Memorandum and Supplement) on 24 November 2004.CRSM’s signature was required as replacement selector on the Portfolio Credit Swap for the Como I notes, and this was obtained on 25 November 2004.
Meanwhile, on 16 November 2004 Mr Micheloni sent Mr Ferrario a summary of the Cernobbio Note which he was proposing to take to a meeting with the Central Bank of San Marino on 18 November, to seek its approval to sell participations in the Note to CRSM’s retail customers. In the email Mr Micheloni asked Mr Ferrario to call him; and it was Mr Micheloni’s recollection (although disputed by Mr Ferrario) that Mr Ferrario contacted him as requested and confirmed he was happy with the contents of the summary. The summary highlighted the AAA credit risk of the CDO and also included information taken from the Como fact sheet about the ability of the CDO to withstand defaults. The summary was submitted to the Central Bank on 18 November 2004. Revised versions (with changes to the wording of the declaration that investors were required to sign) were submitted on 23 November and on 25 November 2004. The Central Bank gave its approval for the placing/selling of the investment on 10 December 2004.
CRSM contended that this email and follow up telephone conversation confirms that, as was the evidence of Mr Montanari and Mr Micheloni, Mr Ferrario was aware that CRSM were selling the Notes to its retail customers. Mr Ferrario denied that this was so, although he stated in evidence that at some stage he came to suspect this. In addition to the evidence of Mr Montanari and Mr Micheloni CRSM contended that it was inherently probable that Mr Ferrario was so told since that was the fact. Further, Mr Ferrario acknowledged having a discussion with Mr Sapignoli in August 2004 about details of the Carnivale Notes and it was Mr Micheloni’s evidence that this concerned the draft prospectus for retail clients. Yet further, the November 2004 email attached a draft prospectus which was clearly directed at retail clients and was then discussed with Mr Micheloni. I accept that Mr Ferrario’s evidence on this issue was not entirely satisfactory. On the other hand, I accept his evidence that if he had known that sales were being made to retail clients it would have raised concerns because of his unfamiliarity with the regulatory requirements of the Central Bank of San Marino and of how that might affect Barclays. I find that Mr Sapignoli did discuss the Carnivale Notes prospectus with Mr Ferrario over the telephone in August 2004. Mr Ferrario understood that this was a document required by the Central Bank but did not appreciate that it was so required for the purpose of sales to retail customers. I also find that Mr Ferrario did discuss the emailed prospectus with Mr Micheloni in November 2004, although he did not consider its terms in detail. He did realise that the Notes might be passed on to retail customers in repo, but not that they would be sold participations in the Notes, although he did become suspicious of this around November 2004. I also find that, contrary to his witness statement, he did not ask Mr Montanari or Mr Sapignoli to confirm that there would be no retail sales.
The Bellagio Note
In early December 2004 Mr Ferrario asked colleagues within Barclays to work both on a further CDO and on a CCO.
On 10 December 2004 Mr Agresta approached a different ratings agency, Fitch, stating that Barclays was “looking to rate an innovative transaction” involving the “new technology” of an “extendable CSO”. The idea was that the issuer would have an option at the end of year 4 to extend the trade by another 2 years; the rating would be based only on the initial 4-year period and the client would take the rating risk at the time of the extension option, if exercised.
On 17 December 2004 Mr Agresta sent a version of the proposed fact sheet (described as the “marketing memo”) to Mr Ferrario, followed by some wording for Mr Ferrario to use setting out how Barclays had “tried to improve the portfolio as much as possible”. Mr Ferrario duly passed on this explanation to CRSM in an email which attached the fact sheet for “the new AAA CDO2”.
Both the email and the fact sheet mentioned that the CDO had an initial period of 4 years but that the issuer had an option to extend this to 7 years. The fact sheet described the offered tranche as “Senior AAA from Fitch” and stated that “[t]he structure has a rating of AAA from Fitch at the date of issue”. Neither in the email nor the fact sheet, however, was it stated that the AAA rating applied only to the initial 4-year period and that it did not cover the option period.
As with the fact sheets for Umbria II and Como I, the “Analysis of Different Scenarios” included the following statements:
“Como is a structure exposed in a highly diversified portfolio able to tolerate different defaults before capital or dividends are notched and eroded.
All of the assets that make up the Como portfolio have a subordination level such as to guarantee protection against a significant number of defaults.”
It also stated as follows (in bold):
“The AAA tranche of Como is particularly robust with respect to defaults. It is possible to compare the default rate of the AAA tranche in the different scenarios with the 7 year historical default rates of a portfolio with an average rating corresponding to that of corporate exposures of CSOs (Baa), assuming, in the worst case, that only corporate exposures are subject to default. Also, in the worst case, it is assumed that historical default rates are 7 year rates. Como’s AAA tranche may withstand a default rate significantly higher than the worst default rate that has occurred in any 7 year period from 1940 to 1991. In scenario 1 “Worst Case”, the AAA tranche can withstand a default rate of 14.14[%] per CSO. This cushion is double the worst historical default rate. In scenario 2 “Intermediate Case”, the AAA tranche can withstand a default rate of 17% per CSO, more than double the worst historical default rate.”
A few minutes after sending the fact sheet (at 19:36 on Friday 17 December 2004) Mr Ferrario sent a follow up email (at 19:44) adding that the terms needed to be finalised by New York close of business on Monday (20 December 2004) in order to meet the Tuesday deadline.
On Tuesday, 21 December 2004 Mr Ferrario sent to CRSM for signature an amended version of the purchase letter and the final version of the fact sheet. The fact sheet included the same statements as quoted above. Mr Ferrario asked for CRSM’s signature on the email, each page of the fact sheet, and the signature pages of the purchase letter. Since Mr Fantini was away, Mr Ferrario asked for Marco Belluzzi, a senior internal lawyer at CRSM, to sign the documents initially, and for Mr Fantini to sign them again on his return. The requested documents duly signed by Mr Belluzzi were returned the same day. The versions signed by Mr Fantini then followed on 23 December 2004. This formed the Bellagio purchase contract.
The purchase letter amended and restated the purchase letter of 28 October 2004. It recorded that the Tranche 1 (i.e. Cernobbio) Notes had been purchased on 22 November 2004, stated that CRSM “hereby agreed to purchase” the Tranche 2 (i.e. Bellagio) Notes, and contained revised agreements as to how and when CRSM would purchase the later Tremezzo and Menaggio notes, for which the CCO option was retained. It contained the same precluding terms as before, and the Termsheet contained the same disclaimers and warnings as before. The indicative Como II termsheet that was contained within it stated that “Expected Rating by Fitch – [AAA]”.
Again, Mr Ferrario had asked Mr Agresta to let him know Barclays’ estimated profit from the transaction. According to an email from Mr Agresta sent that afternoon, the gross profit was 24.2% and the net profit (after all costs and reserves) was 12.4%.
Mr Agresta corresponded with Fitch in order to obtain the credit rating for Como II. On 10 January 2005 he mentioned in an email to Mr Ferrario that Fitch had confirmed the rating. In subsequent correspondence with Fitch, however, Fitch provided a number of comments on the transaction documentation and were concerned to make sure that their rating was properly represented, with the limitation in its scope made clear. Fitch provided its rating letters in relation to Como II on 19 January 2005.
The Tremezzo Note
The Tremezzo Note was the fourth of the five Notes, and the last to contain a CDO² element. Probably because it referenced the same CDO² as the Bellagio Note, there appears to have been relatively little further discussion about the Tremezzo Note.
On 17 January 2005, Mr Ferrario sent to CRSMdocuments for the contract to purchase the Tremezzo notes, namely a Termsheet and fact sheet, which described the anticipated terms of the Tremezzo Notes, calling them a “CLN linked to two Italian corporate loans and a AAA-rated CSO”, and explaining that the Notes would be linked to “the Como II Note in an amount of €[30]m”. Similar fact sheet/Transaction Outline disclaimers were included as before. The Termsheet contained the Amended and Restated Purchase letter for CRSM’s signature which contained precluding terms as before. It referred to “Notes initially due 2009” and stated “Expected Rating by Fitch [AAA]”.Mr Fantini of CSRM signed and returned (at least) the signature pages of the Indicative Fact Sheet and Purchase Letter as requested that same day. No new fact sheet for the Como II notes was sent out to CRSM.
The ratings letter for Como II was issued on 19 January 2005, and CRSM received it on or shortly after that date. It stated: “Fitch Ratings assigns the following rating to the following securities issued by the above-referenced issuer: EUR70,000,000 “Como II” Class A Secured Limited Recourse Extendable CLNs initially due 2009: AAA. The rating only addresses the probability of a claim being made under the master portfolio credit default swap for the period ending on 20 March 2009 (The Original Scheduled Termination Date). The rating will not address any potential exercise of the Extension Option”.
On 19 January 2005, the Como II Offering Documentation was executed. It recorded that the Fitch Rating applied to initial maturity.
On 20 January 2005, the Bellagio constituting documentation was executed and sent through to CRSM, and CRSM bought the Bellagio Notes, according to their terms, with a nominal value of €54m.
The second drawdown under the second loan then occurred on or about 20 January 2005.
On 18 February 2005, CRSM agreed to amend the purchase letter for the later Notes, pursuant to its request to re-allocate the loans as between Plusvalore and Carifin, which required a rewriting of the CDSs, by an Amended and Restated Purchase Letter signed by Mr Vladimiro Renzi. It contained the preclusionary terms.
On 21 February 2005, the Tremezzo constituting documentation was executed, and was probably sent to CRSM on or around this date, and CRSM bought the Tremezzo Notes, according to their terms, with a nominal value of €52m. The Tremezzo documentation referenced the Como II notes as issued previously.
The third draw down under the second loan appears probably to have occurred on or around 21 February 2005.
Barclays’ profits on the Tremezzo Note, as relayed to Mr Ferrario on 19 January 2005, were a gross profit of €9,191,020 (30.6%) and a net profit (after all reserves and costs) €5,203,302 (17.3%).
Barclays’ pricing summary dated 27 January 2005 drew together the net present values (NPV) to Barclays of all three CDO²s, Umbria II, Como and Como II. These totalled approximately €42 million.
The Menaggio Note
The Menaggio Note was the last of the Notes, and included a CCO component rather than a CDO². Mr Ferrario sent the fact sheet and term sheet to CRSM for signature on 24 February 2005 and the signed documents were faxed back the following day.
As it did not include a CDO², the Menaggio Note does not form part of CRSM’s claim.
Restructuring of the Notes
After Mr Buoncompagni joined CRSM in March 2005, Mr Morolli asked him to conduct an analysis of the CDO2 Notes. For this purpose Mr Buoncompagni created a spreadsheet listing the reference entities which he linked to Bloomberg to obtain updated information about spreads and credit ratings.
At this time there was a crisis in the US car industry and general concern about portfolio reference entities connected with the industry. There was some discussion between Mr Ferrario and Mr Buoncompagni about this, which discussion included the possibility of CRSM paying for the removal of certain names. This led to a discussion within Barclays during the course of March 2005 of possible changes to the CRSM portfolios which would enable certain names to be removed.
On 14 March 2005 Mr Micheloni sent an email to Mr Ferrario asking to be sent the final version of all the portfolios of the different CDO2s. This information was required by Mr Buoncompagni for the analysis of the CDO2s which Mr Morolli had asked him to carry out.
Mr Ferrario sent spreadsheets showing the portfolios (and including the current spread for each name) to Mr Micheloni on 18 March 2005. The portfolio spreadsheets were provided to Mr Ferrario by Mr Agresta, who in turn had asked Mr Bechet on 15 March 2005 to prepare this information. In his email asking Mr Bechet to prepare the portfolios Mr Agresta said: “U have to use the existing trades not the optimised ones yet”.
When he forwarded the existing portfolios to Mr Ferrario on 16 March 2005 Mr Agresta said in his email: “I wanted to draft a more detailed proposal to convince them to make the switch from Umbria to Como II.”
At 13.45 on 18 March 2005 Mr Agresta sent an email to Mr Ferrario with the “more detailed proposal” for CRSM that he had wanted to draft. The email starts by referring to “the monitoring that we usually carry out on the CDO positions of San Marino” and goes on to set out a proposal to switch Umbria II and Como I to Como II (on the ground that “this will eliminate all the Intelsat risk and the Bombardier risk”) and to reduce the exposure to Delphi from all 6 to 3 of the CDOs underlying Como II. Mr Agresta re-sent this email to Mr Ferrario at 14.49 on 18 March 2005, asking “What do you think of this?”
When Mr Ferrario sent the portfolios to CRSM on 18 March 2005, he mentioned in his email that the names with the highest spreads were Bombardier, Delphi and Intelsat and that “we had taken away Bomb[ardier] and Intelsat from Como II”.
On 21 March 2005 Mr Agresta sent two more emails to Mr Ferrario on the subject of “Switch Umbria and Como into Como II”. In the first email Mr Agresta stated that “[t]he new Como II portfolio is definitely better than Umbria II and Como”; and in the second email he said that “the Delphi spread has increased to 550bps over 5 years ... we have to act before it is too late”.
On 11 April 2005 Mr Agresta sent an email to other members of the structuring team stating that Mr Ferrario had asked for valuations of the San Marino trades. In his email Mr Agresta said: “We are trying to convince him [i.e. Mr Ferrario] to convince the client to switch Umbria and Como I into Como II”.
On 13 April 2005 Mr Agresta sent an email to Mr Ferrario (and another member of the sales team) attaching the spreads of the CRSM reference entities and summarising the strategy that “I am thinking to adopt”, namely (1) to “improve the portfolio of Como II with 10-15 ‘tactical’ changes for improvement in rating” and (2) to switch Umbria and Como I into Como II.
This was followed by another email from Mr Agresta to the sales team sent on 15 April 2005 in which his approach had now changed. He said: “The spreads have widened and all positions are now below par. For them to repurchase on an equal footing would mean to have a compliance problem (off-market transaction). I suggest you begin to make the switch in Como I and Como II and wait for Umbria to get closer to par (which is quite likely with the passage of time and if spreads narrow). At the same time I would begin working with the client to get across the concept of the switch between Umbria II and Como II, that we are improving.”
On Wednesday, 20 April 2005 a meeting of CRSM’s Finance Committee took place at which CRSM’s exposure to General Motors was discussed and also the progress of Mr Buoncompagni’s CDO study analysis. The minutes also record that Mr Ferrario would be visiting San Marino on 22 April 2005.
On 21 April 2005 Mr Agresta prepared the draft email setting out proposals for the restructuring of Como I and Como II which in its final form was sent to CRSM on 22 April 2005.
The email sent to CRSM on 22 April 2005 stated as follows;
“This is the email with the switch ideas.
During the monitoring which we usually carry out on San Marino’s positions, we believe that it is possible to effect some substitutions aimed at:
1. eliminating specific default risks
2. stabilizing the note rating
These two objectives may be obtained at cost “zero”, by carrying out other “tactical” substitutions for the improvement of the portfolio “spread”, taking also advantage of some specific “axes” of the book both of correlation and of single names. In particular, we would be prepared to “update” the underlying technology to the CDO-squared adding the “cross subordination” which adds a further protection level to the structure (attached you will find some introductory slides on technology and the idea of switch).
For the substitution proposal we have followed 2 “objective” criteria:
1. rating of the single names
2. rating of Barclays Capital Research (according to the Scaramagna model).
It is in fact possible to better both the rating of the single names, and the quantitative rating of Scaramagna.
In addition to substitutions, we have also modified the overlap of the single names in order to minimise the substitution cost. In particular, we have reduced Delphi’s overlap from 5 or 6 times to 3 times for Como I and Como II.
In the attached file you can find a detailed substitution proposal for Como I and for Como II.
Below please find a quantitative summary of the effected substitutions.”
The final version of Mr Agresta’s email which contained these proposals for restructuring was sent to Mr Ferrario at 11.44 on 22 April 2005. It was then sent by Mr Agresta to Mr Micheloni (copied to Mr Ferrario) at 14:24 when Mr Ferrario was in San Marino. I accept that it is likely that the email was sent to Mr Micheloni so that he could print out the email and its attachments for discussion at the meeting with Mr Ferrario.
Attached to Mr Agresta’s email were: (1) spreadsheets listing the entities in the existing portfolios and proposed new portfolios and specifying the reference entities removed and added (with their credit ratings and Scaramanga ratings), and (2) a set of PowerPoint slides entitled “CSO Trade Idea” which described the concept and suggested benefits of cross-subordination. These attachments were discussed at the meeting and Mr Ferrario gave a presentation which he had brought with him of the “Quantitative Credit Relative Value Indicator” developed by Barclays’ analyst, Mr Scaramanga.
There was a factual issue as to whose idea it was to add cross-subordination to the structure of the CDO2s. CRSM’s witnesses recalled that (like the rest of the proposal presented on 22 April) the suggestion came from Barclays. However, Mr Ferrario claimed that it was requested by CRSM. I find that Mr Ferrario’s recollection is wrong on this and that the proposal came from Barclays.
The email of 22 April 2005 stated that the substitutions could be made at “cost zero”. This was to be achieved by “carrying out other ’tactical’ substitutions for the improvement of the portfolio ’spread’, taking advantage also of some specific ’axes’ of the book both of correlation and of single names”. CRSM’s case was that the description “cost zero” conveyed that Barclays was not intending to make any profit from the substitutions and that this was borne out by statements made by Mr Ferrario at the time. I shall address this issue when considering the alleged restructuring representations.
The email of 22 April 2005 made substitution proposals only for Como I and Como II, and not for the other CDO2, Umbria II. In two emails sent on 26 and 27 April 2005, Mr Agresta told Mr Ferrario that he would like to add cross-subordination and to start to make some switches also in Umbria, “especially because that trade has a 4 year maturity and as time passes it will become harder to re-optimise it”. Mr Agresta added that “[i]n a few months Umbria if the spreads shrink they will be very close to equal”.
By 3 May 2005 Mr Agresta had worked up a proposal for changes to the Umbria II CDO². Mr Agresta outlined this proposal in an email sent to Mr Ferrario that day stating:
“We would like to show San Marino a switch also on Umbria because we think that it has much sense. In addition to adding cross-subordination, we would reduce the overlap of Intelsat and Bombardier from 6 to 3 ... We would significantly improve both the ratings and the distribution of ratings of Scaramanga.”
The email then set out statistics showing how both credit ratings and the Scaramanga “relative value ratings” would be improved by the proposal.
On 10 May 2005 Mr Ferrario emailed to Mr Buoncompagni spreadsheets of the portfolios showing the “relative value” rating for each name, which Messrs Bechet and Agresta had prepared. This email indicated that it is being sent ahead of a meeting with CRSM the next day (11 May 2005) when Mr Ferrario would be arriving with “substitutions to suggest”.
On 11 May 2005 (at 17:24) Mr Agresta sent Mr Ferrario an email and attached spreadsheets showing the proposed switches. Mr Agresta sent a slightly revised version of the email at 11:24 the next morning (12 May 2005), which Mr Ferrario sent on to CRSM.
The email sent to CRSM on 12 May 2005 attached the proposed new portfolios for each of the three CDO2s and explained that the switch on the three structures was intended to achieve the following objectives:
“1) improvement of the stability of the structure and technological ’upgrade’ with the addition of cross-subordination.
2) improvement of the quality of the portfolio on the basis of two ’objective’ criteria: agency rating and rating of Scaramagna’s quantitative model.
3) total or partial removal of specific risks you identified.
4) maintaining and stabilizing the explicit rating at AAA.”
The email goes on to comment on the extent to which Barclays had removed particular names in the portfolios about which CRSM had expressed concern. It then refers to what are said to be the “details of the portfolio improvements” and states:
“the new portfolios are of much better quality compared to the current ones, both in terms of their ratings (lower % of high yield names) and Scaramanga ratings.”
The email continues:
“Since the new structures are substantially different from the former ones, in particular because of the addition of the cross-subordination, we have had to change the attachment and detachment points of the underlying CDOs and CDO-squareds. Notwithstanding the portfolio improvement and the addition of cross-subordination, we have managed to increase the subordination of the underlying CDOs both in Umbria and Como I. On the other hand, the subordination of the CDO-squareds has lowered since, thanks to the cross-subordination and greater subordination of the underlying CDOs, the structure benefits from greater protection. ...”
Set out in the remainder of the email were statistics for each of the CDO2s which showed how, as measured by the two stated “objective” criteria of credit ratings and the “relative value ratings” generated by Barclays’ own Scaramanga model, the quality of the reference portfolios had in each case improved. The email ended with various disclaimers.
On 16 May 2005 Mr Agresta sent an email to Mr Ferrario (attaching slides on cross-subordination translated into Italian) in which Mr Agresta stated: “Clearly, from my point of view I would prefer to do the whole amount. If we really have to choose, I would prefer to do the switch on Umbria and Como I (whilst Como II is struggling the most of the three).”
By 18 May 2005 Barclays had produced a draft memorandum relating to the restructuring. Mr Ferrario sent to Mr Agresta a draft of the memorandum which was in due course sent to CRSM on 2 and 3 June 2005. Attached to the 3 June 2005 memorandum were CDS spread figures for the overlap names.
It was Mr Agresta’s evidence, which I accept, that his role in the preparation of the memoranda was limited. The principal draft was mainly produced by compliance and the most senior salespeople and it was carefully reviewed by them. This is borne out by the documents which show a drafting process occurring between Mr Ferrario, Bartolomeo Acquaviva (a senior director in sales), Claudia Tarentino of compliance, and Andrew Whittle (head of trading). In an email of 27 May 2005 from Mr Ferrario to Mr Agresta it was stated that compliance had revised the memorandum and that sales and compliance are “still in the process of discussing if further language is needed for trading”. So far as Mr Agresta is concerned it said: “What we need from you: 1) Get from Scaramanga the latest update on what will be published in his June monthly; 2) work on the portfolio and update the numbers of the memo and scenario analysis”.
Following receipt of the 3 June 2005 memorandum, CRSM’s Finance Committee met that day and the decision was made to proceed with the restructuring.
On 7 June 2005 Mr Ferrario sent updated versions of the memorandum and other documents to Mr Buoncompagni. The changes were said to include “further reduction of Delphi and GMAC”. The attached spreadsheet included a column showing the current CDS spreads of the names in the existing and new portfolios.
Later that day Mr Ferrario forwarded an explanation, prepared by Mr Agresta, to Mr Buoncompagni of the difference between Barclays’ method and CRSM’s method for calculating the average spread of the five names with the highest spreads.
On 7 and 8 June 2005 Mr Agresta and Mr Bechet sent details of the amendments to the CDO²s to the rating agencies for their approval. There followed some discussions with them and some final adjustments to the portfolio by Barclays.
The final versions of the memorandum, portfolios and Forced Default Analyses were produced by Mr Agresta and sent by Mr Ferrario to CRSM on 13 June, followed by the documents for signature to give effect to the restructuring.
On 13 June 2005, Barclays then sent through the legal documents that were necessary to implement the proposed restructuring transaction. They included parts of the constituting documents for the CSO2 notes, which needed to be amended, and also a draft representation letter.
On 14 June 2005, Mr Fantini of CRSM signed on each page and returned the representation letter for the restructuring. This included precluding terms such as those in the purchase contracts.
On 15 June 2005, CRSM signed parts of the transaction documents which were necessary to give effect to the restructuring, including signatures to the Portfolio Supplemental Credit Swaps also Noteholder Resolutions authorising the amendments to the Constituting Instruments of the CSO2 notes. Other consequential revisions to the documentation were implemented on 14-16 June 2005. In each case, the interest rate swaps and credit default swaps were amended to add cross-subordination, adjust attachment points, and amend the reference portfolios.
Barclays’ documents include various P&L breakdowns from 14 June 2005 onwards. Later in the month Mr Ferrario asked for P&L figures so that he could keep track of his sales credits. The figures that he was given on 27 June 2005 for net structured P&L were €3.18m (3.18%) for Umbria II, €2.55m (4.25%) for Como I and -€0.35m (equivalent to -0.85%) for Como II, plus US$652k and €300k flow desk P&L, making a total of approximately €6.1m. The ‘flow desk’ P&L represented profit made by Barclays’ dealers in buying and selling protection on the names removed from and added to the portfolios. Mr Ferrario’s sales credits were €3.18m for Umbria II, €3.18m for Como I, €0.35m for Como II and €0.83m for flow desk profits making a total of €6.9 million.
There was further discussion within Barclays about how much profit to recognise or release immediately; it appears that the final figures for the P&L to be released at the end of June 2005 were -€0.5m for Como II, €3.0m for Como I and €3.7m for Umbria II, giving a total net profit recognised at the end of June 2005 of €6.2m.
Subsequent Events
Even after the restructuring, CRSM remained concerned about the CDO2s and the risks which they might contain. As a result of these concerns, the decision was taken that the participations in the Notes which CRSM had sold to customers should be repurchased (at par), as CRSM was no longer confident that the investments were as safe as it had originally believed and led its customers to believe. The repurchases took place between July and November 2005.
In early October 2005 a meeting was held between CRSM and Barclays to discuss possible further ways of reducing the risks on the CDO²s, such as by removing specific reference entities or increasing the subordination of the outer CDOs (at a cost to CRSM, in either case). Costings were provided on 21 October 2005.
In the meantime, on 14 October 2005 Barclays served a Credit Event notice on CRSM in respect of Delphi, one of the reference assets underlying the CDO²s.
A possible further restructuring was discussed at a meeting with Barclays on 17 November 2005. However, the proposals were not taken up.
At a further meeting on 1 December 2005 Barclays provided bid prices of 77 for the Umbria II CDO² (down from 92 in October), 65 for Como I (down from 80 in October) and 45 for Como II (down from 55 in October). Barclays also put forward various proposals to replace Umbria II with other financial products
On 23 December 2005 S&P downgraded the Umbria II CDO² from AAA to AA negative watch. On the same date Barclays put forward a proposal for discussion to amend the Bellagio and Tremezzo Notes which were linked to Como II so as to guarantee repayment of at least 65% of their principal on maturity in March 2012 but with the CDO² element of those Notes ceasing to pay any coupons. The proposal was not taken up. A similar proposal appears to have been mooted again in March 2006 and again not pursued.
In April 2006 Barclays agreed to repurchase the Umbria II CDO2 underlying the Carnivale Note. The repurchase was effected by complex documentation the essential effect of which was that CRSM as Noteholder authorised Barclays and the Issuers to liquidate the CDO² and pass the proceeds to CRSM. CRSM entered into this transaction subject to an express reservation of all its rights. The “Dirty Purchase Price” (that is, the amount including the accrued but unpaid interest and the unwind cost) of the Umbria II CDO2 was equal to €98,632,711.11 (compared to a par value of €100,000,000).
In November 2006 Barclays agreed to repurchase the Como I CDO2 underlying the Cernobbio Note. The transaction was entered into on the same basis as the repurchase of the Umbria II CDO2, including an express reservation of CRSM’s rights. The Dirty Purchase Price for the Como I CDO2 was equal to €57,505,632 (compared to a par value of €60,000,000).
In January 2007 there appears to have been a further suggestion by Barclays of the possibility of converting the Como II CDO² underlying the Bellagio and Tremezzo Notes into a principal protected note paying zero coupon or a low coupon. Again, CRSM did not take up the proposal.
By the end of July 2007 Barclays’ valuation of Como II had fallen to 58.4% of par. Subsequent defaults eroded the principal and coupons paid by the CDO². In February 2010 Barclays formally notified CRSM that the principal amount of Como II had reduced to nil (compared to a par value of €70,000,000).
III. The Contracts
A. The Purchase Contracts
CRSM’s investment in each Note was effected by its purchase of the Note from Barclays at par, pursuant in each case to an antecedent contract, governed by English law, by which CRSM agreed to buy such Note.
The contracts in question were made as follows:
For Carnivale, the purchase contract was created when CRSM signed and returned to Barclays the “representation letter” that was Appendix IV to a termsheet dated 8 July 2004 concerning the Carnivale Notes;
For Cernobbio, the purchase contract was created when CRSM signed and returned to Barclays on 4 November 2004 a purchase letter dated 28 October 2004, to which had been attached a termsheet dated 28 October 2004 concerning the Cernobbio Notes;
For Bellagio, the purchase contract was created when the October purchase letter referred to in (2) above was supplemented and amended by a purchase letter dated 21 December 2004 which CRSM signed and returned to Barclays, together with copies (each signed by CRSM on every page) of a covering e-mail from Mr Ferrario and a fact sheet concerning the proposed Como II CSO2 (referred to, in fact, as “Como” in the document). One of the attachments to the December purchase letter had been a termsheet dated 21 December 2004 concerning the Bellagio Notes;
For Tremezzo, the purchase contract was created when the October purchase letter contract, as varied in December, was further supplemented and amended by a purchase letter dated 17 January 2005, the signature pages of which CRSM signed and returned to Barclays, together with a copy (signed by CRSM, on every page) of an indicative fact sheet for the Tremezzo Notes. The attachment to the January purchase letter had been a termsheet dated 17 January 2005 concerning the Tremezzo Notes.
It was common ground that CRSM was a party to the contracts contained in the Notes themselves in the capacity of Replacement Selector under the Replacement Annexes to the Notes. Other than in this respect, however, CRSM was not a party to any of the transactions forming part of the structure of the Notes or the CDO2s.
The key contractual documents are therefore the purchase letters dated 8 July 2004 (as regards the Carnivale Notes, which referenced the Umbria II CDO2), 28 October 2004 (as regards the Cernobbio, Bellagio and Tremezzo Notes, which in the case of Cernobbio referenced Como I, and in the case of Bellagio and Tremezzo referenced the Como II CDO2), 21 December 2004 (in the case of the Bellagio and Tremezzo Notes) and 17 January 2005 (in the case of the Tremezzo Notes).
The impact of the letters of 21 December 2004 and 17 January 2005 was the subject of some dispute. It was CRSM’s case that but for the 28 October 2004 purchase letter CRSM would have had no commitment to purchase any further notes and in practice there is no reason to believe the later letters would have been entered into. As such, it was sufficient for it to show that the 28 October 2004 purchase letter contract was induced by misrepresentation for it to be able to recover in respect of Como I and Como II. If that was not shown, but it was established that the later contracts had been induced by misrepresentation, then damages were still recoverable but on a slightly different basis.
Barclays contended that the correct analysis was that Cernobbio was purchased under the contract of 5 November 2004 (slightly amending the contract of 28 October 2004), Bellagio under the contract of 21 December 2004 (re-signing the contract of 21 December 2004), and Tremezzo under the contract of 17 January 2005. If those contracts were not induced by actionable misrepresentation, then there can be no recoverable loss in respect of what was actually purchased, even if, in respect of Bellagio and Tremezzo, the earlier 28 October 2004 purchase letter contract had been induced by actionable misrepresentation. Barclays submitted that this followed from the fact that CRSM was not obliged to sign the contracts of 21 December 2004 and 17 January 2005 by the 28 October 2004 purchase letter contract. It was obliged to purchase certain notes by the earlier contract, but not to enter into new contracts for the purchase of notes. Further, the Notes it agreed to purchase on 21 December 2004 and 17 January 2005 were not the notes which it was obliged to purchase by the 28 October 2004 purchase letter contract. Specifically, the 28 October 2004 letter contract contained obligations to purchase notes referencing either more of Como I, or CCOs, as part of the later note tranches: see terms (c), (h) and (i). There was no reference to the purchase of Como II, which was referenced by Bellagio and Tremezzo as actually purchased.
In the unlikely event it makes any difference on the facts I consider Barclays’ analysis to be correct and that CRSM have to prove that misrepresentations induced the 21 December 2004 and 17 January 2005 contracts.
Barclays relied on three terms of the purchase contracts which it contended preclude CRSM from advancing its claims of misrepresentation under s.2(1) of the Act (but not its claims in deceit). These were clause 5 (“Non-reliance”); clause 6 (“Asessment and Understanding”), and clause 8 (“Terms and Conditions”).
CRSM submitted that none of these clauses affords any defence to a claim for misrepresentation under the Act.
In the Amended Defence, and until the start of the trial, the only terms relied on by Barclays which were alleged to have contractual effect were terms of the purchase letters. The Offering Documentation was only said to be relevant as part of the “background” against which, on Barclays’ case, the purchase letters are to be construed.
In its Opening Submissions, however, Barclays advanced an argument that the terms of the Offering Documentation were incorporated in the purchase contracts. This is said to have been effected via the term sheets, which are themselves referred to in the purchase letters, and/or by clause A8 and the agreement therein that “operative terms and conditions” would be those set out in “Offering Documentation”.
Taking the example of the letter of 8 July 2004 relating to the Carnivale Notes, the “Notes” were defined in the heading of the letter as: “Secured Limited Recourse Credit Linked Notes due 2009 … of Arlo II Limited …, as set forth in the termsheet dated 8 July 2004 … attached hereto”.
The full text of the opening paragraph of the body of the purchase letter was as follows:
“This letter confirms our agreement on the Trade Date to purchase the Notes. In connection with such agreement, we hereby make all the representations, warranties and covenants set forth below. We agree that such representations and warranties shall be deemed to be repeated on the Issue Date by reference to the facts and circumstances then in existence. We will not proceed with a purchase of the Notes on the Issue Date if any of such representations or warranties are no longer true and correct unless otherwise agreed in writing with you.”
The letters then state:
“A. Purchase of the Notes.In connection with our agreement to purchase the Notes, we hereby represent, warrant and covenant as follows: …”
“B. Purchase for Re-sale. If we are purchasing Notes as principal with a view to the re-sale thereof to one or more third parties ..., the following representations, warranties and covenants apply in addition to those set forth in Section A above. ...”
The purchase letters therefore drew a distinction between:
the terms on which CRSM agreed to purchase the Notes from Barclays (which were set out in the letters themselves at Sections A + B), and
the term sheet, which was relevant only because it described the Notes (i.e. the terms of the contract which would, following transfer of the Notes to CRSM, arise between CRSM and the issuer of each Note).
I agree with CRSM that the purchase letters did not incorporate the terms of the term sheet as terms of the purchase contract between CRSM and Barclays. The term sheet was relevant merely in order to identify the characteristics of that which CRSM was agreeing to purchase from Barclays. Further, the term sheet was not a contractual document; it styled itself as a “description” and a “summary only” of the terms of the proposed transaction. According to the term sheet the terms of the proposed transaction were to be found in the Offering Documentation and not in the term sheet. Even if the term sheet had itself been a contractual document (which it was not), it did not incorporate the terms of the Offering Documentation, let alone the entire contents of the Offering Documentation.
The representations, warranties and covenants set out in section A of the purchase letters included clause 8 which provided as follows:
“We acknowledge that the operative terms of and conditions of the Notes will be exclusively those set forth in the relevant Offering Documentation and that we are not entitled to rely on any description of the terms and conditions of the Notes or any undertaking by any other party in respect of the Notes that is not set forth in the relevant Offering documentation, including, without limitation, any such description or undertaking communicated orally or set forth in any pitchbooks or other marketing materials. For purposes of this letter, “Offering Documentation” shall mean the Programme Memorandum and any relevant supplements thereto in respect of the Notes.”
Barclays submitted that this gives rise to incorporation in two ways. First, the incorporation is effected by the agreement that “the operative terms of and conditions of the Notes will be exclusively those set forth in the relevant Offering Documentation”, the intention being to refer to the entirety of the terms and conditions in the Offering Documentation, including the risk factors and disclaimers and not merely the “Terms and Conditions” section of the Programme Memorandum. Although the words “terms and conditions” are used, they are not capitalised. The risk factors and disclaimers are terms or conditions in the relevant sense, capable of being contractual if incorporated, as they were intended to be. Second, in any event, the incorporation is effected by the agreement that “we are not entitled to rely on any description of the terms and conditions of the Notes or any undertaking by any other party in respect of the Notes that is not set forth in the relevant Offering documentation..”
I agree with CRSM that this provision did not incorporate the contents of the Offering Documentation as terms of the purchase contract. It is referring to the terms and conditions of the Notes not the purchase contract. Further, even if it did so it would only be referring to the “Terms and Conditions” of the Offering Documentation, which did not include the disclaimers sought to be relied upon.
It follows that the only precluding terms which are part of the purchase contract are those set out in the purchase letter itself, as Barclays pleaded.
B. The Restructuring Contract
The restructuring contract is contained in a letter dated 14 June 2005, signed on every page by CRSM and sent by it to Barclays. The letter referred to the fact that amendments to the terms of the Carnivale, Cernobbio, Bellagio and Tremezzo Notes had been discussed between Barclays and CRSM. The upshot of those discussions had been summarised in a Memorandum from Mr Ferrario to Messrs Morolli, Micheloni and Buoncompagni sent the previous day, 13 June 2005, together with a spreadsheet setting out the agreed Reference Name alterations and a PowerPoint presentation about cross-subordination.
Barclays relied on three portions of the letter signed by CRSM dated, in different versions, 14 or 15 June 2005 (“the Restructuring Letter”) as precluding the claims made.
First, the body of the Restructuring Letter includes the statement that “In connection with our agreement to amend the Notes, we hereby represent, warrant and covenant as follows with respect to each series of the Notes”. This is followed by a series of provisions closely following clauses 1-6, 8, 10 and 12 of the original 8 July 2004 purchase letter, adapted to refer to the restructuring as opposed to the purchase of the Notes.
These provisions include clauses 6, 7 and 8 headed “Non-Reliance”, “Assessment and Understanding” and “Terms and Conditions”, which are substantively the same as those in the 8 July 2004 purchase letter (save that they refer to the restructuring of the Notes and the “Non-Reliance” clause omits the final sentence denying any assurance or guarantee as to the expected results of the purchase of the Notes).
Second, after the body of the Restructuring Letter and the signature block, is a section headed “Certain Risk Factors”. The body of the letter refers to this section in the following passage:
“We acknowledge and agree that the amendments to the terms of the Notes have been agreed in full consultation with us. We acknowledge each of the Risk Warnings and Disclaimers attached to this letter. We further make all the representations, warranties and covenants set forth below.”
The “representations, warranties and covenants” referred to in the last sentence are the clauses referred to above. The “Risk Factors” are referred to in the middle sentence of the wording quoted above. Thus they are “acknowledge[d]”, but unlike the first sentence, not “acknowledge[d] and agree[d]” nor the subject of representations, warranties or covenants. Accordingly CRSM submitted that they have the status of a notice but no contractual force.
This was disputed by Barclays. It contended that by “acknowledging” the Risk Factors, CRSM was agreeing, with contractual effect, that it is proceeding to contract on the basis of those factors. Thus, for example, by “acknowledging” the “Independent Review and Advice” Risk Factor, CRSM agrees contractually that, as that risk factor states, it is “responsible for making [its] own independent investment decision”; and by “acknowledging” the “Credit Risk” Risk Factor, CRSM agrees contractually that, in order to enter into the transaction, it “should have such knowledge and expertise and credit risk that [it is] capable of evaluating the merits of the transaction”.
The text and format of the “Risk Factors” relied on by Barclays, closely follows that of the “Risk Factors” set out in the fact sheets.
Third, Barclays relies on the “Disclaimers” paragraph which follows the end of the “Risk Factors” section of the restructuring letter. This gave rise to the same issue as to whether it had contractual status as the Risk Factors.
In my judgment there is a distinction between the body of the letter, which comprises the contract and is signed as such, and the “Risk Factors” and “Disclaimers” which follow it. This is supported by the text of the letter itself which draws a seemingly deliberate distinction between the contents of the letter, which are acknowledged and agreed, and the “Risk Factors” and “Disclaimers”, which are merely acknowledged. I accordingly accept CRSM’s case that these have the status of a notice. As appears below, that would be consistent with how the equivalent provisions were treated in respect of the purchase contracts.
Non-contractual disclaimers
Barclays contended that even if the various further disclaimers that they sought to rely upon were not incorporated into the contracts they were nevertheless important background and context against which the allege misrepresentations had to be considered.
Of particular importance are the qualifications and disclaimers contained in the fact sheets since it was the contents of these documents which were central to the alleged Purchase Representations.
The body of the fact sheets were translated into Italian. At the end of the document there followed various qualifications and disclaimers in English.
Under a heading in bold: “Qualification and Disclaimers” the following was set out in capitals:
“Certain Risk Factors
Investors should review the Programme Memorandum of the issuer and the Supplement there under in respect of the notes, including the Risk Factors and Investments Considerations contained in it, prior to making a decision to invest in the Notes. The purchase of the notes involves substantial risks and is suitable only for sophisticated investors who have the knowledge and experience in financial and business matters necessary to enable them to evaluate the risks and the merits of an investment in the Notes. The Notes are not principal protected and purchasers of notes are exposed to full loss of principal. Before making an investment decision, prospective purchasers of Notes should consider carefully, in the light of their own financial circumstances and investment objectives, all the information set out in the Offering Documents for Notes and, in particular, the following considerations..”
Various considerations under different headings were then set out, including the following (with added numbering):
“[II] Volatility. The Market value of the notes (whether indicative or firm) will vary over time and may be significantly less than par (or even zero) in certain circumstances. The Notes may not trade at par.
[III] Highly Leveraged Investment. The reference Portfolio is significantly larger than the principal amount of the Notes. Accordingly, the Notes represent a highly leveraged investment. The use of leverage is a speculative investment technique to enhance returns. However, leverage also will magnify the adverse impact of Credit Events in the Reference Portfolio.
[IV] Limited liquidity. No secondary market exists for the Notes upon issue and there can be no assurance that a secondary market for the Notes will develop…..
[V] Credit Ratings. Credit ratings of debt securities represent the rating agencies opinions regarding their credit quality and are not a guarantee of quality. Rating agencies attempt to evaluate the safety of principal and interest payments and do not evaluate the risks of fluctuations in market value; therefore, credit ratings may not fully reflect the true risks of an investment.
[VI] Default Rates. Reliable sources of statistical information may not exist with respect to the defaults of all types of assets comprising the Reference Portfolio… Actual default rates may exceed historical default rates or the default or other assumptions referenced herein or in the models used by ratings agencies.
[VII] No reliance. Investors may not rely on the issuer or any programme Parties in connection with its determination as to the legality of its acquisition of the Notes or other matters referred to in these Risks Factors. Neither the issuer nor any programme Parties is acting as an investment advisor, or assumes any fiduciary obligation, to any investor of Notes. Neither the issuer nor any programme parties assumes any responsibility for conducting or failing to conduct any investigation into the business, financial condition, prospects, creditworthiness, status and/or affairs of any Reference Entity. Purchasers of the Notes may not rely on the views or advice of the issuer for any information in relation to any person other than the issuer itself.
[VIII] Independent Review and Advice. Each investor is responsible for its own independent appraisal of and investigation into the business, financial condition, prospects, creditworthiness status and affairs of any reference Entity and Reference obligations, as well as the risks in respect of the Notes and their terms, including, without limitation, any tax, accounting, credit, legal and regulatory risks. Each investor is fully responsible for making its own investment decision and whether the Notes (i) are fully consistent with its (or if its acquiring the Notes in a fiduciary capacity, the beneficiary’s) financial needs, objectives and condition, (ii) complies and is fully consistent with all investment policies, guidelines and restrictions applicable to it whether acquiring the Notes as principal or in a fiduciary capacity) and (iii) is a fit, proper and suitable investment for it (or, if it is acquiring the Notes in a fiduciary capacity, for the beneficiary….
[IX] Credit Risk. Investors should have such knowledge and experience in financial and business matters and expertise in assessing credit risk that it is capable of evaluating the merits, risks and suitability of investing in such Notes including any credit risk associated with the Reference Entities and the issuer. None of the Programme Parties will have any responsibility or duty to make any such investigations, to keep any such matters under review or to provide the prospective purchasers of such notes with any information in relation to such matters or to advise as to the attendant risks.
[X] No Representations. None of the Programme Parties make any representations or warranty, express or implied, as to any reference Entity or any Reference Obligation or any information contained in any documents provide by any Reference Entity to any of them or to any other person or filed by any Reference Entity with any exchange or with any governmental entity regulating the offer and sale of securities. “
There then followed a section headed “Disclaimers” in bold capitals. This included the following:
[XI] …. The information is provided for information and discussion purposes only. The information does not constitute an offer, solicitation or recommendation to enter into any transaction or to buy, sell or hold any security…. The information contains a summary of certain proposed terms of a hypothetical offering of securities as currently contemplated in connection with preliminary discussions with potential investors and does not purport to be a complete description of all material terms or of the terms of an offering that may be finally consummated…. You should independently assess the riskand rewardsand suitability of any investment in the light of your goals and experience. You should assess whether it complies with your policies and applicable laws….
[XII] We [Barclays] are solely an arms length contractual counterparty. We are not your advisor or judiciary, and none of our communications constitute financial, investment or other advice of any kind. We may not make a market in future in the investment. We and our affiliates may buy, sell, or hold positions identical to or opposite from the investment….No assurance is given that any indicated returns, performance or results will be achieved… information may be derived from sources generally believed to be reliable, but Barclays does not represent or warrant that such information is accurate or complete, and it should not be relied on as such.
[XIII] The investment is liquid, its value is volatile and can suffer from adverse or unexpected market moves, counterparty default or adverse events involving any underlying reference obligation or entity. You may suffer the loss of your entire investment. This brief statement does not detail all risks of the investments, and you should ensure you fully understand its terms, risks and applicable operational, legal, tax and accounting considerations, prior to transacting.
[XIV] The information is illustrative and is not intended to predict actual results, which may differ substantially fro those reflected in the information. Performance analysis is based on certain assumptions with respect to significant factors that may prove not to be as assumed. Other assumptions may materially impact the results reflected in the information. You should understand the assumptions and evaluate whether they are appropriate for your purposes. Performance results are based on mathematical models that use inputs to calculate results. As with all models, results may vary significantly depending upon the value of the inputs given…
[XV] The information addresses only certain aspects of the applicable investments characteristics and thus does not provide a complete assessment. As such the information may not reflect the impact of all structural characteristics of the investment, including call events and cash flow priorities at all prepayment speeds and/or interest rates. You should consider whether the behaviour of these investments should be tested under assumptions different from those included in this information. The assumptions underlying the information, including structure and collateral, may be modified from time to time to reflect changed circumstances. Any investment decision should be based only on the data in the Offering Documents and then current version of the information.
[XVI] The investment described herein will not be guaranteed by Barclays or its affiliates and neither Barclays nor any of its affiliates nor any director, officer, employee or agent of any of them accepts any liability whatsoever for any direct, indirect, consequential or other loss arising from use of the information.”
It was Barclays’ case that these provisions prevented the fact sheets from containing or implying any representation, or at least any representation to the effect of the alleged purchase representations.
In relation to the restructuring Mr Ferrario’s email of 12 May 2005 concluded with the following:
“The information herein has been obtained from sources believed to be reliable but Barclays Capital does not warrant that it is accurate and complete. Neither Barclays capital, nor any officer or employee thereof accepts any liability whatsoever for any direct or consequential loss arising from any use of this publication ... any modelling ... is not intended to be a statement as to future performance.”
The attached memorandum and all subsequent memoranda ended with a section headed “Qualifications and Disclaimer”. These set out similar considerations under similar headings as the fact sheets, as did the Restructuring Letter itself. The memoranda further stated as follows:
“Barclays is acting solely as principal and not as advisor or fiduciary. Accordingly you must independently determine, with your own advisors, the appropriateness for you of this transaction before transacting. Barclays accepts no liability whatsoever for any consequential losses arising from the use of this document or reliance on the information contained herein.”
“This document is an indicative summary of the terms and conditions of the transaction and may be amended, superseded or replaced in subsequent summaries. The final terms and conditions of the transaction will be set out in full in a binding transaction document. ...”
“Barclays, its affiliates and the individuals associated therewith may (in various capacities) have positions or deal in transactions or securities identical or similar to those described herein.”
“This document does not disclose all the risks and other significant issues related to an investment in the transaction. Prior to transacting, potential investors should ensure that they fully understand the terms of the transaction and any applicable risks, which may not be fully explained above.”
It was Barclays’ case that these provisions prevented there being any representation to the effect of the alleged restructuring representations.
Barclays also had an alternative case of estoppel. It submitted that if the various protective clauses and disclaimers relied upon in the Offering Documentation are not terms of the parties’ respective contracts, they found an estoppel and that a parallel estoppel is derived from the exclusions and disclaimers in the fact sheets which CRSM signed. It submitted that CRSM had represented that the roles of the parties defined thereby would be respected and that no representations inconsistent with such terms would be relied upon, which representations had been relied upon by Barclays, as borne out by its insistence of the fact sheets being signed. If the provisions relied upon neither contractually preclude the claim nor prevent the alleged representation from being made then I am not satisfied that an alternative estoppel case could succeed. In any event it was only explored in a very generalised manner in the evidence and has not been satisfactorily proved. In particular no clear or specific representation or common assumption has been made out.
IV. Misrepresentation: The Law
Deceit
As stated in Clerk & Lindsell on Torts (20th ed.) para 18-01, quoted in AIC Ltd v ITS Testing Services (UK) Ltd (The “Kriti Palm”) [2007] 1 Lloyd’s Rep. 555, 617, para 398, per Buxton LJ:
“The tort [of deceit] involves a perfectly general principle. Where a defendant makes a false representation, knowing it to be untrue, or being reckless as to whether it is true, and intends that the claimant should act in reliance on it, then in so far as the latter does so and suffers loss the defendant is liable for that loss.”
Misrepresentation Act, section 2(1)
Section 2(1) of the Act, on which CRSM relies in the alternative to its claim in deceit, provides as follows:
“2.— Damages for misrepresentation.
(1) Where a person has entered into a contract after a misrepresentation has been made to him by another party thereto and as a result thereof he has suffered loss, then, if the person making the misrepresentation would be liable to damages in respect thereof had the misrepresentation been made fraudulently, that person shall be so liable notwithstanding that the misrepresentation was not made fraudulently, unless he proves that he had reasonable ground to believe and did believe up to the time the contract was made that the facts represented were true.”
In order to establish a right to damages under s.2(1), it is therefore necessary for CRSM to prove (a) a representation made by Barclays to CRSM, which (b) was false, and (c) induced CRSM to enter into the relevant contract, (d) as a result of which CRSM has suffered loss. If these elements are proved, then Barclays would have a defence under s.2(1) if it proves that it had reasonable ground to believe, and did believe, up to the time the contract was made that the facts represented were true.
The requirements for a claim under s.2(1) are therefore the same as for a claim in deceit, subject to the important difference that under s.2(1) it is not necessary for the claimant to prove that the misrepresentation was made fraudulently. Rather, the Act expressly provides that, where the other requirements of the tort of deceit are met, the person making the misrepresentation is liable under s.2(1) “notwithstanding that the misrepresentation was not made fraudulently”, unless he proves that he reasonably believed the facts represented to be true.
A recent case which contains helpful discussion of a number of aspects of the law of misrepresentation is the recent decision of Christopher Clarke J in Raiffeisen Zentralbank v RBS plc [2011] 1 Lloyd’s Rep 123.
Making a representation
A representation is a statement of fact made by the representor to the representee on which the representee is intended and entitled to rely as a positive assertion that the fact is true. In order to determine whether any and if so what representation was made by a statement requires (1) construing the statement in the context in which it was made, and (2) interpreting the statement objectively according to the impact it might be expected to have on a reasonable representee in the position and with the known characteristics of the actual representee: see Raiffeisen, supra, at [81]; Kyle Bay Ltd v Underwriters Subscribing under Policy No. 01957/08/01 [2007] Lloyd’s Rep IR 460, 466, at [30]–[33], per Neuberger LJ.
In order to be actionable a representation must be as to a matter of fact. A statement of opinion is therefore not in itself actionable. However, as stated in Clerk & Lindsell para 18-13:
“A statement of opinion is invariably regarded as incorporating an assertion that the maker does actually hold that opinion; hence the expression of an opinion not honestly entertained and intended to be acted upon amounts to fraud.”
In addition, at least where the facts are not equally well known to both sides, a statement of opinion by one who knows the facts best may carry with it a further implication of fact, namely that the representor by expressing that opinion impliedly states that he believes that facts exist which reasonably justify it – see Clerk and Lindsell para 18-14, citing among other cases Smith v Land and House Property Corp (1884) 28 Ch D 7, 15, per Bowen LJ, and Brown v Raphael [1958] Ch 636.
A statement as to the future may well imply a statement as to present intention: “that which is in form a promise may be in another aspect a representation” - Clerk & Lindsell, para 18-12, quoting Lord Herschell in Clydesdale Bank Ltd v Paton [1896] AC 381, 394.
Silence by itself cannot found a claim in misrepresentation. But an express statement may impliedly represent something. For example, a statement which is literally true may nevertheless involve a misrepresentation because of matters which the representor omits to mention. The old cases about statements made in a company prospectus contain illustrations of this principle – for example, Oakes v Turquand (1867) LR 2 HL 325, where Lord Chelmsford said (at 342-3):
“... it is said that everything that is stated in the prospectus is literally true, and so it is; but the objection to it is, not that it does not state the truth as far as it goes, but that it conceals most material facts with which the public ought to have been made acquainted, the very concealment of which gives to the truth which is told the character of falsehood.”
In relation to implied representations the “court has to consider what a reasonable person would have inferred was being implicitly represented by the representor’s words and conduct in their context”: per Toulson J in IFE v Goldman Sachs [2007] 1 Lloyd’s Rep 264 at para. 50. That involves considering whether a reasonable representee in the position and with the known characteristics of the actual representee would reasonably have understood that an implied representation was being made and being made substantially in the terms or to the effect alleged.
In a deceit case it is also necessary that the representor should understand that he is making the implied representation and that it had the misleading sense alleged. A person cannot make a fraudulent statement unless he is aware that he is making that statement. To establish liability in deceit it is necessary “to show that the representor intended his statement to be understood by the representee in the sense in which it was false” – per Morritt LJ in Goose v Wilson Sandford & Co. [2001] Lloyd’s Rep PN 189 at para. 41. In other cases of misrepresentation this is not a requirement, but one would generally expect it to be reasonably apparent to both representor and representee that the implied representation alleged was being made.
It is necessary for the statement relied on to have the character of a statement upon which the representee was intended, and entitled, to rely. In some cases, for example, the statement in question may have been accompanied by other statements by way of qualification or explanation which would indicate to a reasonable person that the putative representor was not assuming a responsibility for the accuracy or completeness of the statement or was saying that no reliance can be placed upon it. Thus the representor may qualify what might otherwise have been an outright statement of fact by saying that it is only a statement of belief, that it may not be accurate, that he has not verified its accuracy or completeness, or that it is not to be relied on: Raiffeisen,supra, at [86].
In Raiffeisen, supra, at [85], Christopher Clarke J cited a dictum of Rix J that, because of the broad measure of damages currently available under s.2(1) of the Misrepresentation Act in the light of Royscot Trust v Rogerson [1991] 2 QB 297, “where there is room for an exercise of judgement, a misrepresentation should not be too easily found”: Avon Insurance Plc v Swire Fraser Ltd [2000] 1 All ER (Comm) 573 at [200]. Although nothing is likely to turn on the point in this case, I respectfully agree with CRSM that this is questionable in principle. The fact that Parliament (as interpreted by the Court of Appeal) has thought it right to provide a broad measure of compensation where a contract has been made as a result of a misrepresentation should not affect the prior question whether there has been a misrepresentation, all the more so where, as alleged in this case, the damages claimed would be recoverable on ordinary principles, irrespective of the decision in Royscot. Barclays reserved the right to challenge the correctness of the decision in Royscot should this case go further and, as I observed in Cheltenham BC v Laird [2009] IRLR 621 at para. 524, Royscot is a controversial decision bearing in mind in particular:
“The rationale of there being a special rule for damages in fraud is one of morality and deterrence – per Lord Steyn in Smith New Court Securities Ltd v Scrimgeour Vickers (Asset Management) Ltd [1997] AC 254 at 280. Such considerations do not apply, or at least do not apply to anything like the same degree, in cases of mere negligence.
It can be said to be anomalous and unsatisfactory for there to be major differences in the damages recoverable for negligent misrepresentation under section 2(1) of the 1967 Act and for negligent misrepresentation at law.
The wording of section 2(1) – “if the person making the representation would be liable to damages in respect thereof had the representation been made fraudulently, that person shall be so liable notwithstanding that the representation was not made fraudulently, unless he proves that he had reasonable ground to believe and did believe up until the time that the contract was made that the facts represented were true” – does not necessarily compel the conclusion that the liability in damages for misrepresentation under section 2(1) is to be the same as that for fraud. These words can be construed as meaning no more than that whereas before the 1967 Act there would only be a liability in damages for pre-contractual misrepresentation if the misrepresentation was made fraudulently, now there is also to be liability where the misrepresentation is made negligently. It can be considered as referring to the existence of liability rather than its extent.
There is strong, if not unanimous, academic criticism of the decision – see, for example, Chitty on Contracts Vol 1at 6-070; McGregor on Damages (17th edn.) at 41-045-6; Treitel on The Law of Contract (12th edn.) at 9-063; (1991) 107 LQR 547.”
As further observed in Raiffeisen, at [87], the claimant must show that he in fact understood the statement in the sense (so far as material) which the court ascribes to it; and that, having that understanding, he relied on it. Analytically, this is probably not a separate requirement of a misrepresentation claim but rather is part of what the claimant needs to show in order to prove inducement.
Fraud
The classic statement of the mental element required to found a claim in deceit remains that of Lord Herschell in Derry v Peek:
“First, in order to sustain an action of deceit, there must be proof of fraud and nothing short of that will suffice. Secondly, fraud is proved when it is shown that a false representation has been made (1) knowingly, (2) without belief in its truth, or (3) recklessly, careless whether it be true or false. Although I have treated the second and third as distinct cases, I think the third is but an instance of the second, for one who makes a statement under such circumstances can have no real belief in the truth of what he states. To prevent a false statement from being fraudulent, there must, I think, always be an honest belief in its truth.”
As to recklessness, even if the party making the representation may have had no knowledge of its falsehood, he will still be responsible if he had no belief in its truth and made it, “not caring whether it was true or false” - See Clerk & Lindsell, para 18-21. As Lord Herschell put it Derry v Peek, supra, at 368 (and 361):
“Any person making such a statement must always be aware that the person to whom it is made will understand, if not that he who makes it knows, yet at least that he believes it to be true. And if he has no such belief he is as much guilty of fraud as if he had made any other representation which he knew to be false, or did not believe to be true.”
It is not necessary that the maker of the statement was ‘dishonest’ as that word is used in the criminal law - Standard Chartered Bank v Pakistan National Shipping Corp (No. 2) [2000] 1 Lloyd’s Rep. 218, 224. Nor is the defendant’s motive in making the representation relevant: “If fraud be established it is immaterial that there was no intention to cheat or injure the person to whom the false statement was made.” - Clerk & Lindsell, para 18-20, quoting Bradford Third Benefit Building Society v Borders [1941] 2 All ER 205, 211 per Viscount Maugham; and see also Derry v Peek, supra, at 409. What is required is dishonest knowledge, in the sense of an absence of belief in truth - The Kriti Palm, supra, para 257 (Rix LJ); and see also para 258, quoting Armstrong v Strain [1951] TLR 856, 871, per Devlin J (“When Judges say, therefore, that wickedness and dishonesty must be present, they are not requiring a new ingredient for the tort of deceit so much as describing the sort of knowledge that its necessary”).
The ingredient of dishonesty (in the above sense) must not be watered down into something akin to negligence, however gross - The Kriti Palm, supra, para 256. However, the unreasonableness of the grounds of the belief, though not of itself supporting an action for deceit, will be evidence from which fraud may be inferred. As Lord Herschell pointed out in Derry v Peek, supra, at 376, there must be many cases:
“where the fact that an alleged belief was destitute of all reasonable foundation would suffice of itself to convince the court that it was not really entertained, and that the representation was a fraudulent one.”
Where a serious allegation (such as deceit) is in issue, this does not mean the standard of proof is higher. However, the inherent probability or improbability of an event is itself a matter to be taken into account when weighing the probabilities and deciding whether, on balance, the event occurred. The more improbable the event, the stronger must be the evidence that it did occur before, on the balance of probability, its occurrence will be established - The Kriti Palm, supra, para 259, quoting Lord Nicholls in re H (Minors) [1996] AC 563, 586.
An innocent employer may be vicariously liable for a deceit committed by an employee in the course of his employment, as for any other deliberate tort - Clerk & Lindsell, para 18-26; Lloyd v Grace, Smith & Co [1912] AC 716.
The innocent principal is liable where a fraudulent agent passes on a representation indirectly, for example through another (innocent) agent of the same principal - Clerk & Lindsell par 18-25, citing London County Properties v Berkeley Property Co [1936] 2 All ER 1039 (CA), as interpreted by the CA in Armstrong v Strain [1952] 1 KB 232.
Inducement
As analysed by Christopher Clarke JinRaiffeisen, supra, at [153]-[199], to establish inducement for the purpose of a claim under s.2(1) of the Misrepresentation Act, it is necessary to show that, but for the representation, the claimant would not have entered into the contract that he did.
In that case, Christopher Clarke J concluded that where a fraudulent misrepresentation has been made, the requirement is weaker: it is sufficient to show that the representation was a factor in the claimant’s decision and that but for it he might have acted differently - Ibid at [196]-[199], referring to Barton v Armstrong [1976] AC 104 and Barton v County NatWest [1999] Lloyd’s Rep 408; and see also Dadourian Group International v Simms [2009] 1 Lloyd’s Rep 601, 618 at [99] + [101]. This conclusion was challenged by Barclays. It submitted that a fraudulent representation must cause a loss to create a cause of action and to do so it must cause the entry into the contract from which the loss is said to arise. It follows that it must induce the representee to enter into the contract and be a cause of him doing so. It is not necessary to resolve this issue but I propose to proceed on the basis that the approach of Christopher Clarke J is correct.
V. The Representation Case
The context
I have made various findings as the relevant factual background as set out above. Both parties stressed various background matters as being of particular relevance.
CRSM submitted that the relevant context included the following circumstances:
It was impossible to form a rational assessment of the risk of default of the CDO2s without carrying out a statistical analysis which required a mathematical model - It is correct that the complexity of the product meant that statistical modelling was the best method of assessing the risk of default. However, the CRAs used such modelling in order to produce their ratings. Further, CRSM was in a position to assess the likelihood of default of individual names and could relate that to the structural information provided by Barclays and to the number of defaults required to impact that structure.
Very few investors at the time had the ability to model complex structured products such as CDO2s, and CRSM did not – This is correct, although CRSM had the ability to have this carried out by independent experts, which they did for the purpose of valuations for their accounts.
It was normal practice at the time for investors to rely as a principal measure of the credit risk (i.e. risk of default) of the product on a credit rating obtained by the seller from a CRA agency – This is correct. At the time CRA ratings were widely relied upon by investors in products such as the Notes. They were generally perceived as providing a valuable and reasonably reliable assessment by independent experts of the degree of default risk involved.
An investment with a rating of “AAA” is generally considered to have a minimal risk of default – This is correct.
CRSM specifically asked for a product with a AAA rating and did so (as Mr Ferrario knew) because its requirement was for a safe investment which would not expose the bank to significant risk and which it could sell on to its customers – It is correct that CRSM asked for a product with an AAA rating and that Mr Ferrario would have appreciated that they were looking for an investment with a default risk consistent with such a rating. As already found, he did not know that they were looking to sell the product on to their customers.
Barclays created the CDO2s specifically for CRSM, designing every feature of their structure and selecting the proposed reference assets – This is correct although Barclays would have taken features from other like products it had structured.
Barclays also stressed various contextual matters. I accept the main points made with certain additions and changes as set out below:
Mr Ferrario’s commercial relationship with CRSM dated back to 1996, from his time at Deutsche Bank and later Goldman Sachs. By the time he joined Barclays in August 2003, Mr Ferrario’s principal contact at CRSM was Mr Sapignoli. Mr Sapignoli was regarded by Mr Ferrario and within CRSM as being someone with good technical skills who would consider documents in detail.
CRSM was familiar with the fact that trading with major international banks (in particular, London and New York based banks) involved information or presentations on sophisticated financial products coming with detailed qualifications and disclaimers. It was acknowledged by all CRSM’s factual witnesses that such qualifications and disclaimers are invariably present, although their precise terms may differ. This was borne out by the documents and the expert evidence. The essential principle underlying these qualifications and disclaimers is one of caveat emptor. The buyer is meant to make his own assessment of the risks of the transaction and to make an independent decision as to whether to enter into it.
As CRSM would have expected, Barclays’ sale of the Notes was overseen by various departments, in particular compliance and sales. The fact sheets reflected their input, were provided on the basis of the disclaimers set out and acknowledged through signature and would not have been provided on any other basis.
Mr Ferrario was told that CRSM was looking for assets providing a “good” yield, of about 100bp over Euribor, to back the relatively high interest rates it offered to its clients on their deposits. CRSM said they wanted that to be achieved with a AAA rating. In relation to the Notes, CRSM was interested in having a product which could be sold to its clients although it did not make clear to Barclays that that was its intention.
The return of 100 bps above Euribor was significantly more than that being paid on other AAA products at the material time. Part of the difference may have been attributable to the illiquidity of the Notes but it would also be reasonably understood as reflecting a higher risk than other AAA products paying less.
In two separate meetings, one with Messrs Sapignoli and Montanari, the other with Messrs Fantini and Montanari, Mr Ferrario went through the fact sheet scenario analyses as I have found above. Also in that first meeting, Mr Ferrario went through Barclays’ qualifications and disclaimers with Mr Sapignoli and Mr Montanari.
Mr Ferrario did say words to CRSM to the effect that in 10 years in the business, he had not seen a AAA CDO tranche go bust, or words to that effect.
The Purchase Claims
In their trial Opening four representations were alleged by CRSM, one express and three implied, namely:
(expressly) that the CDO2 notes “were secure investments with an extremely low risk of default”;
(impliedly) that Barclays believed the CDO2 notes to be “secure investments with an extremely low risk of default”;
(impliedly) that Barclays “had structured [them] with that intention”.
(impliedly) that Barclays “was acting in good faith and was not knowingly making statements likely to mislead”.
These representations were said by CRSM to have been made by various Barclays documents provided to CRSM, by Mr Ferrario’s assistance in the preparation of two CRSM marketing documents, and by statements allegedly made by Mr Ferrario, to Mr Montanari and Mr Sapignoli at CRSM’s offices in San Marino in the weeks leading up to the purchase of the Carnivale Note, and to Mr Micheloni at CRSM’s offices in San Marino in September or October 2004.
By the end of the trial CRSM’s case had narrowed considerably. It was no longer alleging that oral representations had been made and the case had become one of representations to be implied from documents, and in particular the fact sheets, considered in their context, which included the oral discussions between the parties, as to which I have already made findings.
CRSM’s case was that in selling the CDO2 Notes to CRSM, Barclays impliedly represented that the Notes had (and by implication that Barclays believed and expected them to have) a very low risk of default. That representation was made with the knowledge that it was likely to be relied on by CRSM, and CRSM did rely on it in agreeing to buy the Notes.
The alleged representations (“the Purchase Representations”) were therefore:
that the CDO2 Notes had a very low risk of default;
that Barclays believed and expected the CDO2 Notes to have a very low risk of default
In relation to (1), if, contrary to CRSM’s primary case, that was considered to be a matter of opinion rather than fact, then it was impliedly represented that Barclays had reasonable grounds for such opinion.
CRSM contended that these representations were made:
By describing the Notes to CRSM (in the fact sheets but also in many other communications) as “AAA”; and
By other statements made in the fact sheets.
Although CRSM no longer contended that oral representations had been made by Mr Ferrario, it did contend that he had said nothing to detract from the representations made by the documents, and that what he had said had re-inforced that message. In this connection it relied in particular on his admitted statement of words to the effect that in 10 years of being in the business, he had never seen a AAA tranche of a CDO go bust.
As to (1), CRSM pointed out that the AAA rating of the CDO2s was prominently set out in the fact sheets and other documents provided by Barclays to CRSM. Taking the Umbria II fact sheet as an example, the AAA rating was featured in the “Terms of the operation” box on the first page, and under “Structural characteristics” on the second page. The investment was then referred to as the “AAA tranche” nine times in sections 4 and 5 of the fact sheet.
Further, in his communications with CRSM, Mr Ferrario often referred to the CDO2s simply as “the AAAs”. Examples are:
On 27 April 2004 Mr Ferrario reported on a meeting with CRSM that one of the points accepted by CRSM at the meeting was “to do a AAA for the 50% of the risk that we take Gruppo Delta”.
On 14 July 2004 Mr Ferrario sent to Mr Sapignoli documentation relating to Umbria II and asked him to sign the documents governing the substitution of credits underlying “the AAA”.
On 20 October 2004 Mr Ferrario sent an email to CRSM attaching documents to be discussed at a meeting on 22 October 2004 including “Fact sheet for the AAA”.
On 26 and 28 October 2004 Mr Ferrario sent emails to CRSM attaching various documents for signature. On each occasion one of the documents for which he requested a signature was the “Factsheet in Italian for the AAA discussed last week”.
On 4 November 2004 Mr Ferrario sent an email to CRSM summarising the amounts of the further Notes including the amount in each case “which is AAA”.
On 17 December 2004 Mr Ferrario sent an email to CRSM attaching the “Fact sheet for the new AAA CDO2 operation”.
When questioned about the first two of these examples, Mr Ferrario agreed that he had used the expression “the AAA” as a shorthand to refer to the CDO2 Notes, but denied that he used this shorthand because the AAA rating was the central characteristic of the CDO2s so far as CRSM was concerned. He claimed that he “just happened” to use this particular shorthand of the many that he could have used. CRSM contended that this was implausible.
When considered in context, including in particular the fact that the CDO2s were structured by Barclays and tailor-made to meet CRSM’s expressed requirements for a safe investment, CRSM contended that these descriptions of the CDO2s as ‘AAA’ would naturally have been understood by any reasonable person in CRSM’s position to mean – as it was understood by CRSM to mean – that the Notes had (and were believed and expected by Barclays to have) a very low risk of default.
As to (2), CRSM submitted that aside from the frequent references to AAA rating there were three (or, in the case of Como II, four) further features of the product, as structured and described by Barclays, which would have reinforced the impression that the investment was one with a very low risk of default.
First, a high proportion (82%) of the portfolio of each CDO2 consisted of Asset-Backed Securities (ABS) which were themselves all rated AAA. Indeed, each of the three CDO2s was described in the title of the fact sheet as an “ABS CDO”. The fact sheets also stated, under the heading “Diversified and high quality portfolio”, that the portfolio included “High quality ABS – only AAA-rated ABS are selected”. Further, the ABSs all had a stipulated 90% recovery rate.
Second, the inner CDOs which comprised the remaining 18% of the portfolio themselves all had a minimum rating of A or AA. The fact sheets stated that the “Diversified and high quality portfolio” also included “High quality CSO: minimum rating equal to [A]” (Umbria II and Como I) or “High quality CSO: minimum rating equal to AA of Fitch” (Como II). They further stated that all the assets within the inner CDO portfolios would themselves have investment grade ratings.
According to the rating definitions of S&P, a credit rating of ‘A’ (the rating of the inner CDOs in Umbria II and Como I) denotes:
“strong capacity to meet financial commitments, but somewhat susceptible to adverse economic conditions and changes in circumstances”.
A rating of ‘AA’ (the rating of the inner CDOs in Como II) is defined by Fitch as follows:
“AA: Very high credit quality
‘AA’ ratings denote expectations of very low default risk. They indicate very strong capacity for payment of financial commitments. This capacity is not significantly vulnerable to foreseeable events.”
Thus, these underlying assets of the CDO2s, albeit accounting for only 18% of the portfolio, also appeared themselves to have low or very low default risk.
Third, the structure was designed and described as having two levels of subordination, such that the inner CDOs and the outer CDO each had an attachment point which losses had to reach before the “AAA tranche” was affected. The fact sheets included on their first page a diagram which drew attention to the two levels of subordination from which it was said that the investor “benefits”.
Fourth, in the case of Como II, the tranche offered to CRSM was described in the fact sheet not merely as “AAA” but as “Senior AAA”, and the fact sheet also contained a table which showed a “AAA” tranche which was lower in the structure than the tranche which CRSM was buying.
Fifth, a key part of the information provided to CRSM in the fact sheets was the section headed “Analysis of Different Scenarios” (section 5). There were two elements to this analysis. The first part of the section set out in tables and diagrams how many defaults were needed on various assumptions (worst case, ‘intermediate’ case and best case) to erode (a) the subordination so that the tranche was affected and (b) the whole of the tranche. This part of the section, however, said nothing about how probable or otherwise it was that the specified number and distribution of defaults would occur. The information given concerning the probability of defaults occurring consisted of a diagram and text in the second part of the section.
Taking Umbria II as an example, the text, printed in bold, stated as follows:
“The AAA tranche of Umbria II is particularly robust with respect to defaults. It is possible to compare the default rate of the AAA tranche of Umbria II in the various scenarios with the historic 5-year default rates of a portfolio with a mean rating equivalent to that of the corporate exposures of the CSOs (Baa), and assuming in the worst case that only the corporate exposures are subject to default. The AAA tranche of Umbria II can withstand a default significantly higher than the worst default rate recorded in 5 years from 1940 to 1991 (five-year period 1986-1991). In scenario 1 “Worst Case”, the AAA tranche can withstand a default rate of 11.1% per CSO. This cushion is twice as high as the worst historical default rate recorded in the five year period 1986-1991. In scenario 2 “Intermediate Case”, the AAA tranche can withstand a default rate of 27.7% per CSO, which is 5 times higher than the worst historic default rate.”
CRSM submitted that this clearly conveyed that the risk of default of the “AAA tranche” was very low in that, in order for the tranche to be affected, the default rate over the 5 year period of the investment would have to be several times higher than the worst 5-year historical default rate recorded for entities with the average rating of the entities in the CDO portfolios (Baa) during the 51 years from 1940 to 1991. This message was said to be reinforced by the diagram which immediately preceded the text. This was entitled “Distribution of historical defaults on a cumulative basis over 5 years 1940-1991 (Moody’s) – default rate equal to (number of defaults) / (number of companies under observation) average rating of Baa*”. The footnote referenced by the asterisk after “Baa” stated:
“* Average rating Baa equal to average rating of the corporate exposures within the CSO of Umbria II.”
The diagram showed, in graphic form, the number of 5 year periods in which the default rate for companies rated Baa had reached different levels. The distribution was weighted heavily towards the left hand side of the diagram (the lowest default rates). Thus, in more than half (over 25) of the 51 5-year periods shown in the diagram, the default rate for Baa rated entities was less than 0.5%; and in only 2 of the 51 periods was it higher than 3.5%, with the highest recorded rate in any 5-year period being 5%. This distribution was then compared in the diagram with (i) the default rate (of 11.1%) per CSO required for a default of the AAA tranche to occur in the worst case scenario, and (ii) the default rate (of 27.7%) per CSO required for a default of the AAA tranche to occur in the intermediate scenario. Each of these rates of default is depicted by a vertical line on the diagram.
CRSM submitted that the clear purpose of the diagram was to represent graphically that, even in the worst case scenario, the default rate for the 5 year period of the CDO2 would have to be of wholly unprecedented severity in order for the CDO2 tranche to suffer a default.
Taking these features together, and in particular the comparative historical analysis of default rates, they would have been understood by any reasonable person in CRSM’s position to mean – as it was understood by CRSM to mean – that the Notes had (and were believed and expected by Barclays to have) a very low risk of default.
Barclays contended that there was simply no basis for implying the alleged representations from the factsheet or other documents. As to the fact sheets they contended:
The fact sheets described a proposed transaction that Barclays understood CRSM to wish to conclude. As such, it was not, by nature, a representation of any kind about anything, but an indicative statement of the terms of the transaction Barclays believed that CRSM wished to conclude.
It contained express qualifications/risk disclosures and disclaimers that prevented it from containing, or implying, any representation of any kind about anything, or at least prevented it from being read as making, or implying, any statement to the effect of any of the Purchase Representations. Barclays relied particularly in that regard on the text of the qualifications/risk disclosures entitled “Certain risk factors”, “Risk of Loss from Credit Events”, “Volatility”, “Highly Leveraged Investment”, “Credit Ratings”, “Default Rates”, “No reliance”, “Independent Review and Advice”, “Credit Risk”, “No Representations”, “Conflicts of interest”, and to the text of the “disclaimers”.
Although written in the present tense the fact sheets did not refer to anything then in existence. The “disclaimers” make that even clearer: “The Information contains a summary of certain proposed terms of a hypothetical offering of securities as currently contemplated in connection with preliminary discussions with potential investors and does not purport to be a complete description of all material terms or of the terms of an offering that may be finally consummated”.
If, contrary to (2) and (3) above, something stated in text of the fact sheets could amount to a representation, that text did not include any statement to the effect of any of the Purchase Representations. (Nothing that was stated in that text is alleged to have been inaccurate).
Nor, if capable of conveying representations at all, did the preceding text imply a statement to the effect of the Purchase Representations, or any of them. In particular:
the references to a AAA rating were no more than indications that Barclays was anticipating that any Umbria II note actually issued would be rated AAA, or, to put it the other way round, that Barclays believed the Umbria II note would not be issued unless the CRA rated it AAA;
even had Barclays been making a statement that an existing note was AAA rated, the idea that this, and no more, should carry with it an implied representation as to Barclays’ opinions as to the risks of default of the asset, is unsound; and is in any event precluded, in context, by the warnings and disclaimers here, and specifically “Credit Ratings” and “Credit Risk”;
specific scenario analyses were made of behaviour it was anticipated the Umbria II note, if issued, would display, upon stated hypotheses (about individual reference name defaults, or prevailing default rates), but no statement was made, or implied, as to how risky that would or would not make the note generally, and it was certainly not stated, or implied, that the Umbria II note, if issued, would be an investment with a very low risk of default.
There is a difference between the effect of defaults of individual reference names (which the scenario analyses considered) and the probability of defaults of such names, which the fact sheet did not consider, and which was a matter for CRSM to assess: as the “Credit Risk” warning made clear. Thus, although the fact sheets stated that Umbria II was “particularly robust to default”, particularising this by reference to details of the number of individual defaults that would be necessary to erode the subordination and comparing that to historical default rates, that is not an assertion as to the probability of such individual defaults (particularly when read together with the “Default Rates” risk warning). Mr Ferrario indeed explained to Messrs Montanari and Sapignoli in terms that “they would have to use their own judgment to assess how many names were likely to default”
In relation to references to AAA Barclays stressed that a statement that a CSO2 note had been rated AAA by one of the CRAs does not imply anything more than, or other than, that the note had been rated AAA by that agency. The CRAs’ definitions and methods were publicly available. The rating was a statement of the rating agencies’ expert opinion. Nothing more, or less, was implied by the rating. A statement that an instrument was so rated (or was expected to be so rated) was, in turn, merely a statement about what the rating agencies’ opinion was (or about what it was expected to be). Nothing more, or less, was implied by a statement that an instrument had been rated AAA.
A statement that a CSO2 note had been (or was expected to be) rated AAA by a CRA thus was not some general or abstract statement about default probabilities or risk. It was not a statement about market opinions concerning default probabilities or risk. It was certainly not a statement about a structuring or selling bank’s estimate of modelled (spread-implied) default probabilities or risk. Indeed, one of the key functions of credit ratings is, in arm’s length purchase transactions, to provide the objective opinion of a trusted third party on the prospects of non-performance of an obligation, in circumstances where a seller bank may well not wish to make any representation as to default probabilities or risk; and the purchaser bank will prefer to rely, in any event, on the ratings agency rather than the seller.
I find that no representations were made in the terms or to the effect of the Purchase Representations for the reasons given by Barclays and in particular the following:
A generalised statement that an instrument has a very low risk of default is an inherently unlikely statement to be made. As both parties would have appreciated, no one can know what the default risk of an instrument actually is. The most that can done is to estimate what that risk is, and if that is to be done then it is likely to be by reference to some stated criteria or yardstick rather than by way of a generalised or abstract statement.
It is inherently unlikely that a bank in the position of Barclays would make a generalised statement in relation to default risk, whether by way of opinion, estimation, expectation or belief. As CRSM well knew, and as would be expected, selling banks in Barclays’ position are generally extremely careful about what they say and about what reliance can be placed upon what may have been said. Regardless of the qualifications and disclaimers in the fact sheets themselves, CRSM knew from its general experience, and the earlier Rafal/Falco trade with Barclays, that banks in Barclays’ position present information and transact on the basis of wide ranging disclaimers, the general effect of which is to stress that it is for the investor to make its own assessment of the risks involved. The fact sheets themselves were clearly carefully drafted documents which were likely to have involved input from compliance and sales and which one would not expect to make generalised statements of the kind alleged.
No generalised statement about default risk is made in the fact sheets. The closest that one comes to such a statement is the statement that the tranche is “particularly robust with respect to defaults”. However, that statement is clearly made on the basis of the historical comparison which then follows. It is not a general or unqualified statement.
The reference to AAA in the fact sheets is as part of the description of the characteristics which the Notes are expected and indeed required to have. It is simply a reflection of the Notes’ required rating by the relevant rating agency.
The first four fact sheet features relied upon by CRSM are similarly descriptions of the characteristics which the Notes are expected to have. In so far as they involve any qualitative statement it is always by reference to a stated criteria (e.g. the stated rating).
The section headed Analysis of Diverse Scenarios does go beyond a description of the expected characteristics of the Notes and provides information relevant to an assessment of the risks involved, particularly in the comparative historical analysis. However, all the information provided was accurate.
For similar reasons no particular significance can be attached to Mr Ferrario’s references to the CDO2s as “AAA”s. It was a reflection of the Notes’ required rating by the relevant rating agency. In any event, this was simply a shorthand, as CRSM would have appreciated.
The key risk as far as CRSM was concerned was the default risk. If there was any risk about which one would expect the investor to have to make up his own mind it was the default risk. If there was any doubt about this, which there should not have been, it was a point expressly made by Mr Ferrario to Mr Sapignoli and Mr Montanari: “I went very carefully through the scenario analysis with CRSM, showing it that if a certain number of names (i.e. reference entities) defaulted, given recovery assumptions, the Notes would lose their entire value. I explained to Mr Montanari and Mr Sapignoli that they would have to use their own judgment to assess how many names were likely to default”. One would not reasonably expect Barclays to be making any statement about what CRSM had to decide for itself, even more so when this had been specifically emphasised to those involved.
The position is made even clearer by the fact sheet “Qualifications and Disclaimers”. These made it plain that CRSM was to be “responsible for its own independent appraisal” of “the risks in respect of the Notes” and that it may not rely on Barclays “in connection with” the “matters referred to in these Risks Factors”. The identified risks and risks factors included “default rates” of “assets comprising the Reference Portfolio” and the “credit risk” “associated with the Reference Entities”. Regardless of the detail of these terms CRSM would have known very well that it had to make its own evaluation and appraisal of the default risk. Further, before the first transaction Mr Ferrario went through the fact sheet qualifications and disclaimers with Mr Sapignoli and Mr Montanari.
The reasons given in the evidence of Mr Agresta and Mr Ferrario as to why they would not have made the representations alleged and why they did not understand that they had done so (see paragraphs 350-355; 392-396 below).
CRSM’s own oral evidence, which was inconsistent with any understanding that the Purchase Representations had been made. Thus it was Mr Montanari’s evidence that what he relied upon was the AAA rating, not what Barclays had said or was understood to have impliedly said. Similarly, Mr Micheloni’s evidence was that he did not form any view from the fact sheets about how probable it was that there would be defaults; if so, he cannot have understood Barclays to be making thereby any representation as to that.
I accordingly find that the Purchase Representations were not made.
Alternatively, if any representations were made as alleged I find that they involved no representation of fact. I find that there is no such thing as an “actual” default risk. Default risk is not a matter of fact; it is a matter of estimation. There are various ways in which that estimation may be made and various different yardsticks which may used for that purpose. It follows that if any representation was made it was one of opinion and/or expectation and/or belief. I am prepared to assume that if it was a representation of opinion it is to be inferred that it was made on reasonable grounds.
The Restructuring Claims
In Opening four representations were alleged to have been made by various documents, at a meeting in San Marino on 22 April 2005 where Mr Ferrario discussed the first (in time) of those with CRSM and/or at lunch meetings between Mr Ferrario and Mr Morolli. These alleged “Restructuring Representations” were that:
The substitutions (i.e. individual reference name substitutions) “had, and were intended and believed by Barclays to have, the effect of improving the quality (i.e. reducing the risk) of the reference portfolios of the CDO2s”;
Credit ratings and Barclays’ “Scaramanga” scores “were regarded by Barclays as appropriate objective criteria which Barclays themselves had used to assess the credit risk of the entities proposed for substitution”;
The addition of cross-subordination and other changes to the structure of the CSO2s “also had the effect, objectively and in Barclays’ belief, of reducing the level of default risk of the CDO2s” ;
the restructuring “was being proposed for CRSM’s benefit and the substitutions would be made at “zero cost” such that Barclays would not make any profit from the restructuring”.
By the end of the trial CRSM’s case had narrowed and to an extent shifted. Their case had become as follows (“the Restructuring Representations”):
Misrepresentation of intention to profit - CRSM was told that the restructuring would be carried out at “cost zero” – that is, on the basis that Barclays would be left no worse off but also would not profit from the restructuring, and that the only cost to CRSM would be the bid/offer spread incurred on substitutions. In fact, as Mr Agresta and Mr Ferrario knew, Barclays did intend and expect to make a profit from the restructuring.
Misrepresentation of the Selection Criteria - In formulating Barclays’ proposals for substitutions to the CDO portfolios Mr Agresta (assisted by Mr Bechet) had pursued three objectives: he had sought to find names to substitute which, in general, had (1) higher credit ratings, (2) higher ratings according to Mr Scaramanga’s model, and (3) higher spreads for their ratings than the names which were replaced (leading, overall, to higher expected loss). By stating that the first two criteria had been followed and omitting reference to the third, the emails and memoranda which described the proposals were deceptive.
Misrepresentation relating to Scaramanga model - Despite being informed of its limitations, Mr Agresta and Mr Ferrario chose to present ratings from Mr Scaramanga’s model to CRSM as an “objective” measure of portfolio improvement and did so in a way which ignored the limitations of the model and was actively misleading.
Misrepresentation of intention to profit
In Barclays’ email of 22 April 2005 which I accept was discussed at the meeting that day it was said that the substitutions may be obtained at “cost “zero””. CRSM submitted that that meant with no profit to Barclays.
CRSM further submitted that this was borne out by the oral evidence.
It was Mr Morolli’s evidence that there was a conversation with Mr Ferrario over lunch at the Ristorante dei Lavoratori in San Marino in which Mr Ferrario said that the restructuring would be “cost zero” for CRSM. Mr Morolli recalled that in this conversation he expressly asked Mr Ferrario whether Barclays would make a profit out of the restructuring and Mr Ferrario answered ‘no’.
Although he did not remember when the meeting in question was, Mr Buoncompagni remembered asking Mr Ferrario whether there would be any cost to CRSM of the restructuring and that Mr Ferrario replied that the only cost to CRSM would be the bid/offer spread of the substitutions. Mr Micheloni had no recollection of any discussion about Barclays’ profits or of bid-offer costs.
Mr Ferrario denied that he had ever said that Barclays would not make any profit out of the transaction. Consistently with the evidence of Mr Buoncompagni, it was his evidence that he had explained at a meeting that CRSM would have to pay the bid/offer spread on the substitutions.
CRSM submitted that what Mr Ferrario was thereby saying was that if there were substitutions there would have to be compensating changes for Barclays. In other words that the net effect for Barclays would be neutral, which meant no profits being made. It further submitted that this was consistent with Mr Ferrario’s use of the words “zero cost” in his email of 27 April 2005 and some of his answers in cross examination.
It was Barclays’ case that “cost zero” simply meant that no cash payment would be required to be made. It said nothing about whether and if so what profits would be made by Barclays, nor was anything ever said to the effect that they would not. On the contrary it was implicit in the admitted statement that there would be a bid/offer cost to CRSM that Barclays would be making profits – the spread between the “bid” and the “offer” being the bank’s (gross) profit on mark to mark valuations.
I find that the only statements made relating to this matter are the references to “cost “zero””/”zero cost” in the emails of 22 and 27 April 2005 respectively and Mr Ferrario’s explanation at the 22 April 2005 meeting that CRSM would have to pay the bid/offer spread on the substitutions. Although I accept that there was a lunch meeting on 22 April 2005 I am not satisfied that anything was said at that or at any other meeting to the effect that Barclays would not make a profit out of the transaction. In oral evidence Mr Morolli sought to expand upon the position set out in his first witness statement and referred to a number of occasions on which Mr Ferrario allegedly assured him that Barclays would make no profit; he then put it somewhat differently in re-examination when he said that this was to be implied from Mr Ferrario telling him that the restructuring would have been carried out at no cost to CRSM. I do not accept his evidence on the point and find that the only reference to the issue was on the three occasions set out above.
I further find that no representation was made that Barclays would be left no worse off or would not profit from the restructuring. In particular:
The natural meaning of the expression “cost zero” is that it is addressing the costs to CRSM. Whilst there may be a link between costs to CRSM and profit to Barclays they are not the same thing.
The expression was used against the background of a discussion which had earlier taken place between Mr Ferrario and Mr Buoncompagni at which the possibility of CRSM making a payment of money to make substitutions or improvements had been raised.
That “cost zero” would be reasonably understood as referring to there being no cash payment required is borne out by Mr Ferrario’s email of 27 April 2005 which refers to the possibility of effecting changes at “zero cost” in contradistinction to CRSM paying cash (c.3.5% of notional to remove GM entirely). In cross examination Mr Micheloni accepted that “zero cost” there meant that CRSM did not have to pay cash out.
It is implicit in the admitted fact that CRSM were told that there would be a bid/offer spread cost for CRSM that there might be some profit for Barclays.
It was inherently unlikely that Barclays would say that they would not make any profit given the extensive amount of work that was potentially involved and the fact that even if this was the intention there could be no guarantee that it would be achieved. The changes involved in making substitutions were bound to involve changes to the MTM position. Precisely what the effect would be could not be known until the changes were actually made but the possibility of an improved MTM position for Barclays was always there.
Even if a representation to the effect alleged had been made in April 2005 I find that it was not continuing and/or that it would not have been reasonable to rely on it by the time the contract was concluded in June 2005. No mention of this matter had been made in any document or discussion after 27 April 2005. It was not included in any of the deal memoranda presented or in any of the transaction documentation. There had been substantial changes to what was proposed in the meantime as well as known changes in market conditions. Further, before the contract was concluded CRSM had the credit spread information relating to the changes and could see for themselves the potential for profit for Barclays through widening spreads. The qualifications and disclaimers set out in the memoranda and the Restructuring Letter provide a further reason why a long antecedent representation of this kind could not be relied upon as a continuing representation.
Misrepresentation of the Selection Criteria
As formulated in closing CRSM’s case was that the evidence demonstrated that in formulating Barclays’ proposals for substitutions to the CDO portfolios Mr Agresta (assisted by Mr Bechet) had pursued an objective of higher spreads for their ratings than the names which were replaced. By stating that the two identified criteria had been followed and omitting reference to the third and allegedly real criterion, the emails and memoranda which described the proposals were deceptive.
CRSM relied by analogy on share prospectus cases such as Oakes v Turquand. It submitted that this was one of those cases where a statement which is literally true involves a misrepresentation because of matters which the representor omits to mention.
At times it appeared, and was so understood by Barclays, that it was CRSM’s case that a representation was made that higher credit ratings and higher Scaramanga ratings were the only selection criteria being used by Barclays. I did not understand that to be CRSM’s case, at least by the end of the trial. If that was its case then I find that no such representation was made. Nothing was said about these being the only criteria used. In any event it would have been apparent to CRSM from the known widening of credit spreads that some other criteria was being used.
It is the case that Barclays represented that higher credit ratings and higher Scaramanga ratings were criteria that would be and had been used in relation to portfolio selection. Whether that involved deception or was rendered untrue by what is alleged to have been the real criterion used of higher spread for ratings depends on the case on fraud and will be considered under that heading.
Misrepresentation relating to Scaramanga model
It was CRSM’s case that Mr Agresta and Mr Ferrario chose to present ratings from Mr Scaramanga’s model to CRSM as an appropriate “objective” measure of portfolio improvement and did so in a way which ignored the limitations of the model and was actively misleading. In support of the alleged representation that Barclays had represented the model to be an appropriate “objective” measure of portfolio improvement CRSM relied in particular on the following:
the email dated 22 April 2005 which stated that “[f]or the substitution proposal we have followed 2 ‘objective’ criteria” – one of these being:
“rating of Barclays Capital Research (according to the Scaramagna model).”
The email set out tables comparing the distribution of Scaramanga ratings in the “current” and “new” portfolios of Como I and Como II, with, it was said, the object of showing that the portfolios were being improved.
The email of 12 May 2005, which extended Barclays’ restructuring proposal to include Umbria, identified “improvement in the quality of the portfolio on the basis of ... the rating of Scaramagna’s quantitative model” as one of the objectives of the restructuring, and stated:
“As far as Scaramagna’s model is concerned, we mainly reduced the rating of names rated -2 and increased the weighting of credits rated +2.”
The email again set out, for each of the CDO2s, tables showing that “the new portfolios have a considerably better quality than the existing ones in terms of Scaramanga ratings”.
The June memoranda, including the final version dated 13 June 2005, presented as one of the objectives of the restructuring “improvement in the quality of the portfolio in conformity with ... the rating based on the quantitative model of Scaramanga” and went on to state:
“Details are provided below of the improvement of the portfolio. The new portfolios display a significantly better quality than the existing ones, both of the basis of their official rating ... and calculated on the basis of the model developed by our analyst, Scaramanga. With regard to the Scaramanga model, we have acted above all on names with ratings of -2, reducing their weight, and on credits with ratings of +2, increasing their weight.”
As the 22 April 2005 email makes clear the Scaramanga model was initially put forward as a “proposal” by Barclays. A description of the model was set out in a summary document Quantitative Credit Relative Value Indicator (“the QRV presentation”) which Mr Scaramanga had prepared. Mr Ferrario went through this with CRSM. It was not suggested that there was anything inaccurate in the QRV presentation or Mr Ferrario’s explanation of it. Having had that presentation it was up to CRSM to decide whether they wanted to proceed with the model as one of the selection criterion for the restructuring. It was content for it to be so. Thereafter it was described in the documents as one of the objectives of the restructuring and also as being an improvement.
The model had been described in the 22 April email as an “objective” criteria. Once it was agreed that it would be used for the purpose of the restructuring it was referred to as being an improvement in quality. However, that was not a generalised statement. It was simply a statement that measured by reference to that agreed criterion the portfolio had been improved. The appropriateness of the use of the criteria and its qualitative effect was a matter for CRSM, as the memoranda and Restructuring Letter qualifications and disclaimers served to emphasise.
Against the above background I find that no representation was made that the Scaramanga model was an “appropriate” criterion of portfolio selection or that it involved any improvement in the portfolio other than as reflected in the Scaramanga scores. It was represented as being an “objective” criterion, but this was true. Whether its proposal and use nevertheless involved some deception will be considered in the context of the fraud case.
VI. Fraud/Negligence
CRSM has to prove that any representations made were fraudulent for its claim under the tort of deceit. As CRSM accepts, that means showing that either or both Mr Ferrario or Mr Agresta were fraudulent: either or both must individually have the necessary subjective understanding and intentions for fraud. It is not possible for Barclays to be held liable for fraud by combining the understandings of individually innocent people.
For the claim under s.2(1) of the Act, it is formally for Barclays to prove that the relevant individuals had an honest and reasonable belief in the truth of the representations they were making.
The technical issues
At the heart of many of the issues between the parties lay a major dispute as to whether an implied probability of default derived from EL figures taken from the output of a pricing model such as that used by Barclays provides a reliable measure of “real world” probability of default.
It was CRSM’s case that it is the most reliable measure, that Barclays knew this, that Barclays knew that it showed that the Notes had a very substantial risk of default, and that Barclays also knew that this was contrary to the impression given by the Notes’ “AAA” rating and other statements made that the investment had a very low risk of default.
It was Barclays’ case that this was misconceived. EL figures derived from a pricing model do not provide a real world measure of probability of default. It was well known that they would produce figures of an order of magnitude greater than the long term average default rate based on historical default rates (at least for higher grade credits). The two are simply incomparable. There was therefore no inconsistency between high projected EL figures and an AAA rating. On any basis this was a view which Barclays, and in particular Mr Agresta and Mr Ferrario, could reasonably and honestly hold at the material time.
The pricing model used by Barclays and other banks at the time was a similar statistical model to that used by the CRAs. The major difference between them was the inputs used. Banks would use information derived from the credit spreads for the reference entities. The CRAs would use historical default rates for the ratings ascribed to the reference entities. These produce (at least for higher grade credits) an order of magnitude difference in the probability of default derived therefrom with the consequence that the model will produce an output with an order of magnitude difference between the probability of default implied therefrom and that implied from the rating derived from the CRA’s rating model.
CRSM submitted that credit spreads are and were generally regarded by Barclays itself, as well as more widely by market professionals and by the experts in this case, as the best measure of default risk. They stressed the following points in particular:
Barclays used credit spreads as inputs to its model to measure the probabilities of default of the CDO reference entities. There is no evidence that Barclays made any attempt to adjust these figures for any supposed non-default component of the spreads when using the model for any of the various purposes for which it was used.
As Mr Nasr pointed out, Barclays could not have used data from its pricing model, derived from spreads, for the calculation of its profits had it not regarded them as an accurate measure of default risk. The net profits so calculated must have been regarded as sufficiently certain to meet the accounting standards of revenue recognition, and therefore cannot have been regarded as subject to large uncertainty.
There is no disclosed documentation or other evidence of any contemporaneous discussion or suggestion within Barclays that credit spreads were an inaccurate or inadequate measure of default risk.
Similar models which used credit spreads to measure default risk were used at the relevant time not only by other major banks as well as Barclays but by some sophisticated investors. Dr Ellis agreed that the use of credit spreads to measure default risk was standard market practice and best valuation practice at the time. There is no evidence that any of these entities, any more than Barclays, ever made any attempt to adjust their assessments of default risk on account of any supposed non-default component of credit spreads, or considered that it might be necessary or appropriate to do so.
The experts in this case used credit spreads to measure the probabilities of default of the CDO reference entities in valuing and assessing the risks of the CDO2s. Neither expert sought to make any adjustment to any of their calculations on account of any non-default component of credit spreads.
Barclays took fundamental issue with CRSM on this question. It stressed the following in particular:
A single name CDS spread is a price at which a buyer of credit protection buys protection on that name for a given period by way of periodic premium payable to the protection seller under a CDS referencing that name. It reflects the market at a point in time. It is not, in itself, a prediction.
Single name CDS spreads vary substantially from name to name, and for different protection periods. They vary substantially from name to name within any given rating, although average spreads will increase as ratings decrease (i.e. the average credit spread for AAA names will be smaller than the average spread for AA names; the average credit spread for BBB+ names will be smaller than that for BBB names, which will in turn be smaller than that for BBB- names; and so on).
Single name CDS spreads exhibit significant volatility unrelated to change in underlying credit-worthiness. They also vary over time, and by geography, industry, supply and demand, and other considerations irrespective of whether those factors have any real impact on underlying credit-worthiness. In respect of any particular spread, the extent to which that spread reflects perceptions of credit-worthiness, and the extent to which it reflects other matters, is unknowable.
The default component of a price is the portion of the price that can be said or assessed to be referable to the default risk in the instrument in question, i.e. the portion of the price that is being charged for taking on that risk rather than for taking on other types of risk or for other reasons not related to the default risk. It does not measure or quantify the default risk.
Suppose by way of example a name attracting a 5-year CDS spread of 100bp. If a 40% recovery rate in case of default is assumed, and using Mr Agresta’s slight refinement on Mr Nasr’s formula for calculating an implied probability of default from credit spreads a probability of 7.5% can be calculated: (1 x 4.5) / 0.6 = 7.5. However, that does not mean that there is (or the market thinks there is), in the real world, a 7.5% chance of that name defaulting during the next 5 years (or anything like it). (In 2004/2005) a single name spread of 100bp would look, stating it very broadly, like a BBB- spread. The market would not be expecting anything like 1 in 12 BBB- names to default over 5 years. The 100bp spread means that the market is paying, for taking on the 5-year CDS risk of the name in question, a price that a mathematician would say was equivalent to there being a 7.5% chance of paying out 60% of the nominal of the trade at some point during the period of the trade. But that is not the same thing at all.
Similarly, pursuant to the CSO2 notes, CRSM was paid c.100bp over a risk-free rate for taking on the single-tranche portfolio credit risk created thereby for the relevant period, in line with market prices for that type of risk. That does not mean that CRSM thought, or the market thought, or Barclays thought, that there was c.7.5% chance of a default on the CSO2 notes. That was, however, the market price at the time for taking on the (very small) risk of (very large) loss that an investment in the CSO2 notes represented.
An implied probability of default for the CSO2 tranche calculated from an EL figure calculated by, or with the help of, a structuring banks’ correlation pricing model, is conceptually similar. Just as the implied probability of default for a single name CDS, or a bond, does not measure the real world chance of the reference name, or issuer, defaulting, and may be an order or orders of magnitude greater than the default rate experienced by debt of the rating in question, so the implied probability of default for the CSO2 tranche did not measure the real world chance of the tranche defaulting and was an order or orders of magnitude greater than the default rate historically experienced for AAA-rated debt. Just as the premium on a single name CDS would not be expected to cover the loss incurred if the name defaulted, so for the CSO2 tranche the coupon (even though very generous for a AAA-rated instrument) would not be expected to cover the loss that would be incurred if the CSO2 tranche defaulted.
Why the premium required for taking on credit risk is (proportionately) so large is the “credit spread puzzle” referred to by various analysts and commentators.The credit spread puzzle is a known phenomenon that exists because implied probabilities of default, calculated from CDS credit spreads, are not, and are not perceived by anyone in the market as, measuring (quantifying) the prospect that names will default.
Barclays’ hedging/pricing model used single name CDS prices as a key input because the single name CDS market was the primary market in which Barclays’ delta hedge would be calculated and managed with the help of the model. It had essentially nothing to do with the market price for the CSO2 notes expressed as the yield that the notes might be expected to offer the investor.
Because implied probabilities of default derived from single name CDS prices were used in Barclays’ hedging/pricing model as default probabilities for the individual reference names in the CSO2 portfolios, the credit spread puzzle was built into the pricing and profit calculation methodology used by Barclays for its purposes. An “Expected Loss” or implied probability of default derived from the model’s delta hedging calculations was thus bound to be out of proportion to the actual chances of the CSO2s defaulting. At all events, a large disproportion between such an implied probability of default and historical default rates for AAA-rated debt was (i) nothing surprising or untoward, and in particular (ii) not a reason to doubt the validity or reliability of the credit rating agency’s expert assessment that the CSO2s qualified for its AAA rating.
Having carefully considered the expert and factual evidence on this issue, I find on the evidence before the court that an implied probability of default derived from EL figures taken from the output of a pricing model such as that used by Barclays does not provide a reliable measure of the real world probability of default, and that on any view that was a reasonable view for banks such as Barclays to take at the time. I essentially accept the various reasons advanced by Barclays for reaching this conclusion. In particular:
The CRAs, being the industry experts in credit risk analysis, regarded historically derived default data as the most reliable information to be input into their rating models. They regarded the output as being appropriate, reliable and fit for purpose.
Historical default data input is based on actual experience over long periods reflecting all phases of the business cycle and may be regarded as a reliable source for estimating expected future default rates in uncertain economic conditions.
It was chosen as the appropriate input by the CRAs for rating purposes – i.e. for making a long term risk assessment.
By contrast, input data using credit spreads reflects the market price for credit protection at a particular moment in time. It will fluctuate considerably between names and over time for reasons which do not reflect differences in underlying credit-worthiness.
It was chosen as the appropriate input by banks for pricing and hedging purposes rather than for rating purposes.
CRAs were aware of the modelling carried out by banks, that it used implied default probabilities derived from current CDS market credit spreads, and that arranging banks used such modelling to hedge/manage the banks’ risk, book MTM valuations and to assess P&L. It was the CRAs’ expert view that structuring banks’ modelled EL did not undermine their AAA rating opinions
The order of magnitude difference between the default risk derived from the two different inputs means that they cannot both be reliable measures of default risk. I find that if one was seeking to provide a quantifiable measure of real world default risk, it is the historical input data which is the more reliable measure. Credit spread implied default probabilities reflect perceived default risk but do not lead to a reliable quantitative assessment of that risk.
That modelled EL figures derived from credit spreads do not measure real world default risks was a view shared by Barclays (Mr Agresta and Mr Ferrario), other banks (Ms Duhon) and experts (Dr Ellis).
Mr Nasr suggested that the credit spread puzzle might be explained as the CDS market somehow pricing in the (tiny, but not zero) chance that at any given time, a once in 80+ years market event or correction could occur within the relevant time period. However, as Barclays pointed out, if that were so it would in fact be consistent with the basic point it makes. It would indicate that a CDS price is much larger than the spread that would be equivalent, as a matter of arithmetic, to the actual probability of a default occurring because that is what it takes to persuade people to take on even a vanishingly small risk of losing a lot of money, not because that actual probability is measured by, or is perceived as being measured by, or as being even close to, the implied probability of default that can be calculated from the CDS price.
That it was reasonable at the time for banks such as Barclays to take the view that modelled EL figures did not reflect real world probabilities of default was borne out by the evidence of Ms Duhon. Although CRSM suggested that she had little relevant experience I accept that she was in a position to provide valuable evidence of the perceptions and practices of reasonable bankers at the material time. Her largely unchallenged evidence was that:
“11.1 Importantly, single tranche structurers and traders weren’t credit analysts nor were we credit investors. We simply used market inputs as we knew them to calculate market to market and to assess our hedging strategy …
11.2 In terms of my own views at the time, I remember working on some CSO2 trades in 2004 and 2005 and thinking that the AAA was “bullet proof” …
11.3. Also, we never assumed that the entire securitisation market had gotten it wrong and we had gotten it right …
11.4 … we did not consider the model price to be an accurate indicator of expected loss or gain on a single tranche …
12.3 … in my experience we did not consider the MTM anything other than potentially a way of taking advantage of the arbitrage in that if we were clever in our hedges we could make more money out of the trade … we didn’t consider the rating methodology as inconsistent with our models …
12.4 Thus, if … on day 1 of a AAA rated CSO2 tranche a bank had a gross profit of 20-30% (before the adjustments I mention … below) that did not mean, based on my experience, the bank somehow perceived the risk on the tranche to be inconsistent with the rating. The two things were entirely different.”
Although Mr Nasr took a different view, he confirmed in evidence that the view that AAA CSO2 trades were “bullet proof” was one shared by competent, experienced and reasonable market participants at the time. As Barclays stressed, Mr Nasr also accepted in evidence that, absent market catastrophe, default rates based on historic performance by rating are accurate predictors, and that a AAA rating can be justified and sound, and a very low historic probability of default can be experienced and expected, even though the spread-implied probability of default is greater by one or two orders of magnitude.
In relation to the points stressed by CRSM as summarised above I find as follows.
As to (1), it was the evidence of Barclays’ witnesses, which I accept, that the main reason that single name credit spreads were used as inputs to the model was because the single name CDS market was the primary market in which Barclays’ delta hedge would be calculated and managed with the help of the model. The primary purpose of the model was to calculate the hedges, to mark to market and to assess potential P&L; it was not used to measure long term probabilities of default. As Ms Duhon explained, in her experience the term “EL modelling” was never used. The model was referred to as a “pricing model” not a “rating model”. That was a reflection of its primary function. One would not therefore have expected Barclays to make adjustments to credit spread figures for non-default elements since this was not the rationale of their use. Further, even if Barclays had given consideration to this, the expert evidence showed that there is a considerable debate as to how much of a credit spread represents the default element of the price. Even if the answer had been known it still would not have produced a reliable measure or quantification of the probability of default for the reasons already stated. This is a further reason why there would have been no reason for Barclays to make adjustment for this, even if it had been known how to do so.
As to (2), it is correct that Barclays must have regarded the output of their pricing model as producing a sufficiently reliable value to justify being reflected in their accounts. However, it was the valuation figure that Barclays was focussing on, not the EL figure that might be derived from it. CRSM is correct that the valuation method worked by reference to the probability that the investor would not be paid back the price at the end of the relevant period, but it was the valuation figure that banks such as Barclays were interested in for hedging, MTM and P&L purposes. It was not viewing or using the model as an estimation of long term default risk.
As to (3), this may be correct, but it simply reflects the fact that Barclays did not regard credit spreads as providing a quantifiable measure of default risk. There was therefore no particular need for any such discussion.
As to (4), as the evidence of Ms Duhon confirmed, other banks, like Barclays, were using these models for hedging, MTM and P&L purposes; not to measure long term default risk. Further, the evidence was that different models used different methodologies and produced different results, thus highlighting the variability of any derived EL figure. Although Dr Ellis agreed that the use of such models was market practice and best valuation practice at the time he did not consider that the EL figure derived therefrom reflected ““real world” probabilities sometimes called “physical probabilities””.
As to (5), this is correct. However, it is not clear what adjustment could be made and, even if that was known, it would not, on Dr Ellis’s evidence, reflect the “real world” probabilities of default.
CRSM contended that the difference between default risk derived from historical default rates and credit spreads could not explain the very large EL figures of 20-30% allegedly shown by Barclays pricing model in this case. It contended that by far the largest part of the difference reflected credit ratings arbitrage on the part of Barclays. By this it meant not simply taking advantage of the difference between the price reflective of the default risk implied by CRA ratings and that reflective of the default risk implied from credit spreads, but accentuating that difference through the selection of reference entities with high spreads for their ratings. It suggested that it was this practice which explained the great majority of the difference.
It submitted that even if credit spreads do not provide a reliable measure of default risk they do reflect default risk. That being so the very high EL figures shown by the pricing model showed, as Barclays must have realised, that the default risk cannot have been at the minimal levels suggested by the Notes AAA rating. This will be considered further when addressing the evidence of the Barclays’ witnesses, but some preliminary observations may be made:
If the credit spread derived figures of default risk are not regarded as being comparable to historically derived figures, nor as providing a reliable measure of default risk, then taking full advantage of the resulting price difference does not necessarily change that. It may be regarded as an extension of the same point. If the arbitrage itself is reasonable and justifiable, so may be taking full advantage of the arbitrage opportunity provided.
To put the matter another way, if the implied probability of default in the reference portfolio calculated from single name CDS prices does not mean that the real world chances of default are different to those indicated by ratings, the fact that the reference portfolios were largely populated by names paying higher-than-average spreads (for their ratings) similarly did not mean that the real world chances of default were different to those indicated by the ratings. The contrary suggestion assumes both that implied probabilities of default derived from single name CDS prices measure real world default prospects and that today’s higher-than-average spreads are tomorrow’s higher-than-average spreads.
Before turning to the factual evidence I should make findings as to what are the relevant EL figures for the Notes.
I consider that the most relevant figures to use are those calculated by Barclays at the time. They were the figures (or approximations of them) which the witnesses would have been aware of and they were similar to those calculated by the experts. These figures should then be adjusted to take into account skew valuation adjustment, for the reasons given by Dr Ellis. On the basis of his Table 18 that produces the following EL figures:
Umbria II: 20.26%;
Como I: 15.81%;
Como II (Bellagio): 19.55%;
Como II (Tremezzo): 23.74%.
If these are the relevant EL figures then the probability of default figures to be derived therefrom are likely to be higher and so these can be treated as minimum figures for probability of default.
Assuming that a comparison with the probability of default derived from ratings is appropriate then even on these figures the difference is marked. According to all the ratings agencies’ figures the 5-year default rates for AAA assets were well below 0.5%, and often as low as 0.1%, whether for corporate issuers, sovereigns or structured finance securities. An order of magnitude difference between the figures would be explained by the credit spread puzzle. The remainder (which may be high in percentage as opposed to multiple terms) would be explained by Barclays’ use of high spread names for their ratings.
CRSM’s general case
CRSM’s general case was that the way in which Barclays was able to structure each of the CDO2s so as to obtain a AAA rating for a product but with a high expected profit/loss was through what has been described as ‘credit ratings arbitrage’. The ‘arbitrage’ involved taking advantage of two aspects of the method by which the rating agencies modelled the probability of default of the CDO2s:
First, the fact that, in estimating the probabilities of default of the CDO reference entities, the rating agencies used historical data rather than credit spreads; and
Second, the fact that the rating agencies used as a measure of the probability of default of each individual reference entity statistics which related not to the particular entity but to all entities which shared its rating; in other words, there was no selection process within each rating category.
The existence of a range of entities within each ratings category at any given time with widely different probabilities of default as reflected in their credit spreads allowed Barclays to increase its profits, without affecting the credit rating of the CDO2s. This was possible because the rating agency’s assessment took no account of differences in risk between the particular entities included in the reference portfolios and other entities with the same credit rating. It looked only at the data for all entities in the given rating category and treated them all as if they were of identical risk.
The evidence established that this is what Barclays did. It deliberately selected entities for inclusion in the reference portfolios of the CDO2s which had high spreads for their ratings. This allowed Barclays to increase its profits, whilst still obtaining a AAA rating for the CDO2s, because the rating agencies took no account of differences in risk between reference entities with the same credit rating (e.g. BBB+) and estimated the probability of default of each such entity by reference to the default rate for all entities with that credit rating.
Both the fact that Barclays chose for inclusion in the reference portfolios of the CDO2s names which had high spreads for their ratings and the extent to which they did so was evidenced by statistical analyses carried out by both experts.
It is not conceivable that the concentration found by the experts of names with high spreads for their ratings could have occurred other than by design. The results of the experts’ analyses are only explicable on the basis of a deliberate policy in structuring the CDO2s to include in their reference portfolios names which had high spreads for their ratings.
That there was such a policy is further borne out by Barclays’ internal documents. There are more internal emails which evidence the arbitrage employed by Barclays in the restructuring of the CDO2s than in the case of their initial structuring but CRSM relied on two emails sent during the structuring of, respectively, Umbria II and Como I in particular.
The first is the email sent by Mr Agresta to his colleague, Mr Monkkonen, on 21 May 2004 during the process of structuring Umbria II. The subject of the email is “San Marino universe”. Mr Agresta wrote:
“● pool is 95% the final pool - we have room to manoeuvre if rating/spreads changes
● we are still looking for efficient names! Highly rated names with high spreads or in rare industries (recent example: Centerpoint Energy Inc ...)”
CRSM submitted that this email provides confirmation that, in constructing the reference portfolios, Mr Agresta was deliberately looking for names which not only had high ratings but also had high spreads, and were therefore “efficient” in making greater profits for Barclays, whilst at the same time enabling Barclays to secure a AAA rating for the CDO2.
The second is Mr Monkkonen’s email of 14 October 2004, sent in response to Mr Agresta’s enquiry about how best to try to make more profit on Como I than on Umbria II despite the fact that spreads were generally lower. Mr Monkkonen’s commented:
“The more we arb this the more likely it is that we have downgrades/defaults before next years anticipated 90m increase.
CRSM submitted that it is clear that the point which Mr Monkkonen was making was that, the more Barclays “arbitraged” Como I by including in its portfolios names with high spreads for their ratings, the greater the risk of default and the more likely it was that defaults would occur even before CRSM had agreed to purchase the remainder of the planned series of Notes.
CRSM contended that in the light of the documentary and expert evidence, Mr Agresta could hardly have denied that the inclusion of names with high spreads for their ratings was the result of a deliberate strategy, nor did he do so. In particular:
He accepted that that as a structurer, one way he was required to make the transaction profitable for Barclays was by finding names to include in the reference portfolios which had high spreads for their ratings. This was “one of the several constraints when we were building the portfolio”.
He agreed that the selection of names with spreads in the top quartiles of spreads for their ratings was not random, and that “[a] lot of the names were in the higher spectrum”,whereas if the names in the portfolio had been randomly selected one would expect roughly half to have probabilities of default below the median and roughly half above it.
He agreed that one way in which Barclays was able to generate profit from the transactions with CRSM was by putting into the portfolios names which had high spreads for their ratings and that this was “one component of the process”.
He agreed that Barclays’ pricing model was used to “tell us how big the arbitrage is” and that, in the case of these CDO2s, “the pricing model was telling us” that the arbitrage was of the order of 300 or 400 basis points higher than the 100 basis points paid to CRSM, and that the expected losses calculated by the model were higher than those implied by the rating agency.
With reference to Mr Monkkonen’s email of 12 October 2004, he explained that there was an “arbitrage spectrum” and that every portfolio was positioned somewhere in this spectrum, though he claimed that in this case “it was never our intention to position the portfolio at the very end of the extreme, towards the maximum level of profitability”.
He confirmed that Barclays’ management wanted the P&L for the subsequent CDOs to be around the same as for Umbria II, and that it was part of his job as a structurer to try to arrange the structure to help Barclays achieve that level of profit.
CRSM stressed that this adverse selection greatly increased the risks of default of the CDO2s. The way in which it did this involved two further important features of the CDO2 structure as designed by Barclays.
The first was the amount of leverage built into the CDO2s. The leverage was created by the fact that each of the CDO2s contained within it six inner CDOs, each of which had a reference portfolio which was not much smaller in size than the notional amount of the CDO2 itself. It was for mainly this reason that Barclays’ hedges needed to be much larger than the notional amount of the tranche sold to CRSM. Moreover, the fact that CRSM was exposed to the risk of defaults in such a large portfolio meant that any increase in the average default probability of the portfolio was multiplied several times over in its effect on the default probability of the CDO2 itself. Dr Ellis agreed that the leveraging effects of the structure meant that a small increase in the default probabilities of the reference entities had a much bigger effect on the default probability of the CDO2 itself.
The second significant feature of the structure was that, as Dr Ellis observed, the relationship between increasing the probability of default of the reference entities and the effect on the probability of default of the CDO2 itself was not linear. This had the result, as Dr Ellis agreed, that “once you start getting defaults, the loss rate accelerates quite rapidly”, and that “once you can increase the average probability of default of the [reference] assets beyond a certain point, the overall probability of default of the CDO-squared ... takes off”.
CRSM placed particular reliance upon a passage in his evidence, in which Dr Ellis agreed that the difference between the probability of default implied by the spread of 1% above Euribor paid on the CDO2s and the much higher probabilities of default which he had calculated (as reflected, for example, in his breakeven spread figure of around 6%) cannot be accounted for by the difference between “real-world” and spread-based calculations; rather, it is due to the adverse selection of the reference assets in combination with the structure of the CDO2s.
It was CSRM’s case that the effect of the adverse selection, in combination with the structural features of the CDO2, was thus to make the AAA ratings of the CDO2s misleading. The AAA rating was in each case based on the historical default rate of all assets within each relevant rating band, whereas Barclays had put into the CDO portfolios assets with probabilities of default significantly higher than the average for their ratings. The leveraged and tranched structure of the CDO2s magnified the effect of this arbitrage strategy.
The adverse selection also meant that, for the same reason, the information given in Barclays’ fact sheets about the probability of default of the CDO2s was misleading. The analyses in the fact sheets compared the number of defaults in the portfolio needed to cause a loss to the CDO2 tranche with historical rates of default for companies with the same average rating as those in the portfolio; this comparison was then used to imply that wholly unprecedented events would have to occur in order for the CDO2 tranche to suffer loss. However, the comparison was misleading in circumstances where the entities included in the portfolio were much riskier than average for their ratings.
For all these reasons CRSM contended that the deliberate heavy weighting of the portfolios by Barclays in favour of names which had high spreads for their ratings had a massive distorting effect on the risk of default of the CDO2s, and made Barclays’ representations false.
Barclays’ answer to CRSM’s general case was essentially that set out in discussing the technical issues between the parties. It accepted that it had sought to make a profit out of the transaction by selecting high spread names for their ratings. This was the basis of the business. However, it did not consider those profits to mean that there was an equivalent or high real world probability of default. It did not analyse its profit calculations in terms of EL. That was not its interest or concern. Even if it had done so it would not have regarded it as being inconsistent with the low probability of default implied from the instrument’s AAA rating. That is to compare the incomparable. That was its view and it was an honest and reasonable view.
Barclays emphasised in particular the evidence of Ms Duhon that “if on day 1 of a AAA rated CSO2 tranche a bank had a gross profit of 20-30% …, that did not mean … the bank somehow perceived the risk on the tranche to be inconsistent with the rating. The two things were entirely different.” They also emphasised Dr Ellis’ evidence in re-examination that that could be reasonable a view.
The Purchase Representations
The case against Mr Agresta
In support of its allegation of fraud against Mr Agresta CRSM stressed the following points in particular.
(1) Mr Agresta’s role - CRSM submitted that it is clear from the contemporaneous documents that, even though relatively junior, Mr Agresta played a central role in arranging the transactions, and was the main architect of the CDO2s. He, with assistance from others, produced the fact sheets and the scenario analyses which were included in the fact sheets. He was also involved in producing the comparisons with historical rates of default.
In addition, it was Mr Agresta who provided the fact sheets to Mr Ferrario. He knew the scenario analysis was going to be provided to CRSM and he agreed that he knew that CRSM was likely to rely on the information as being accurate and was entitled to expect that it was correct. He also had direct contact with CRSM in relation at least to the Cernobbio Note, in the form of a telephone conversation with Mr Micheloni, as evidenced by the emails of 27 and 28 October 2004.
(2) Mr Agresta’s understanding of the alleged representations - CRSM submitted that as a practitioner in financial services specialising in structured products, Mr Agresta knew the significance of a AAA rating and the very low risk of default which this would naturally be understood to imply. He confirmed in his oral evidence that he knew that a AAA rating meant the lowest expectation of default risk.
It was submitted that Mr Agresta must also have been aware, since he produced or was involved in producing the scenario analyses, that the comparisons made with historical default rates further reinforced the impression given to CRSM that the risk of loss to the ‘AAA tranche’ of the CDO2s was remote.
(3) Mr Agresta’s knowledge of Barclays’ profits - CRSM submitted that Mr Agresta knew that Barclays was aiming and expecting to make very large gross profits on the CDO2s. He was involved at the start of each transaction in discussing profit targets with Mr Ferrario and how those could be achieved. He also prepared or forwarded estimates of the day 1 profits at the time of closing each trade.
It was submitted that Mr Agresta had a very good understanding of Barclays’ pricing model and how it related to the structure of the CDO2s. Thus, he was very well aware that whether there were losses on the Notes “is purely driven by credit events in the underlying entities”, and that in valuing a CDO2 the need to form a view about whether there will be such defaults so that the notional amount will reduce is in principle “one of the building blocks of the valuation”. He further agreed that the way in which Barclays’ model calculated Barclays’ P&L involved calculating the expected losses caused by such defaults. He also had good familiarity with the relationship between spreads and default probabilities or expected loss.
(4) Mr Agresta’s knowledge of the arbitrage by which the profits were achieved - Mr Agresta understood better than anyone how Barclays’ profits were achieved, as he was the person who did most of the work of structuring the CDO2s to achieve those profits. Mr Agresta accepted that one way in which this was done in this case was by including in the reference portfolios names which had high spreads for their rating.
CRSM submitted that it followed that at the time Mr Agresta must have known and calculated in detail the extent to which the composition of the reference portfolios was distributed in this way and the impact which it had on the expected loss and that he therefore knew that, as a result of the portfolio selection which he himself had devised, the CDO2s had much higher probabilities of default than their AAA ratings implied. He must also have appreciated that, the comparisons made in the fact sheets with historical default rates were misleading, as were the statements that the CDO2s were “particularly robust with respect to defaults”.
(5) Mr Agresta’s evidence that he regarded credit spreads as unrelated, or only weakly related, to default risk should be rejected - It was pointed out that this was not supported by the contemporaneous documents and that two documents in particular show that he believed and understood that the two were directly related; namely (1) his email to Mr Ferrario of 8 July 2004, and (2) his email exchange with Mr Monkkonen on 14 October 2004.
In his 8 July 2004 email Mr Agresta set out figures which showed how much profit would be lost if either the 2 or 4 names with the highest spreads were taken out of CRSM’s portfolio. It was submitted that the only possible reason why the removal of these names was being considered at all was that it was recognised that their inclusion in the portfolio significantly increased the risk of default to which CRSM was exposed.
I have already set out the terms of Mr Monkkonen’s email of 14 October 2004 (see paragraph 320 above). It was said that this makes explicit the direct relationship that there was understood to be between “arbitraging” the CDO2 more, by including more names at the top end of the scale of spreads for their ratings and the likelihood of defaults occurring even within a few months.
(6) Mr Agresta was prepared to say things which he knew to be untrue – CRSM submitted that there were instances apparent from the documents in which and which therefore reflect on his integrity. In particular:
He concealed, in the fact sheet for Como II, the fact that the rating, stated there as simply as ‘AAA’, in fact related only to part of the potential duration of the investment.
Mr Agresta had an exchange of emails with Fitch about the terms in which the rating would be expressed and whether it would be public. In one of these emails Mr Agresta requested a public rating “because the client does not know how to monitor the rating of such a transaction”. When asked in cross-examination how he knew this, Mr Agresta said that he did not and that he was “bluffing with the rating agency to get the best deal”.
In the last of the emails in which Mr Agresta was putting pressure on S&P, Mr Agresta told S&P: “We are trying to rate this with another agency.”In cross-examination he stated “I really don’t remember if we had some other rating agency waiting to rate the deal”.It was submitted that it is plain from the documents that Barclays did not, and Mr Agresta must have known this.
In summary, it was submitted that it is evident that Mr Agresta was an enthusiastic member of Barclays’ structuring team who was seeking to act in what he perceived to be his employer’s best interests by finding ways to maximise its profits. In pursuing that aim, however, he produced or helped to produce fact sheets to assist in selling the products that he structured which were designed to convey the impression that the products carried far less risk than he knew and believed to be true.
I shall address each of these points in turn.
(1) Mr Agresta’s role – Although Mr Agresta was closely involved in the structuring of the Notes and the putting together of the fact sheets, he was working with others and being overseen by both sales and compliance, who were ultimately responsible for what was put forward. His role was essentially a technical one and to follow instructions given to him.
Mr Agresta’s limited role made it inherently unlikely that he would make statements with any intention to deceive. Mr Agresta was a junior structurer. It was his job to structure products in accordance with the instructions provided to him. Mr Agresta was not a “client-facing” member of the team. He had no direct contact with clients and consequently had no real knowledge of how matters may have been presented to them, what they may have been told or of what their interests and understanding may have been. His role was first to structure the Notes subject to the guidance and targets he received from his bosses, and second to provide technical input into the documentation which the sales department were to present to the client. But as he made clear, the choice of which types of information was to be provided and how it was to be presented was not his. It was the compliance department, in particular, that mandated what information was necessary and how it should be presented.
Mr Agresta’s junior role also meant that he had no real motive to seek to mislead CRSM by making knowingly false representations. It was not his job to sell the product to CRSM and he had no knowledge of what might or might not be needed to achieve that. Although it was part of his role to seek to ensure that Barclays made profits on the transactions he earned no sales commission. On the other hand, the professional and reputational risks for him of making knowingly false representations would be great. Even an inadvertent inaccuracy would be likely to earn him the criticism of his bosses in structuring, the sales team, and compliance. He would therefore have every incentive to ensure that, to the extent he was involved in putting together documentation to go to CRSM, or assisting with information, any input he added was accurate and factually correct, and not misleading.
(2) Mr Agresta’s understanding of the alleged representations – Mr Agresta knew what a AAA rating means but he had no particular reason to consider or know what significance this would have had for CRSM. As he explained in evidence - “I didn’t have any understanding of the real reasons why the client wanted a AAA. I’ve never been in business meetings with the client.” As far as he was concerned the AAA rating was a requirement for the structure which he had to work to. As he said in evidence - “To me the AAA was a constraint of the structuring problem that I have to solve”.
As Mr Agresta understood it, the purpose of the relevant section of the fact sheets was to provide the principal factualcharacteristics of the trade. He did not consider that it would or could be construed as containing any wider credit analysis. There is no reason why he would believe that by referring to a AAA rating in documents of this kind Barclays was doing any more than stating that the notes were AAA rated.
As Mr Agresta explained in evidence, as a structurer he had no reason to consider issues such as “expected losses” or “probability of default” - “I wasn’t a trader and I’m not a trader myself, so while I was structuring the transactions, I never made considerations around the expected losses as implied by the pricing model. I was using the pricing model as a way to understand the feasibility of the transaction …” ; “My role as structurer simply did not lend itself to assessing credit risk for an end investor, and it would not have occurred to me to use the day 1 P&L to do so in any event.”
Since he had no reason to assess the credit and default risk for the investor, he equally had no reason to make generalised representations about that risk. It is also unlikely that Mr Agresta can have seen himself as making any such representations when he saw his role as providing factual data and analysis, with broader questions of presentation being for others. As he said – “I never made subjective statements on their riskiness” ; “I never conveyed to our sales people, or encouraged them to convey to CRSM, that the investments were in any subjective way, “secure””.
For similar reasons, it is unlikely that Mr Agresta understood that the paragraph about historical default rates in the fact sheets, and the words “particularly robust”, meant that he was making a generalised representation that the notes had a low risk of default. As far as he was concerned his role was to provide factual information relating to the trade. As he said, the scenario analyses “were provided to the client in order to help them assess the portfolio, I believe I provided information that accurately reflected the economics of the transaction. It was up to CRSM to decide what further analysis it wished to see and to make its own decisions about whether to invest …”; “I note that the fact sheet states that the structure was “particularly robust with respect to defaults”. I am not sure whether it was me or someone in sales or compliance who drafted this sentence …. In any event this was just general wording introducing a detailed technical paragraph. The paragraph read as a whole was intended to describe in words what is shown in the graph.”
His role was to set out the facts as instructed by sales and compliance; it was the sales department’s role to present them; it was the investor’s decision what to make of them. As he said in evidence: “I never thought that an investor in 230 million of CDO squared …. could have made a decision solely or predominantly on a five-page fact sheet. I never thought that that was possible. The fact sheet was something to summarise in five pages the principal characteristics of the trade, which is what you have in the first two pages, and to show the worst case scenarios. That was the purpose of the fact sheet. It was just one of the many information that the investors in CRSM had to make the decision” ; [when asked whether he was “expecting the investor to … do other analysis”] “I didn’t expect anything. If the investor – any time the investors were asking the salespeople questions or analysis, my job was to deliver the analysis in a correct way. If the internal control function and compliance was asking to incorporate other analysis, I was happy to be helpful and to help make the analysis”.
(3) Mr Agresta’s knowledge of Barclays’ profits – It was Mr Agresta’s evidence that ensuring that Barclays made a profit was one of the “constraints” to which he had to work. The internal emails at the time of the trades also show that he had an awareness of the expected level of such profit. However, he had no particular view as to the level of these profits. As he explained in evidence - “I did not have any opinion about the day 1 profits or whether they were out of line with what I would have expected from my limited experience of similar trades at the time. I had certainly seen net day 1 profits of over 10 per cent on other transactions … However, as I had only been working in a bank for one or two years at that point, I was not in a position to determine what an appropriate level of profit would have been.”
Although Mr Agresta had a good understanding of how the pricing model worked, as I have already found, he was not considering credit or default risk or what the model may have implied about such risk, nor did he consider the model to be a reliable indicator of such risk. As he said -“I did not consider that there was any clear or deterministic relationship between the day 1 P&L and the rating or credit risk of the transaction. In my mind these were largely unrelated things.” ; “the pricing model was not a model that gave a reliable insight into the credit riskiness of the structure”;“The day 1 P&L figure was an estimate only, and would have changed dramatically had it been calculated by another bank with a different methodology. I recall that Dresdner, a competitor bank, booked profit much more conservatively at the time, on an accrual basis. If Barclays had booked day 1 P&L in this way it would have made around 3% on day 1 but the risk to the investor would have been exactly the same.”
(4) Mr Agresta’s knowledge of the arbitrage by which the profits were achieved – Mr Agresta knew that profit for Barclays was to be achieved by including in the reference portfolios names which had high spreads for their rating. As far as he was concerned that was the basis of the arbitrage CDO business. He saw nothing wrong or deceptive about that. He did not consider that this impacted on the actual or factual probability of default. As he said: “…this was the basis of the arbitrage CDO business…the CDO business in general and also the single tranche CDO business in particular existed because historically there had been a consistent discrepancy between the market implied default probability as implied by credit spreads in the market and the historical, which is the actual…factual default probability..”
Mr Agresta’s evidence, which I accept, was that he did not consider that credit spreads were a reliable indicator of the real world probability of default. As he said: “The main concept was that, for a number of reasons, including regulatory and risk managing, Barclays on a single tranche CDO or a CDO squared had to hedge itself, and the most readily available instruments were CDS spreads”. “Spreads were used in the pricing model for a hedging reason because Barclays had to mark to market the positions on day one. But I also say it is debatable what spreads actually represent. They don’t just represent a premium to compensate investors for true credit risk. They represent other factors, like irrational supply and demand, liquidity, excess premium to the asymmetric profile of a credit investment”; “We need to distinguish the pure default risk from the market-perceived risk. That is why the ratings agencies didn’t use spread in their models, because spreads were measuring a number of other factors and sometimes it is pure irrationality from the market … you have all these factors which are incorporated in the spreads, but from the rating agency point of view they are only interested in the true default rates.”
(5) Mr Agresta’s evidence that he regarded credit spreads as unrelated, or only weakly related, to default risk should be rejected - I accept Mr Agresta’s evidence on this issue. I found him to be a credible witness both generally and on this issue in particular.
It is correct that the emails relied upon indicate an awareness that other things being equal a similarly rated name with a large credit spread may pose a higher risk of default than one with a small spread. However, that is a market snapshot. It may be influenced by a number of market factors and may change significantly over time. Further, it does not tell one anything reliable about what the difference in probability of default (if any) actually is. As Mr Agresta explained when addressing the fact sheet historical scenario analyses: “We were never doing this analysis taking into account the spreads, because we know that the spreads can indicate a number of different factors and can be very volatile. Here we are completely immersed in an historical overview, looking at the actual realised default rate…”
It was not Mr Agresta’s job to make assessments of credit risk based on credit spreads, and he did not do so. He viewed figures of expected loss derived from credit spreads as essentially non-comparable with what the documents he was involved in, the fact sheets, were discussing.
There is nothing surprising or implausible about this being Mr Agresta’s view and it is supported by Ms Duhon’s evidence of the perception at the time of reasonable practising bankers as set out in paragraph 298 above.
By contrast, it would be both surprising and implausible, as CRSM’s case supposes, for Mr Agresta, a junior structurer of 1-2 years’ experience in the job, to have decided to second guess that market perception, and also the practice of the ratings agencies, and formed a view that what the ratings agencies and the historical data was saying was wrong, even though credit analysis was no part of his job.
(6) Mr Agresta was prepared to say things which he knew to be untrue – Mr Agresta was thrust into the proceedings at very short notice following CRSM’s late amendment to plead fraud against him. He was cross-examined at length but very few inroads were made into his credibility and I found him to be a generally honest and straightforward witness.
In relation to the fact sheet description of the rating for Como II, whilst it could have been more clearly expressed Mr Agresta did not consider it to be inaccurate. As he pointed out, the translation should have been “The structure has a rating of AAA rate from Fitch on the date of issue” and, as he read it, what was said naturally applied to the initially maturity period. In any event, a ratings letter was sent to CRSM making the point clear on 19 January 2005 before the Bellagio transaction closed. This is inconsistent with there being any intent to deceive on the part of Mr Agresta with respect to the duration of the short rating.
In relation to the emails with Fitch and S&P it does appear that Mr Agresta was not being entirely truthful, but I do not consider that that impacts on his credibility generally, which was tested extensively but largely unsuccessfully in cross examination.
For all these reasons I reject CRSM’s case of fraud against Mr Agresta in respect of the Purchase Representations and, in the light of the findings set out above, I find as follows:
Mr Agresta did not understand that he or Barclays’ was representing that the Notes had a very low risk of default, or that that was his or Barclay’s belief or expectation.
If such a representation had been made, and been understood to have been made, Mr Agresta would have considered it to be true.
As a matter of his personal opinion, Mr Agresta believed that these were good notes and not excessively risky. As he said: “I believed in the product”. He believed that they did have a low risk of default, and the fact that the ratings agencies rated the Notes AAA supported that conclusion.
I further find that Mr Agresta had no intention to mislead CRSM. As he said: “I never wanted to mislead CRSM …. Even with the benefit of hindsight, I cannot find any wrongdoing or anything misleading in the way I structured the notes”. I accept that evidence.
I find that if the alleged representations were made they were made with reasonable grounds for belief in their truth. In particular:
The Notes were rated AAA by the credit rating experts, the CRAs.
The structure of the Notes did involve considerable protection against defaults.
That such AAA notes were effectively “bullet proof” was a view shared by reasonable practising bankers at the time.
It was reasonable to take the view that credit spreads and an implied probability of default derived from the pricing model did not provide a reliable measure of the real world probability of default.
The profits that Barclays were making were not understood to be abnormal. Further, on Ms Duhon’s evidence, they were neither abnormal nor inconsistent with the AAA rating - “the two things were entirely different”.
The fact that the average credit spread was higher than the average for the ratings category, was a normal part of the trade. It was the arbitrage that the CDO business was based on.
All the information Mr Agresta did provide was carefully checked, was based on data and was factually accurate.
The facts and matters set out at paragraphs 296 to 299 above.
The case against Mr Ferrario
In support of its allegation of fraud against Mr Agresta CRSM stressed the following points in particular.
(1) Mr Ferrario’s role - CRSM submitted that Mr Ferrario was the individual who had almost all the direct communications with CRSM and was responsible for making the statements which were the alleged misrepresentations. He was described by Mr Agresta as “a very smart salesperson”with a good general understanding of the products.
It was further submitted that Mr Ferrario knew what the fact sheets said, and that they were intended to be relied on by CRSM, because he received them from Mr Agresta (sometimes having worked with Mr Agresta on them), provided them to CRSM himself and discussed them with CRSM’s representatives. He knew that CRSM was relying on the AAA rating and other statements made to CRSM and that the context of those statements was CRSM’s requirement for a safe investment with a low risk of default. Mr Ferrario also knew that CRSM had no models and no ability to check the pricing or structure of the CDO2s or properly to assess their risks.
(2) Mr Ferrario’s knowledge of Barclays’ profits – CRSM submitted that Mr Ferrario knew that Barclays was aiming and expecting to make very large gross profits from the sales of the CDO2s. He confirmed in his oral evidence that, at the start of each deal, when the CDO2 was about to be put together, he would have discussions with the structurers about the form it would take and about the expected profit from the deal. This is borne out by his various emails requesting P&L figures for the transactions. He also had a direct interest in the profits achieved because it bore on his sales commission.
CRSM submitted that in evidence Mr Ferrario was defensive about the size of Barclays’ profits and his active involvement in seeking to maximise those profits (in which he had a vested interest) and that the reason for this was:
As a banking professional with more than 10 years experience, Mr Ferrario cannot have imagined that it was possible for Barclays to make gross profits of 20-25% (before deduction of its costs and reserves) from the sale at par of a security paying a coupon of 1% above Euribor without exposing CRSM to more than a very low risk of default.
It must have been all the more apparent to him that Barclays’ profit was being achieved by exposing CRSM to substantial risk when the nature of the transaction was such that the amount of Barclays’ gross profit was directly equal to the amount of CRSM’s expected loss.
CRSM further submitted that the fact that, as Mr Ferrario knew, investors such as CRSM did not assess the risk of a complex structured credit transaction using similar analytical methods to those used by investments banks, and were reliant on far more simplistic methods, serves to emphasise his awareness of the disparity between the parties in terms of sophistication and expertise which enabled Barclays to make such large profits at CRSM’s expense.
(3) Mr Ferrario’s understanding of the arbitrage – CRSM submitted that Mr Ferrario understood how it was that the structurers were able to achieve such large profits for Barclays while also obtaining AAA ratings for the CDO2s. They pointed out that Mr Agresta confirmed in evidence that he would have expected Mr Ferrario to have been aware of this.
Mr Ferrario acknowledged in evidence that he was indeed aware that one of the things the structuring team did, in order to achieve value for Barclays, was to include in the reference portfolios assets which had high spreads for their ratings. It was submitted that this evidence stands in contrast to Mr Ferrario’s claim in his fourth witness statement that he did not know whether the practice of selecting assets with high credit spreads for their rating occurred.
CRSM further submitted that Mr Ferrario accepted in oral evidence that, in general terms, the higher the spreads included with given ratings, the greater the expected profit for Barclays that would be generated.
Understanding as he did that the way in which very large profits were being generated for Barclays was by including in the CDOs names with high credit spreads for their ratings, CRSM submitted that Mr Ferrario must therefore have realised that this involved exposing CRSM to considerably greater risk than was reflected in the AAA rating.
(4) Mr Ferrario’s understanding of connection between spread and risk – CRSMsubmitted that Mr Ferrario’s denial of knowledge of the link between spread and risk should be rejected.
CRSM submitted that it was plain that he did understand the link between higher spread and increased risk from some of his own contemporaneous emails. For example, on 20 September 2004 Mr Ferrario sent an email to Mr Agresta and others in which he wrote:
“I am worried about Bombardier and Intelsat ... Market has tightened and they continue to widen a lot ... plse advise if we have to substitute ... would rather not do in [sic] before the next trade but can’t afford a default before November.”
It was submitted that in this email Mr Ferrario was expressing concern that the widening spreads of Bombardier and Intelsat, particularly when spreads generally were tightening, indicated they were at risk of defaulting even in the next two months. It was further submitted that this was eventually accepted by Mr Ferrario in evidence, but only after a number of evasive answers had been given.
CRSM submitted that Mr Ferrario was again evasive when asked about the email sent by Mr Agresta to him and others on 8 July 2004 setting out what the impact on Barclays’ P&L would be if names with spreads higher than 250 basis points were taken out of the Umbria II portfolio. The background to this email was Mr Ferrario’s request the night before to “see the economics in writing before I trade”. He received the projected P&L figures early the next morning, to which he had replied in his email sent at 9.36am, referred to above:
“Thanks will be in the office shortly and we can discuss what to do. No marks should ever be asked by the client on this trade ...”
It was about an hour after this, at 10.36am, that Mr Agresta sent to Mr Ferrario and the other recipients of Mr Ferrario’s earlier email figures showing how much profit would be lost if either the 2 or 4 names with the highest spreads were taken out of CRSM’s portfolio.
It was submitted that it is clear from the context as well as the content of Mr Agresta’s email that a discussion took place about whether Barclays should forego some of the large profit which it was expecting to make on the Umbria II transaction by removing from the portfolio before the trade closed some of the names with the highest spreads, so as to reduce the risk of default to which CRSM was exposed. Since none of the names specified in Mr Agresta’s email was in fact removed, it is apparent that the outcome of the discussion was a decision to maximise the profit rather than reduce the risk.
When asked about this email, Mr Ferrario’s response was to claim that he was effectively just a passive recipient of the email and that the matters in it were not really to do with him. However, CRSM submitted that it is obvious that the reason any such change was being contemplated was precisely because the higher spread names were seen as the most risky. Indeed, in his first witness statement Mr Ferrario had said, in the context of the restructuring, that spread figures were eventually provided to CRSM at Mr Buoncompagni’s request “as he was aware of the significance of high spreads on the reference assets”.
CRSM submitted that Mr Ferrario was again evasive in evidence when he was asked about another of his own emails, sent to CRSM on 17 December 2004 and referring to “improvements” to the portfolio including, in particular, the removal of Intelsat, the name “with the highest spread” in the portfolio.
It was CRSM’s case that the reason why Mr Ferrario was continually reluctant to admit the connection between a high spread and a high risk of default was because he understood its implication: that he was aware at the time of the transactions that the inclusion of so many names with high spreads for their ratings created a substantial default risk which made the AAA ratings, and the statements made to CRSM about risk of default on the CDO2s, misleading.
In summary, CRSM contended that on the figures that he gave for 2004, more than 90% of Mr Ferrario’s total remuneration consisted of bonus, which in turn reflected his sales credits. Try as he did to downplay it, the extent to which Mr Ferrario was pushing for very high profits in the transactions with CRSM is clear from the documents. In the pursuit of those profits, he was prepared to ignore, or at least turn a blind eye to, the obvious fact that those profits were being achieved by building into the transactions with CRSM a much greater risk of loss on the investments than he led CRSM to believe.
Barclays objected to CRSM putting a case of fraud against Mr Ferrario on any basis other than his knowledge of Barclays’ profits. In particular, objection was taken to any case being put on the basis of credit ratings arbitrage and the structuring of the transaction. It was pointed out that CRSM’s plea of knowledge of falsity as against Mr Ferrario (as opposed to Mr Agresta) is limited in this way. If there was any doubt about this it is confirmed by the explanations provided by CRSM’s counsel at the hearing before Mr Justice Blair at which permission to amend was obtained. Having reviewed the pleading and how it was explained at the hearing before Mr Justice Blair I agree with Barclays that CRSM’s case against Mr Ferrario is so limited. Nevertheless, without prejudice to that ruling I shall address the wider case which has been put. I shall address each of CRSM’s points in turn.
(1) Mr Ferrario’s role – It is correct that Mr Ferrario was the salesman who presented the transaction documents to CRSM. Although he did not draft those documents he would have been familiar with their contents and have ensured that he was to able explain them as may be required. As found above, he also had various discussions with CRSM representatives about the transactions and the documents.
Mr Ferrario knew that CRSM required the Notes to be AAA rated and he also knew what an AAA rating means. However, he did not understand that he or Barclays were making any generalised representation of risk of default or that the description AAA conveyed anything more than the fact of the AAA rating.
As Mr Ferrario explained “It also never occurred to me that Barclays or I might be making implied representations that Barclays believed the CDOs to be “secure investments” with “an extremely low risk of default” and had structured them with that intention. As I have explained, we went out of our way to make the risks very clear to CRSM. I did not understand anything I said to imply the specific further representation now alleged. All that CRSM should have understood from what I said to them and the documents provided by Barclays was that the CDOs would be structured to achieve a AAA rating.” I accept this evidence which I found to be credible and consistent with the inherent probabilities. In particular:
Mr Ferrario was a skilled and astute salesman. He knew the importance of sticking to the facts and not making vague or generalised statements about the transaction or the risks it involved.
Mr Ferrario drew CRSM’s attention to the fact sheet disclaimers which included a Credits Rating disclaimer.
He drew CRSM’s attention to various risks involved and in particular MTM risk, volatility risk and liquidity risk.
He rejected the suggestion that he ever said anything to suggest that the Notes had anything other than a very low risk of default. As he said: “…I actually said the opposite, my lord, on many different occasions. We talked about leverage, we talked about cliff risk, we talked about a lot of elements, and a lot of the representatives of San Marino were fully aware of the fact that this was a highly leveraged investment whose default risk might not have occurred with a few defaults within the portfolio but would have accelerated with – over a certain number of risks.”
As I have already found, he explained to Mr Montanari and Mr Sapignoli that they would have to use their own judgment to assess how many names were likely to default.
I make similar findings for similar reasons in relation to the fact sheets and the scenario analyses. He understood these to be a factual analysis and no more. It was up to CRSM to decide what, if any, reliance to place upon it. He did not understand that it conveyed any generalised representation as to risk of default, and he stressed that it was up to CRSM to assess the likelihood of default by the names. As he said: “The fact sheets provided in relation to each of the CDOs made a reference to each series being particularly robust in relation to defaults. I understood that to be a comment on their robustness to default compared to historical default rates. I believed the analysis to be correct, and from my own experience, I thought that the transactions were not out of line with other deals of this nature”;“it was intended to give a factual analysis of the worst case scenario to help the investor …. It was really up to the investor to decide whether they liked the risk associated with the scenarios and it was up to the investor to give a probabilistic view about this. All we could do was to give a factual analysis. I think there is a fairly clear factual analysis …”
It is correct that he was aware that CRSM placed reliance on the AAA rating. However, that was reliance on the rating – it was not reliance on a representation by Mr Ferrario as to what the AAA rating meant. As he said: “Montanari was always saying “OK, OK, I understand this is how it is going to behave, but I have the comfort it is AAA rated.””
It is also correct that he was aware that CSRM placed reliance upon the scenario analyses. However, again that was on the basis of the facts of the analysis. Any wider consideration was a matter for CSRM, as Mr Ferrario made clear.
It is also correct that he knew that CRSM did not have a pricing model. However, as Mr Ferrario knew, if MTM value was regarded as being a matter of significance CRSM could always have asked Barclays for a MTM value. When, at later stages, such values were requested they were always provided. This was never something Barclays could somehow hide.
(2) Mr Ferrario’s knowledge of Barclays’ profits – Mr Ferrario did have an awareness of the likely level of profits to be made on the transactions and he did have a financial interest in such profits due to the sales credits he could earn.
It was not, however, for him to fix targets for the profits to be made. Further, the sales credits on this transaction were only a part of his overall sales credits for the year and there was no direct relationship between sales credit and bonus, as bonus was a discretionary decision. This suggested financial motive is slight compared to the professional and reputational risks involved in making any knowingly false representations. Moreover, this would also risk undermining a long standing and commercially satisfactory relationship with CRSM, and endangering his future with Barclays (whom he had joined only in August 2003) and in the industry.
Although the gross projected profits were large Mr Ferrario understood that it was the net profits which were the figure which matter and it was his evidence and that of Mr Agresta that day 1 profits of over 10% were not abnormal on transactions of this kind.
As to the alleged relationship between profits and expected loss this was not something which he appreciated or focused on at the time. As he explained in evidence, he thought that day 1 P&L numbers had “very little, if not nothing, to do with the chances of the transaction themselves defaulting at maturity”; “I never thought at the time, and I am still convinced now – I never knew at the time that these numbers would reflect expected losses because I wasn’t using numbers myself, and I never thought of those numbers in that way. But most importantly, I don’t think it has anything to do, in the real world [with] whether the notes would default.” In his mind, there was a clear distinction between the relevance of the levels of the credit spreads to “mark to market” risks and to real world default risks.
That he did not draw a link between profits and risk of default is borne out by his evidence, which I accept, that he did indeed believe that the Notes had a very low risk of default. As he said: “I honestly believed the notes would go to par unless the default levels in the market reached a level shown in the scenario analyses provided to CRSM. Those default levels would have represented an unprecedented and huge change in market conditions …. No-one within the structuring or trading teams suggested to me that they thought the CDOs were weak either”; “I did not have any belief or suspicion that CRSM would lose money on the CDO notes at maturity. Barclays took great care to explain exactly what would happen, in terms of defaults on the relevant reference portfolios, before it would lose money”; “CRSM was not expected to lose that kind of money. CRSM was expected to earn the coupons paid plus spread. Unless unprecedented clearly described events would have occurred, they were expected to get par back on maturity”.
(3) Mr Ferrario’s understanding of the arbitrage – It is correct that Mr Ferrario accepted that he was aware that it was in Barclays’ interests to use names with high spreads for their ratings. It is not, however, necessarily inconsistent with his witness statement. His witness statement acknowledged that there would be nothing “unexpected or wrong about that” and that it would be “more efficient” for Barclays. However, he did not know whether that had been done in this case as he was not a structurer and could not comment in detail on asset selection. Whilst Mr Ferrario knew of the generally expected levels of profits he had no detailed knowledge of how this had been achieved.
(4) Mr Ferrario’s understanding of connection between spread and risk- I accept that Mr Ferrario was aware that there is a connection between spread and risk and that a name with a large credit spread is perceived by the market as posing a higher risk of default than one with a small spread. However, he did not consider that there was a direct or quantifiable relationship between spread and real world default risk.
The credibility of that evidence is supported by the evidence of Ms Duhon to which I have already referred.
For all these reasons I reject CRSM’s case of fraud against Mr Ferrario in respect of the purchase representations. I found Mr Ferrario to be a generally honest and straightforward witness. In the light of the findings set out above, I find as follows:
Mr Ferrario did not understand that he or Barclays’ was representing that the Notes had a very low risk of default, or that that was his or Barclay’s belief or expectation.
If such a representation had been made, and been understood to have been made, Mr Ferrario would have considered it to be true.
I further find that Mr Ferrario had no intention to mislead CRSM.
I further find that if the alleged representations were made they were made with reasonable grounds for belief in their truth for essentially the same reasons as given in respect of Mr Agresta. Further, as a salesman who was not involved in the structuring of the transaction Mr Ferrario had little knowledge of many of the detailed matters relied upon by CRSM in support of its case.
The Restructuring Representations
The case against Mr Agresta
Misrepresentation of intention to profit
In his witness statement Mr Agresta stated that: “I have addressed the zero cost allegation … above. This is an untenable allegation. I never discussed with CRSM what zero cost meant. However, the idea that Barclays would not have required some day 1 P&L (even a very small amount) is nonsensical, because the volatility in the market would have meant that Barclays would have been at risk of losing money and unable to close the transaction if the market moved against it. … We meant by “zero cost” that CRSM would not have to pay an upfront cash sum and that its coupon would remain unchanged. I believed this at the time and I did not represent otherwise to CRSM.” I accept that this is what he understood “cost zero” to mean.
In cross examination it was put to Mr Agresta that the reference to “costo zero” “clearly meant that the changes were going to be made for free, and not at a cost to CRSM”. Mr Agresta’s response was to disagree and to explain how he had actually understood it: “In our mind, it meant that the client didn’t have to pay any cash, and also it meant that its economics were the same as before, in particular that we didn’t have to lower the coupon.” I accept that evidence.
It follows that even if, contrary to my findings above, “cost zero” involved a representation that Barclays would be left no worse off and would not profit from the restructuring, Mr Agresta did not understand that any such representation had been made and did not intend that it should be understood in that sense.
Misrepresentation of the Selection Criteria
CRSM relied in particular on the following:
It submitted that the evidence demonstrated that in formulating Barclays’ proposals for substitutions to the CDO portfolios Mr Agresta (assisted by Mr Bechet) had pursued three objectives: he had sought to find names to substitute which, in general, had (1) higher credit ratings, (2) higher ratings according to Mr Scaramanga’s model, and (3) higher spreads for their ratings than the names which were replaced (leading, overall, to higher expected loss). By stating that the first two criteria had been followed and omitting reference to the third, the emails and memoranda which described the proposals were deceptive.
Mr Agresta was the person principally responsible for drafting the communications which described Barclays’ proposals. In particular he had drafted the 22 April email (in all its versions) and the 12 May email (apart from 2 amendments made by Mr Ferrario), and he was jointly responsible with Mr Ferrario for preparing the June memorandum (which was itself based on the 12 May email).
Mr Agresta was also the structurer principally responsible for selecting the names proposed for substitution. He knew that the names proposed for substitution had been chosen (by him, with the assistance of a more junior structurer) so as to meet not only the two stated criteria but also the third (unstated) criterion of higher spreads for their ratings than the names removed – with the aim of increasing the expected loss to CRSM on the CDO2s. It was submitted that his statement in these circumstances that two criteria had been followed and omission of the third can only have been deliberate and the result of a conscious choice to conceal from CRSM Barclays’ objective of increasing the arbitrage on the CDO2s at CRSM’s expense.
It was further submitted that the inference of deliberate deception is supported by the removal of the spread information from the email sent by Mr Agresta to CRSM on 22 April. It was submitted that this deletion can only have been the result of a deliberate decision and that the only plausible reason for the deletion is that Mr Agresta did not want CRSM to see that the spreads of the names added were in general significantly higher than the spreads of the names removed.
Reliance was placed on reassurances given to CRSM that they would not be harmed by the increase in spread following this being discovered by Mr Buoncompagni in early June 2005. On 7 June 2005 Mr Ferrario forwarded an email, prepared by Mr Agresta, to Mr Buoncompagni which discussed the calculation of the average spread of the names with the highest spreads. Mr Agresta’s analysis compared two methods of calculating the average spread of the five names with the highest spreads – described as the “San Marino” method and the “Barclays” method – with reference to Como II. Whilst Mr Agresta suggested that Barclays’ method (which produced lower averages than CRSM’s method) was to be preferred, CRSM stressed that it is notable that both methods showed that the average spread of the five highest spread names was lower in the new Como II than in the current Como II portfolio. On CRSM’s calculation, the average spread decreased from 5.45% to 4.758%; and on Barclays’ calculation, the average spread (while lower than on CRSM’s calculation) also decreased from 4.91% to 3.81%. Although, the email does not explain how it came about that methods of calculating the average spread of the highest spread names were being discussed, CRSM submitted that the likely explanation is that Barclays had responded to Mr Buoncompagni’s observation that the average spread was increasing by making a point that, even though this was the case for the portfolio as a whole, the average spread of the highest spread (and therefore riskiest) names was in fact decreasing.
In all the circumstances, it was submitted that Mr Agresta concealed the manner in which, and object with which, Barclays was arbitraging the CDO2s and increasing the expected losses, by substituting names which had high spreads for their ratings, remained misleading. It was submitted that the evidence shows that Mr Agresta intended to mislead CRSM in this respect.
Barclays made a number of general points about CRSM’s case on this issue with which I agree.
First, Mr Agresta made a number of points about the restructuring, which evidence I accept. In particular he said:
He always assumed that Barclays would be making profits on the restructuring: “Not only would it not have been commercially sensible for Barclays to do a large amount of work, and to make substitutions that would be costly to it, without obtaining some benefit, but also market conditions were changing rapidly at the time and we had to build in some profitability to make the switches executable. Otherwise, Barclays would lose money from the restructuring.” “If profit was not built into the structure the restructuring would not have been executable”.
“It was my belief that the CDO2s were better structures after the restructuring.” “The new structures offered CRSM greater default protection, as the Forced Default Analysis demonstrates”.
“My recollection was that the profit on the restructuring was not very high, which is supported by an email I have been shown dated 14 June 2005which shows the new day 1 P&L on the restructuring was actually 2 to 4%.”
“Our restructuring was successful in the short term as we decreased the risk of Delphi in the underlying CDOs and therefore its negative effect on the tranche subordination due to erosion. We also improved the portfolio across other objective dimensions and we were able to achieve all these goals without asking CRSM to pay any cash upfront.”
“Cross-subordination … was very popular across the market. It would have decreased certain CDO2 risks, such as the overlap risk and idiosyncratic risk, and increased default resilience. The presentations at the time clearly set out these risks and the benefits of cross-subordination. I believed that this was a beneficial technology and that specifically it would have improved the structure of CRSM’s investment”.
Second, Barclays rightly emphasised that Mr Agresta had only a very limited role in the preparation of the actual memoranda which were the presentations on which the transaction was actually concluded. As he explained, the memoranda, starting with the draft of 2 June 2005, were “very different” to his earlier emails as to both “philosophy and the context”. It was a “more robust and fundamentally different memorandum”. But his role in the preparation of the memoranda was limited. The principal draft was essentially produced by compliance and the most senior salespeople and it was carefully reviewed by them - his contribution was essentially confined to providing data for inclusion in the tables. He was a “very passive” contributor to the memorandum. The decisions as to what to include in it, and as to the representations it was making, were not made by him.
Third, although Mr Agresta prepared the earlier emails sent to CRSM he did not regard them as making presentations to the client. They were instead, from his perspective, emails to the salesforce, which they could use to present the transaction to the client. He was aware that Mr Ferrario was in some cases passing them on direct to the client, but he had no broader understanding of the context of the “communication flow” between Mr Ferrario and CRSM, so he did not know what CRSM understood. They were also very much preliminary proposals. Thus, as he said about his email of 11 May 2005 to Mr Ferrario which Mr Ferrario turned into his own email and sent to CRSM on 12 May 2005: “This to me was just the basis to start the discussion with the client; for Matteo in particular to present the business case”.
Turning to consider each of the main points made by CRSM.
As to (1), it is correct that building in a profit for Barclays was one of the criteria which Mr Agresta had to meet. For reasons already stated, he did not think of that in terms of credit or default risk and did not understand it to have any real world bearing on expected losses.
As to (2), it is correct that he drafted the 22 April and 12 May 2005 emails but his role in relation to the later memoranda was limited, as found above.
As to (3), it is correct that Mr Agresta sought to include names with high spreads for their ratings in order to ensure that some profit was achieved. However, he did not see anything wrong or misleading about this. Barclays were doing a great deal of work in the restructuring. To prevent Barclays losing money a profit cushion was required given changing market conditions. In any event the anticipated profit was small.
As to (4), this was not put, or at least not clearly put, to Mr Agresta. His evidence was that he could not remember why the names (and it was only the removed names) were left out and there could be any number of innocent explanations for this. In any event it was always likely that at some stage credit spread information for both the added and removed names would be provided, as indeed it was.
As to (5), again deception in this regard was not put, or at least not clearly put to Mr Agresta. In any event, there is no suggestion that the analysis Barclays put forward was wrong. Further, what was set out in it was two different ways of calculating the average spread, which produced different results for the impact of removing particular names. It is apparent that there had been a discussion of how to calculate the average spread, with reference to the wider spread names. Mr Agresta made clear that on Barclays’ calculation, “the average final spread is higher” (compared to the CRSM calculation). Thus, far from concealing the effect of the spread changes, Mr Agresta’s email, forwarded by Mr Ferrario, gave CRSM an alternative way to calculate average spreads which would accentuate their average increase. There was a decrease in spreads on the widest spread names on both calculations, but less of a decrease on Barclays’ calculation. CRSM’s suggestion that Barclays had responded to Mr Buoncompagni’s query by making a point that, even though this was the case for the portfolio as a whole, the average spread of the highest spread (and therefore riskiest) names was decreasing is speculation. It is also noteworthy that Mr Buoncompagni made no attempt to argue that he had been misled about the effect of the spread changes. I agree with Barclays that, if anything, what this email shows is that when questioned about spreads, Barclays gave prompt, accurate and factual answers.
For all these reasons I reject CRSM’s case that there was any dishonesty or deception involved on Mr Agresta’s part in any representation made as to the criteria to be followed in the selection of names. I repeat my earlier findings that I found Mr Agresta to be an honest and straightforward witness. He was doing his job to the best of his ability. He was not setting out or trying to deceive CRSM and had no obvious motive for doing so.
Misrepresentation relating to Scaramanga model
CRSM submitted that the evidence demonstrated that Mr Agresta deceived CRSM in the use of ratings from Mr Scaramanga’s model to persuade CRSM that the proposed substitutions would improve the quality of the portfolios.
CRSM relied in particular on the following.
(1) Limitations of the Model - Mr Scaramanga in his witness evidence explained that there were two inherent limitations on the use which could properly be made of his model; namely, (a) The model only looked at the value of a rated name relative to other names trading at similar credit spreads; it therefore could not be used to make meaningful comparisons between names trading at different spread levels (“the spread bucket point”); (b) The model was “specifically created in order to monitor the market for short-term buying opportunities”. Thus, it might be useful in indicating to an investor which name in a cohort of credits trading at similar credit spreads was a “good buy” at the moment of investing, in the sense that it might do better in the short-term; but the model “would not tell an investor anything about the likely long-term performance of the entity” (“the short term point”).
It was submitted that the second of these limitations meant that the Scaramanga model was fundamentally unsuitable for use in restructuring the CDO2s, since CRSM had bought the CDO2s as a “buy and hold” investor whose interest lay in the long-term performance and likelihood of default of the entities in the CDO portfolios over a period of some 4 – 7 years (and not in short-term trading).
In addition, it was submitted that the first limitation meant that, even if the Scaramanga model had otherwise been a suitable criterion to use, it was misleading to apply the model to the proposed substitutions and to present improvements in Scaramanga ratings as a benefit unless the entities being compared had similar spreads.
It was further submitted that the evidence showed that both limitations of the model were explained at the time by Mr Scaramanga to Mr Agresta. In his witness statement Mr Scaramanga recalled that, following initial discussions of his model with Mr Ferrario, “I spoke to him and Antonio [Agresta] about the QRV and how it worked frequently over the next few weeks”. In his oral evidence Mr Scaramanga specifically confirmed that these discussions would have included an explanation that his model was designed for short-term trading and not for long-term investment.
Mr Agresta confirmed that he had some internal meetings with Mr Scaramanga at which Mr Scaramanga explained how his method worked – although Mr Agresta could no longer remember how the method worked at the time of the trial.
(2) Misrepresentations – CRSM contended that despite being informed of its limitations, Mr Agresta chose to present ratings from Mr Scaramanga’s model to CRSM as an “objective” measure of portfolio improvement and did so in a way which ignored the limitations of the model and was actively misleading. In this connection, CRSM relied in particular on the emails of 22 April and 12 May 2005 and the June memoranda.
It was submitted that in all of these communications, higher Scaramanga ratings (in particular an increase in names with ratings of +2 and a reduction in names with ratings of -2) across the portfolio as a whole were presented to CRSM as an unqualified improvement. No mention was made of the fact that Scaramanga ratings said nothing about the risk of default of the reference entities over the duration of the CDO2s and were therefore of little or no relevance for a ‘buy and hold’ investor such as of CRSM.
Nor was it explained that, even as a predictor of short-term performance, the Scaramanga model could only properly be used to compare names trading at similar credit spreads so that, unless the names being added to the portfolios had similar spreads to those being removed, it was not legitimate to compare them on the basis of their Scaramanga ratings. In fact, as the average spread of the portfolios was increasing as a result of the substitutions, it was submitted that the comparison was almost certainly illegitimate.
(3) Knowledge of Mr Agresta – CRSM submitted that in his evidence Mr Agresta did not want to take responsibility for the decision to present Scaramanga ratings to CRSM as a measure of improvement of the CDO portfolios. In relation to his evidence CRSM stressed that Mr Agresta denied that it was his idea to use the ratings, and claimed that he was not keen to use them and only did so because he was asked to use them by Mr Ferrario. His response to the suggestion that he must have realised that his use of Scaramanga ratings was misleading was to claim that he “was never interested to understand this rating” and “spent as little time as possible to understand these ratings” – despite his having already acknowledged that he had had a meeting with Mr Scaramanga to understand the ratings and that Mr Scaramanga would have explained how the model operated. Mr Agresta’s claims that he did not properly grasp the meaning of Scaramanga ratings or understand their limitations are implausible, given his manifestly high level of technical expertise, and are contradicted by the evidence that the model was fully explained to him.
CRSM submitted that it is likely that the use of Scaramanga ratings was in fact (contrary to his evidence) Mr Agresta’s idea. Thus, Barclays’ internal email chain on 20-21 April, in which Mr Scaramanga’s model is first mentioned, starts with an email from Mr Scaramanga to Mr Agresta (at 14:19 on 20 April) attaching a list of names showing their ratings according to Mr Scaramanga’s model. Mr Agresta forwarded this email to Mr Ferrario at 17:02. At 17.56 Mr Bechet sent to Mr Agresta some “missing” Scaramanga ratings for Como I and Como II. Then at 8:24 on 21 April Mr Ferrario contacted Mr Scaramanga to ask to speak to him to discuss his methodology. This sequence of events – as well as the fact that it was Mr Agresta who formulated all the other aspects of the restructuring proposals and the fact that Mr Ferrario is less likely to have been aware of recent technical research strongly suggests that it was Mr Agresta who thought of using the Scaramanga model and drew it to the attention of Mr Ferrario for this purpose.
If this is right, CRSM submitted that it is relevant to ask why Mr Agresta suggested using Scaramanga ratings as an additional criterion when, as he pointed out in evidence, the effect was to make his job as a structurer “even more labour-intensive”, as it represented an additional constraint to take into account in the portfolio construction, and that the obvious answer is that he saw such ratings as a useful means of presenting to CRSM as an improvement in portfolio quality substitutions which would increase the break-even spread.
Even if that is wrong and Mr Agresta was only “a passive user” of the Scaramanga ratings, it was submitted that he at the very least acquiesced in their use and was actively involved in drafting or helping to draft a series of communications to CRSM which he must have known were misleading (or at any rate been reckless as whether they were misleading or not).
In response to these points, Barclays repeated the general points made above, which I have accepted, about Mr Agresta’s understanding of and limited role in the restructuring.
Further I make the following findings about the Scaramanga model:
The QRV was a useful tool to identify buying and selling opportunities which could therefore be meaningfully used in the construction of portfolios (and also in the monitoring of substitutions). This was Mr Scaramanga’s evidence. Further, Mr Nasr accepted that the Scaramanga methodology would identify names that were likely to perform better, in credit spread terms, compared to others at a similar level of spread.
Mr Scaramanga (who was the Barclays expert on it) regarded it as an appropriate objective criterion which could be used for the purposes of the restructuring. That is why he was happy for it to be used.
In relation to the short term point, although Mr Scaramanga did not regard the QRV as something that could be used to predict the possibility of default of an entity “over a period of years”, that means only that it was not a predictor of long-term default risk. It did, however, contain relevant metrics for the assessment of short-term credit risk because (as he said) it “provided a good entry point for assessing exposure to the relevant entity”; and because it enabled the investor “to identify superior risk/return opportunities in relation to companies”. These considerations are not irrelevant to CRSM as a long term investor because:
Even a long term investor will be assisted, at the moment of selection, if he buys relatively cheaply compared to relatively expensively. He will be assisted by the identification of “good buying opportunities”. Further, it was in particular a useful metric in that context when combined with others. As Mr Scaramanga said: “I would say that certainly in the building of a credit portfolio, my model will not have been enough for a long-term investment in a credit portfolio, but I think it was a good complement to other measures, like credit rating.”
Although Mr Scaramanga accepted that his model was not a predictor of long term probabilities of default, that does not mean it was no use over significant periods. In fact, over a one year period a buy and hold investor following the indicator would show a return in excess of the market (as was explained in the documents provided to CRSM).
Even a long term investor will have short term horizons, especially when he is a replacement selector, as CRSM was. The short-term operation of the QRV was useful to CRSM in its role as replacement selector to enable it to consider when to substitute in or out certain credits. Knowing the initial Scaramanga scores would be useful in this regard.
In relation to the spread bucket point, Barclays pointed out that it was not pleaded as a ground of falsity and that it had not been put, or at least clearly put, to Mr Agresta. I agree with Barclays. The point is accordingly not open to CRSM. In any event:
The Scaramanga methodology does operate by identifying the richness or cheapness of entities compared to their spreads. It is also correct that its +2/+1/0/-1/+2 scores are allocated by reference to entities within the particular spread category.
However, the scope of comparability is relatively broad. As Mr Scaramanga explained, “it does not aim to directly compare high yield vs investment grade names”. But that does not mean that comparisons within the high yield category, on the one hand, or comparisons within the investment grade category, on the other hand, are meaningless. Mr Scaramanga’s research demonstrates that his methodology was conceived as giving meaningful information across wide spread ranges.
Consequently, when, as here, a list of particular proposed substitutions is advanced, with a list of Scaramanga scores, a recipient can, if he wishes, check those substitutions against other names within the same spread category. This will be easy to do since the spreads will be available to him, and he can also check comparator spreads on, for example, Bloomberg. The QRV will then give him a relative fine grained indication of the richness or cheapness of the name compared to other names within that category which he specifically considers. Indeed, information enabling CRSM to do precisely this without even having to do its own research was provided in Mr Scaramanga’s roundups which Mr Buoncompagni saw. Similarly, Mr Scaramanga’s bespoke analysis of 31 May provided all the spread data and the QRV scores and made recommendations to buy or sell.
Even if the recipient does not check recommended names against the relevant spread bucket, he will still know that the QRV is giving him valuable information. The list of proposed substitutions will, necessarily contain names across a range of spreads (as would have been obvious, and as was also clear from the spread data available to and provided to CRSM). The recipient will know for any name proposed, by reference to its QRV score, that it is rich or cheap compared to its spread category. He will thus know, in every case, that if he follows the QRV’s recommendation he will be getting a better name than one of similar spreads (and he will know that the other candidates he might have considered would in many cases have had similar spreads). He will not know whether there are names outside that spread category which might be better or worse than each particular name. But he will also know, even without comparing each high scored name to the spread category on an individual basis, that if his portfolio contains a higher proportion of high scored names, it will systematically be improved by reference to the QRV.
The fact that application of the QRV is meaningful as a predictor of quality in the aggregate, even without discrimination between spread category, is illustrated by Mr Scaramanga’s research paper, justifying the QRV. The back test he performed against market performance was performed on an aggregate basis across all spread categories but nevertheless showed that positive QRV scores were correlated with superior performance. Another example is given by how he himself deployed his scores, in one of his research papers, across a wide range of spreads (min 9, max 206) with the conclusion that: “In total, 12 of the 17 firms had a -1 or -2 rating as of February 2004. By choosing not to invest in the -2 category, investors would have avoided 41% of the firms that had the largest spread-widening vs 20% with a random selection.” He therefore, did not consider that the “spread bucket” point prevented the QRV from providing useful information to investors across a wide portfolio, even without definition of the relevant spread buckets.
The “spread bucket” feature therefore did not prevent the QRV being a useful buy or sell indicator. It was for that reason that Mr Scaramanga was perfectly willing to provide to CRSM his bespoke analysis making “overweight”/ “underweight” recommendations on the basis of the QRV principles.
Against the background of those findings I shall consider each of the points made by CRSM.
(1) Limitations of the Model - It is correct that these are limitations in the model. However, for the reasons outlined above, neither means that the model is not a useful buy or sell indicator when building or changing a portfolio even in the context of a long term investment. The QRV is therefore a useful and relevant indicator, even once its constraints are taken into account. It is useful both for monitoring substitutions and for selecting initial portfolios.
(2) Misrepresentations - The main presentation of the Scaramanga model was made in the QRV presentation, not Mr Agresta’s emails. That presentation, which Mr Ferrario went through at the 22 April meeting, was not criticised by CRSM and was the basis on which CRSM decided it was a tool it would like to use (because, in particular, Mr Buoncompagni decided it would be helpful and useful). The presentation, and Mr Scaramanga’s research papers (which Mr Buoncompagni saw), made clear that it “represents a forward looking P&L risk/return measure of holding a five year credit over the next six months”. It also went on to explain, in the attached graphs, its degree of effectiveness for the future. As to spread buckets, it explained clearly that it “signals those firms in a desired spread range that are expected to have the best returns per unit of risk”. The limitations in the model were therefore expressly identified.
(3) Knowledge of Mr Agresta – Although he had a familiarity with the model from his discussions with Mr Scaramanga I am not satisfied that he had any particular awareness of the spread bucket point, nor was this explored in evidence with him.
It is correct that Mr Agresta was not particularly interested in the Scaramanga model and regarded its inclusion as a criterion as generally making his job more difficult. This serves to emphasise how his focus was on the practicalities of the job of portfolio selection rather than how this was or might be perceived by CRSM, as to which he had no direct knowledge. I reject the speculative suggestion that its inclusion was his idea. He was not a salesman and it was not for him to decide matters such as this.
For all these reasons I reject CRSM’s case that anything said by Mr Agresta about the Scaramanga model involved any dishonesty or deception on his part.
I therefore reject CRSM’s case of fraud against Mr Agresta in respect of the Restructuring Representations and, in the light of the findings set out above, I find as follows:
Mr Agresta did not understand that he or Barclays was making any of the representations alleged by CRSM.
Mr Agresta had no intention to mislead in relation to the alleged misrepresentations and concealments.
In relation to negligence, in respect of the second and third alleged misrepresentations by the end of the trial CRSM’s case rested on proof of deception. If that is wrong and negligence is still in issue I find that Mr Agresta had reasonable grounds to believe in the truth of any representation he can be understood actually to have made. In particular:
He had reasonable grounds to believe that the notes were being improved by the restructuring, both generally, and according to the parameters defined. The forced default analysis explained exactly how cross-subordination was improving the notes and it was not criticised. Mr Agresta was entitled to rely on it.
He had reasonable grounds to believe that credit ratings and Scaramanga ratings were objective measures and (if relevant) appropriate and useful measures. Credit ratings are obviously objective (and useful). So far as Scaramanga scores were concerned, he could reasonably rely on the fact that Mr Scaramanga, the expert, thought they were relevant to the assessment of CDOs.
In respect of the “cost zero” representation, if, contrary to my findings and Mr Agresta’s belief and understanding, this involved a representation that Barclays would not make any profit from the restructuring he had no reasonable grounds for so believing since he knew that some profit was intended.
The case against Mr Ferrario
As CRSM accepted, its only pleaded case of fraud against Mr Ferrario in relation to the restructuring representations related to the alleged representation that no profit would be made from the restructuring. In Closing they applied for permission to amend to allege fraud against Mr Ferrario in respect of the two other alleged representations. They relied in particular upon the fact that it was already alleged that such representations were made negligently by Mr Ferrario and that the fraud case had been explored in cross examination of Mr Ferrario so that he had had a fair opportunity to deal with it. It was submitted that it was in the interests of justice that the amendment be allowed so that the Court can decide the case on the merits.
Unsurprisingly Barclays took strong objection to this very late application to amend. I am satisfied that it is far too late to introduce an amendment of this kind and that it would be unfair to Barclays and Mr Ferrario for this to be permitted. CRSM’s case on fraud was only introduced shortly before trial. CRSM successfully resisted Barclays’ application to adjourn the trial in the light of the lateness of the amendment. In such circumstances it was particularly important that it was absolutely clear what fraud case Barclays was required to meet. CRSM confirmed in Opening that the case was as expressed in their Written Opening and Barclays approached the trial on that basis. Having done so it would be unfair to allow CRSM to introduce a wider case now. It is obviously important that a party and a witness against whom fraud is alleged should know what is being alleged before (at the latest) giving evidence. It is wholly unsatisfactory for such a case to be sprung on the witness for the first time in cross examination and it is no answer to say that no objection was taken to such questioning at the time. Barclays were entitled to proceed on the basis that the only case being advanced was that pleaded and confirmed in Opening. Nor is it fair on the witness for this to be done – having to address a fraud case for the first time in cross examination is not a “fair opportunity” of dealing with the case.
I am accordingly satisfied that serious prejudice and unfairness would be caused if the amendment was permitted and in the exercise of my discretion I refuse permission to amend. I should add that I would in any case have rejected CRSM’s case on the facts, largely for the reasons why the like case was rejected as against Mr Agresta. It follows that the only case that needs to be considered as against Mr Ferrario is in relation to the “cost zero” representation.
Mr Ferrario, like Mr Agresta, did not understand the reference to “cost zero” in Mr Agresta’s email of 22 April 2005, or in the email of 27 April 2005, as any more than a statement that CRSM would not be charged cash for the restructuring changes. As he said:“…Antonio also explained that CRSM would be able to make some changes without having to pay any more money”. This is consistent with the use of the phrase in the 27 April 2005 email.
Mr Ferrario also did not believe himself to have represented that Barclays would not be making profits by his reference to bid/offer costs. He understood himself to have been thereby referring to the bid/offer costs of the restructuring as a whole. He did not understand that as inconsistent with any notion that Barclays was going to be making profits on the transaction. As he said when the point was put to him in cross examination: “I disagree, my Lord. For any person that has worked on the financial markets and has traded even simple instruments, it is clear that bid/offer costs would impact on the mark to market of the transactions. It is very clear that it could originate a profit for the dealer you are dealing with, and CRSM understood at the time that we were spending a lot of time and effort and would have expected us to make some profits on the transactions … on the restructuring … It was implied by discussing that we were charging bid/offer costs, my Lord. As I said, we didn’t get too much into the detail, also because I wouldn’t have had the details. I wasn’t aware of the entirety of those bid/offer costs. It would have been impossible for me to enter into details” “Dealers capture, in executing transactions, bid/offer. That is part of the nature of our business. If you capture bid/offer costs for the client, you were implicitly making money …. Their mark to market valuation is going to go down by those bid offer costs.”
I accept Mr Ferrario’s evidence on this issue and find that he did not understand that he or Barclays was making any representation that it would make no profit or be no worse off as a result of the restructuring. It follows that even if, contrary to my findings, such a representation was made, Mr Ferrario did not understand that any such representation had been made and did not intend that it should be understood in that sense.
I further find that Mr Ferrario had no intention to mislead in this regard or indeed with respect to any aspect of the restructuring.
As with Mr Agresta, if, contrary to my findings and Mr Ferrario’s belief and understanding, a representation was made that Barclays would not make any profit from the restructuring Mr Ferrario had no reasonable grounds for so believing since he knew that some profit was intended.
VIII. Falsity
The Purchase Representations
I have already found that if, contrary to my findings, a representation was made that the Notes would have a very low risk of default, it was one of opinion and/or expectation and/or belief. I shall assume that if it was a representation of opinion it is to be inferred that it was made on reasonable grounds. On that basis I shall address the issue of falsity.
As to whether Barclays, and in particular Mr Ferrario and Mr Agresta, did have the expectation or belief that the Notes would have a very low risk of default I have already found that they did. As Mr Ferrario and Mr Agresta said (consistently with Ms Duhon’s evidence) they and Barclays took the AAA rating at face value, “as a fact”. It was also their evidence, which I accept, that they believed these were good, safe products.
As to whether there were reasonable grounds for such an opinion to be held, I find that there were was for the reasons set out in paragraph 371 above.
I accordingly find that if, contrary to my findings, a representation was made that the Notes would have a very low risk of default, it was true.
The Restructuring Representations
Misrepresentation of intention to profit
If contrary to my findings, such a representation was made, and if, contrary to my findings it was a continuing representation at the time that the restructuring contract was made, it would have been false. By the time that the contract was made this may well have been a representation of fact since it would be likely to be established by then that some profit would be made. Even if it was only a representation of expectation or belief it would be untrue since Barclays always expected and intended that some profit should be made.
Misrepresentation of the Selection Criteria
As I have already found no representation was made as alleged and there was no deception such as to render what was represented untrue.
Misrepresentation relating to Scaramanga model
As I have already found no representation was made as alleged and there was no deception such as to render what was represented untrue.
VII. Inducement/reliance
It was common ground that a “but for” test for reliance applies in relation to a claim under s.2(1) of the Act 1967. It follows that if a representee would have entered into the transaction even had the misrepresentation not been made, there is no inducement. It is not sufficient for the representee merely to show that he was supported or encouraged in reaching his decision by the representation in question.
CRSM contended that there is a weaker test in fraud and that the misrepresentation need only be a factor relevant to a decision. Although Barclays disputed this I shall address the issue on the basis that CRSM is correct.
Barclays stressed that CRSM had not adduced evidence from the two key individuals. There was no evidence from Mr Sapignoli, whose analysis of the information provided by Barclays appears to have been crucial in CRSM’s decision-making process in relation to the initial purchases of the Notes. Nor was there evidence from Mr Fantini, the Chief Executive Officer of CRSM, who signed the purchase agreements and the restructuring agreement on behalf of CRSM, and was the ultimate decision-maker.
Whilst I accept that the absence of such evidence makes CRSM’s task of proving inducement and reliance more difficult, I do not consider it to be fatal. CRSM had witnesses who were able to and did give evidence relating to the decision making process in relation to all the transactions in question.
The Purchase Representations
CRSM submitted that it was clear from the evidence of its witnesses that they understood what they were told as to the CDO2s being ‘AAA’ and through the information contained in the fact sheets, as meaning that the CDO2s were investments with a very low risk of default; and that they relied on that understanding when CRSM agreed to buy the Notes.
CRSM further submitted that in so far as actions may be said to speak louder than words, the clearest possible confirmation that CRSM understood that the CDO2s had a very low risk of default is that it sold the Notes to its customers (and, in Mr Montanari’s case, advised his parents, as well as other family and friends, to buy interests in the Notes).
Conversely, when CRSM came to realise the CDO2s were not as low risk as it had believed, it took steps to buy back (at par) all the participations sold to its customers, and to exit from the investments where it was able to do so at an acceptable price.
Further, Mr Ferrario confirmed that the two principal matters on which CRSM was relying when it decided to invest in the CDO2 Notes were the AAA rating and the information contained in the fact sheets that Barclays prepared. He also said that CRSM’s representatives looked carefully at the fact sheets in order to assess the risk of default and that they “expected exactly the risk of default that was clearly showed to them and represented to them in the scenario analysis”.
I accept that CRSM’s evidence establishes that reliance was placed upon the AAA rating of the Notes and the scenario analysis. I also accept that CRSM regarded the notes as having a very low risk of default and that this was an important factor in the decision to buy the Notes. However, the key issue is whether CRSM relied on a representation by Barclays to that effect. Reliance upon the AAA rating and the scenario analyses in themselves is not sufficient. It was CRSM’s case that Barclays was making a representation that the Notes would have a very low risk of default and it is that alleged representation upon which reliance must be established.
In relation to Mr Montanari his evidence included the following:
In his witness statement he stated that whilst CRSM trusted Barclays, he personally did not perceive Barclays as an advisor. He said that in relation to the transactions with Barclays, it was the AAA rating that made CRSM feel comfortable with Barclays’ proposal; “As this rating was the highest and therefore indicated that the relevant investment was the safest that one could buy, I took it to mean that it was almost impossible for the investment to which it applied to fail”. He did address what CRSM would have done had the express Purchase Representation not been made, but he did not say there that any such representation was made to him or that he personally relied on it, or conveyed it to anyone else.
In cross examination it became very clear that the key matter as far as Mr Montanari was concerned was the AAA rating rather than anything that might have been said by Barclays. For instance: “Q: In your activities on behalf of the bank, you placed trust and reliance in those credit ratings; is that right? A: Yes, we took it into consideration. Q: You were, in fact, in general terms, happy to buy on the basis of a good credit rating? A: Yes.”;“Q: The AAA rating was something which you found satisfactory as confirmation of low credit risk in the product; correct? A: Correct Q: You believed, based on your experience, that a AAA rating indicated a safe and secure investment; correct? A: Exactly” ; “Q: But you did realise that you had the credit risk on the CDO note, that if there were enough defaults in the reference portfolios, the protection built into the notes could be lost? A: I didn’t know the structure of the CDO, and when I was told that it was a AAA -- it had a AAA rating, I believed that there were no immediate default risks.” Finally, and most significantly he stated as follows in re-examination; “Q: When you had meetings with Mr Ferrario at which the ARLO notes were discussed, did he say anything to you about the credit risk or risk of default of the ARLO notes? A: No, we never discussed it because AAA to us was synonymous of that tranquillo, safe and transparent that I was looking for.”
In relation to the Carnivale Note it was apparent from his evidence that Mr Montanari relied upon Mr Sapignoli to read and digest the documents that Barclays sent. He relied on Mr Sapignoli to conduct the analysis and he left Mr Sapignoli to get on with it, which included discussing matters with Mr Ferrario as he (Sapignoli) might feel appropriate, as illustrated by the following exchange in evidence: “Q: To be fair to you, Mr Montanari, Mr Ferrario will say that he had the impression that you did not pay much attention to any of these details because you were leaving it to others. A: That is true because I trusted the people who worked with me.”;…. “I always entrusted Mr Sapignoli with the evaluation of the notes.”
Mr Montanari did not give any evidence to the effect of ‘Mr Sapignoli reviewed a particular document, or had a particular conversation, and told him that Barclays said that the CDO notes were secure and had an extremely low risk of default’.
Although Mr Montanari stressed that he trusted Barclays and that he believed that the Notes had a very low risk of default I am not satisfied that he gave any clear or satisfactory evidence to the effect that he understood Barclays or Mr Ferrario to be saying (independently of the AAA rating) that the Notes had a very low risk of default or that he relied on any statement by them to that effect. He relied on the AAA rating and Mr Sapignoli.
In relation to the later transactions CRSM relied primarily on the evidence of Mr Micheloni. As to that:
Mr Micheloni said little in his witness statement about the basis of any decision to proceed with the further purchases. He indicated that after Mr Sapignoli departed in September 2004, Mr Montanari just left him to fill those shoes, whilst Mr Montanari was largely out of the office. His (incorrect) understanding was that CRSM had already agreed to purchase the second batch of Notes and he was never asked to consider the rationale for the second purchase and did not do so.
There is no evidence, therefore, that Mr Micheloni personally relied on any of the Purchase Representations, whether made afresh to him by Mr Ferrario orally or in new documents provided by Barclays, or as continuing representations made originally in connection with the Carnivale Notes.
In re-examination Mr Micheloni said that he did not recall looking at the historical analysis provided by Barclays’ Fact Sheets or that any part of the Fact Sheet caused him to form any idea of the probability of default. CRSM submitted that this was not consistent with the documents and that Mr Micheloni must have reviewed the fact sheets and the scenario analysis. Even if that were so, his oral evidence indicates that he did not place any particular reliance upon the fact sheets, still less upon representations made thereby by Barclays, as opposed to the facts set out therein.
In effect Mr Micheloni was executing transactions and trying to follow as closely as possible what had gone on before. There was no satisfactory evidence that he was relying on anything that Barclays had told him.
Mr Morolli’s evidence was that he had nothing whatsoever to do with the decision to purchase the Notes.
I accordingly find that CRSM has not, therefore, shown that it relied on the Purchase Representations, if made, in any conclusion that the CSO2 notes had a very low risk of default, or that the decision to invest in the Notes was made upon the basis of any such conclusion. There is little, if any, evidence about why it decided to buy Cernobbio, Bellagio and Tremezzo, other than that it chose to follow the path set when Carnivale was purchased. I find that CRSM has not proved that the investments would not have gone ahead if the Purchase Representations had not been made or even that the Purchase Representations were a factor in its decision.
The Restructuring Representations
CRSM’s witnesses gave evidence that, in agreeing to Barclays’ proposals for the restructuring of the CDO2s, they relied on Barclays and that they were acting in the belief, based on what they had been told by Barclays, that the restructuring was beneficial to CRSM and improved the default risk of the CDO2s. Again, however, what needs to be proved is reliance upon the individual representations allegedly made.
In relation to the alleged misrepresentation that no profit would be made CRSM submitted that this was clearly an important matter for CRSM, which is why Mr Morolli expressly raised it and sought assurance on the point from Mr Ferrario. Had I accepted that evidence that would be a strong point. However, I rejected that evidence and found that the only relevant statements were those made in the 22 and 27 April emails and Mr Ferrario’s statement about bid/offer costs (as to which Mr Micheloni had no recollection). If these amounted to a representation that no profit would be made I am not satisfied by the evidence that it was relied upon. If such reliance had been placed one would have expected it to be raised at some stage thereafter and to feature in later documents; it does not.
In relation to the alleged misrepresentation of the selection criteria I accept that the evidence shows that CRSM, and in particular Mr Buoncompagni (upon whom others relied), did rely on those criteria as positive reasons for going ahead with the restructuring. On CRSM’s case a truthful account of how the names proposed for substitution had been selected would have had to disclose that the real criteria applied was to find names with high spreads for their ratings with the aim of increasing the break-even spread on the CDO2s. If CRSM had been told that (or alternatively told nothing at all about the criteria followed in proposing substitutions), I find that CRSM would not have proceeded with the restructuring.
CRSM accepted that it is less certain what the effect of the misrepresentation relating to the Scaramanga model was, considered on its own. However, I am satisfied that CRSM did place reliance on the Scaramanga ratings. As Mr Buoncompagni stated: : “I deemed these criteria -- that is to say the rating, lower presence of high-return names and the Scaramanga model, were deemed by Barclays and by ourselves as an indication of a better quality portfolio.” On CRSM’s case Barclays deliberately obscured the limitations of the model and a truthful account of the use of the Scaramanga model would have highlighted those limitations. Had that been done I find that CRSM would have rejected the use of the model but I am not satisfied that this would have meant that the restructuring was not proceeded with. I accordingly find that this alleged misrepresentation was relied upon and did encourage or support CRSM’s decision to restructure. I do not, however, find that the restructuring contract would not have been entered into had such a misrepresentation not been made.
X. The Contractual Defences
Barclays contended that CRSM is precluded by certain terms of, respectively, the purchase contracts and the restructuring contract from advancing its case of misrepresentation under the Act (but not in fraud) in relation to each of those contracts.
Barclays relied on recent authorities to the effect that parties to a contract may agree that a particular state of affairs is to be the basis upon which they are contracting, regardless of whether or not that state of affairs is true, and that such an agreement may give rise to a “contractual estoppel”, precluding the assertion of facts inconsistent with those which have been agreed to form the basis of the contract. The most recent of these authorities is the decision of the Court of Appeal in Springwell Navigation v JP Morgan Chase Bank[2010] EWCA Civ 1221.
CRSM submitted that none of the provisions relied upon by Barclays purports, or is effective, to negate the prior making of any representation by Barclays and that no question of any defence of contractual estoppel therefore arises in this case.
CRSM further submitted that there is no inconsistency between (a) an expectation (or even agreement) that a purchaser will make its own independent assessment of the risks and (b) an expectation that the purchaser is entitled to rely in making its assessment on information about the proposed transaction provided by the selling bank. The purchaser will naturally and rightly rely on such information unless the document in question makes clear that no reliance should be placed on it. CRSM submitted that none of the terms of the purchase letters have such an effect.
The terms relied on by Barclays were the “Non-reliance”, “Assessment and Understanding” and “Terms and Conditions” clauses which CRSM had agreed “we hereby represent, warrant and covenant”.
The Authorities
Two principal authorities were relied on by Barclays in support of its arguments that the terms of the purchase contracts in this case had the effect of precluding CRSM’s claim. These were Peekay v Australia & New Zealand Banking Group [2006] 2 Lloyd’s Rep 511 and Springwell Navigation v JP Morgan Chase Bank [2008] EWHC 1186 (Comm) (Gloster J);[2010] EWCA Civ 1221 (CA).
In Peekaythe first claimant (Peekay), a company controlled by the second claimant (Mr Pawani), bought an investment product from the defendant (ANZ). The product was a note linked to Russian Government bonds known as GKOs. In a telephone conversation Mr Pawani was given (as the judge found) a “rough and ready” description of the product by an employee of ANZ which led him to believe that, if it bought the product, Peekay would acquire an interest in a GKO. Subsequently, Mr Pawani was sent documentation consisting of final terms and conditions (FTCs) for the investment to which a Risk Disclosure Statement was attached which he was asked to sign. These documents made it clear that the investment being offered was a derivative, consisting of a deposit linked to GKO bonds, rather than a share in the bonds themselves.
The Risk Disclosure Statement included the following:
“You should also ensure that you fully understand the nature of the transaction and contractual relationship into which you are entering … The issuer assumes that the customer is aware of the risks and practices described herein, and that prior to each transaction the customer has determined that such transaction is suitable for him.”
Immediately above the space for signature there appeared:
“[Client] confirms it has read and understood the terms of the Emerging Markets Risk Disclosure Statement as set out above.”
Mr Pawani signed the documents and initialled every page of them but the trial judge found that he did no more than glance at their contents, with the result that he did not appreciate that he was buying a structured product. (He also did not read an Indicative Term Sheet previously sent to him which accurately described the nature of the investment.) On these facts the judge concluded that Peekay was induced to make the investment by the misrepresentation made in the initial telephone conversation and was entitled to damages under s.2(1) of the Act.
The Court of Appeal allowed an appeal by ANZ on the issue of causation. Moore-Bick LJ (with whose judgment Lawrence Collins J and Chadwick LJ agreed) held, in effect, that the chain of causation was broken by Mr Pawani’s signature of the documents without reading them and that Peekay was induced to enter into the contract not by what had been said previously on the telephone, but by Mr Pawani’s own assumption that the investment product to which the documents related corresponded to the description he had previously been given (see para 52).
The Court of Appeal went on to hold that ANZ was also entitled to succeed on the alternative basis that the signature of the Risk Disclosure Statement gave rise to a contractual estoppel (see para 54). Moore-Bick LJ said:
“There is no reason in principle why parties to a contract should not agree that a certain state of affairs should form the basis for the transaction, whether it be the case or not. For example, it may be desirable to settle a disagreement as to an existing state of affairs in order to establish a clear basis for the contract itself and its subsequent performance. Where parties express an agreement of that kind in a contractual document neither can subsequently deny the existence of the facts and matters upon which they have agreed, at least so far as concerns those aspects of their relationship to which the agreement was directed. The contract itself gives rise to an estoppel” (para. 56)
…..
“I can see no reason in principle why it should not be possible for parties to an agreement to give up any right to assert that they were induced to enter into it by misrepresentation, provided that they make their intention clear, or why a clause of that kind, if properly drafted, should not give rise to a contractual estoppel….” (para. 57)
Moore-Bick LJ concluded (at para 60) that, by signing the declaration that he had read and understood the Risk Disclosure Statement, Mr Pawani agreed contractually on behalf of Peekay that Peekay understood the nature of the transaction described in the FTCs, with the result that Peekay could not assert that it was induced to enter into the contract by a misunderstanding of the nature of the investment derived from what ANZ’s employee had said about the product some days earlier.
Chadwick LJ in his concurring judgment characterised the estoppel (at para 70) somewhat differently – as being that “Peekay could not be heard to say that Mr Pawani had assumed that the FTCs which he had signed on its behalf did not need to be read and understood”.
In Springwellthe principal claim advanced by the claimant was that Chase had been engaged to give general investment advice to it, and that this gave rise to duties of care in contract and tort together with fiduciary duties.
In addition, more specific allegations were made of negligent misrepresentation, under s. 2(1) of the Act and under the Hedley Byrneprinciple. The three broad categories of alleged misrepresentation (of which the first and second were pursued on appeal) were to the effect that (1) the GKO investments at issue were a “conservative” investment, (2) they had a high degree of liquidity and (3) investment in them involved no currency risk.
The trial judge, Gloster J, held on the facts that the alleged misrepresentations had not been made, and in any event would not have been actionable (there being no basis, on the facts found, for concluding that Chase’s salesman’s expressions of opinion involved any implied representation about his having objectively reasonable grounds for his views); and these findings were upheld on appeal.
Since full argument had been heard and one point of general importance arose, the Court of Appeal went on to make findings about the effect on Springwell’s misrepresentation claims of a number of provisions in the contractual documentation. These included the following:
“In addition, the Holder has not relied on, and acknowledges that neither CMSCI nor CMIL has made, any representation or warranty with respect to the advisability of purchasing this Note.” [Section 6(c) of the GKO Note]
“CMSCI has not made any representations and warranties whatsoever, either expressed or implied, including, without limitation, any representation or warranty as to (I) the due execution, legality, validity, adequacy or enforceability of the Designated GKO Assets or any other Transaction or any document relating thereto; (ii) the financial condition of any party to a Transaction or the performance of any party to a Transaction of any of their obligations related to any Transactions or that it has made, or will make, any inquiries concerning any such parties; and (iii) as to any tax matters related to the Transactions or investments in S Accounts; and (iv) as to the content of or the applicability of the S Account Rules.” [Section 5(e) of the GKO Note]
“Neither CMB nor CMIL nor CMp are required to give you investment advice generally or in relation to specific investments, make any enquiries about, or to consider, your particular financial circumstances or investment objectives. By placing an order with CMB or CMIL, or CMp, you represent that you are a sophisticated investor, are purchasing the instrument concerned for your own account for investment purposes and not with a view to any distribution thereof, and that you have independently, without reliance on CMB or CMIL or CMp, or any associated person, made a decision to acquire the instrument having examined such information relating to the instrument and the issuer thereof as you deem relevant and appropriate. You have represented to CMB and CMrt, and CMp, and therefore they have assumed that, you are fully familiar with and able to evaluate the merits and risks associated with such instruments and any consequence of these instruments forming part of a portfolio of investments and are able to assume the risk of loss associated with such instruments. You should therefore consider whether an instrument is appropriate in your particular financial circumstances or in the light of your investment objectives. Neither CMB nor CMIL nor CMp are liable for any loss which you may incur arising out of any investment decision made by you in consequence of any service contemplated in this letter unless such loss is caused by its gross negligence or wilful misconduct. Neither CMB nor CMIL nor CMp shall be obliged to provide any dealing services to you and may within its absolute discretion decline to do so at any time.” [Clause/Paragraph 4 of the DDCS Letters]
“When providing you with any circular, information memorandum, investment advertisement, published recommendation or any other written or oral information regarding any instrument or investment opportunity neither CMB nor CMIL nor CMp, will have taken any independent steps to verify the document or information and no representation or warranty, express or implied, is or will be made by either CMB or CMIL or CMp, their representative officers, servants or agents or those of their associated companies in or in relation to such documents or information nor will CMB or CMIL or CMp or any of their associated companies be responsible or liable (save to the extent required under the applicable law, rules or regulations) for the fairness, accuracy or completeness of such documents or information.” [Clause/Paragraph 6 of the DDCS Letters]
The Court of Appeal, focussing initially on the first clause quoted above (“Section 6(c)”), held that:
Section 6(c) applied to Springwell as holder of the notes (paras. 132-140).
The statements of Moore-Bick LJ in Peekay about contractual estoppels were consistent with principle and authority, and represented the law (paras. 141-169).
Applying Peekay, the effect of Section 6(c) was to bind Springwell to its statement or acknowledgment that Chase had not made “any representation or warranty with respect to the advisability of purchasing this Note” (para. 170).
Similarly, Springwell must be bound by the terms of Section 5(e), “which means that it accepts that CMSCI has not made any representations or warranties of the kind set out there” (para. 170).
The effect of Sections 6(c) and 5(e) was, subject to any argument based on section 3 of the Misrepresentation Act 1967 and the Unfair Contract Terms Act 1977, first, that Springwell was contractually estopped from contending that there were any actionable representations made by Chase on which Springwell could base its current claims. Secondly, those provisions pointed strongly against any statement by Chase’s salesman being actionable (para. 171).
Clause 4 of the DDCS letters meant that Springwell was contractually bound by the representations that it was a sophisticated investor, familiar with and able to evaluate the merits and risks associated with the instruments with which those letters were concerned and able to assume the risk of loss associated with such instruments (para. 172).
Clause 6 of the DDCS letters meant Springwell accepted that no express or implied representations were made concerning any written or oral information regarding any instrument or investment opportunity offered by the terms of the letters (para. 172). Accordingly Springwell had agreed that statements by Chase’s salesman in the numerous telephone calls which had taken place were to be treated as expressions of opinion, and not to be susceptible to scrutiny with a view to seeing if they constituted potential actionable misrepresentations (para. 173).
The authorities accordingly establish that:
It is possible for parties to agree that one party has not made any pre-contract representations to the other about a particular matter, or that any such representations have not been relied on by the other party, even if they both know that such representations have in fact been made or relied on, and that such an agreement may give rise to a contractual estoppel.
If a term is to be construed as having this effect (and thereby prevent from arising the ordinary consequences which would otherwise follow as a matter of law) clear words are necessary – see Peekay para. 57; Board of Trade v Steel Brothers & Co. Ltd. [1952] 1 Lloyd’s Rep. 87 at p95.
Whether or not a clause or collection of clauses has this effect is a matter of construction of the contract.
The principle may not apply where there has been a misrepresentation as to the effect of the contractual documents which give rise to the estoppel – see Peekay para. 60; Springwell para. 166.
The cases provide clear examples of clauses which will be construed as having the effect of precluding claims for misrepresentation - see, for example, in the banking context the provisions in the Raiffeisen case (para. 229):
“RZB acknowledges and agrees that ... RBS and its Affiliates, officers, employees, agents, and professional advisers do not make any representation or warranty, express or implied as to, or assume any responsibility for, the accuracy, adequacy, reliability or completeness of any of the Confidential Information.”
“The contents of this Memorandum have not been independently verified. No representation, warranty or undertaking (express or implied) is made, and no responsibility is accepted as to the adequacy, accuracy, completeness or reasonableness of this Memorandum or any further information, notice or other document at any time supplied in connection with the Facility.”
As Christopher Clarke J said in Raiffeisen, those provisions “specify that no representation is made or that RBS does not make any. These provisions are unambiguous.”
The relevant clauses in Springwell also involved widely drafted provisions, including clauses stating expressly that no representations had been made in relation to the relevant matters, no representations had been relied on and no liability was accepted for documents or information provided. I agree with CRSM that those were the principal provisions which the Court of Appeal held would have defeated Springwell’s claims for misrepresentation, had those claims otherwise been good.
I also agree with CRSM that in so far as the Court of Appeal considered that the provisions of the DDCS precluded Springwell from denying that it was a sophisticated investor, familiar with and able to evaluate the merits and risks associated with the relevant investment, there is no indication that the Court of Appeal considered that that provision by itself would have defeated Springwell’s claims.
Armed with the guidance provided by the authorities I turn to consider each of the contractual provisions relied upon in this case.
The contractual clauses
Clause 5 (“Non-reliance”)
This provided as follows:
“5. Non-reliance. We are acting for our own account and have made our own independent decisions to enter into this letter and purchase the Notes and as to whether the Note purchase is appropriate or proper for us based upon our own judgment and upon advice from such advisors as we deem necessary. We are not relying on any communication (written or oral) from you as investment advice or as a recommendation to enter into this letter or to purchase the Notes, it being understood that information and explanations related to the terms and conditions of the Notes shall not be considered investment advice or a recommendation to purchase the Notes. No communication (written or oral) received from you shall be deemed to be an assurance or guarantee as to the expected results of the purchase of the Notes.”
CRSM submitted as follows:
The gist of Clause 5 (“Non-reliance”) is to make clear that Barclays is not providing investment advice or making an investment recommendation. An investment adviser’s responsibilities go well beyond a duty not to make misrepresentations: it has a positive duty to advise the customer as to the type of investment he ought to be making, and as to the characteristics of proposed investments. Moreover, “advising on investments” was a defined regulatory term with specific consequences (See definition in the FSA Glossary and Conduct of Business rule 5.1.). Similarly, a range of regulatory requirements applied to the case in which an investment business made a “personal recommendation” to a private customer (for example, the FSA’s conduct of business rule COB 5.3). Bearing in mind that these were Barclays’ standard terms, it is not surprising that Barclays would wish to define its role by reference to prevailing regulatory criteria, even in a document sent to a non-private customer.
It is, in any event, not CRSM’s case that it relied on Barclays’ statements as “investment advice” or as a “recommendation to ... purchase the Notes”. There is no inconsistency between a claim for misrepresentation and an agreement that Barclays has not acted as advisor or made a “recommendation” to purchase.
Equally, CRSM did make its own independent decision to purchase the Notes – it does not claim to have acted on advice from Barclays; but in making that decision, it was entitled to rely on the representations of fact which Barclays, as the structurer and seller of the Notes, had made to it.
As to the last sentence of Clause 5, CRSM’s claim does not involve any allegation that Barclays gave any assurance or guarantee as to the expected results of the purchase of the Notes. Rather, it is that Barclays made false representations of fact which were liable to mislead CRSM.
Barclays submitted that:
The effect of the first sentence was to make it the agreed basis of each purchase transaction, and it was CRSM’s contractual promise, that CRSM had made its own independent decisions to conclude the contract, and purchase the Notes, in question, and as to whether that Note purchase was appropriate or proper for CRSM based upon its own judgment and upon advice from such advisors as it deemed necessary (Clause A5). In context, the “independent decisions” CRSM promised it had made were decisions made independently, in particular, of anything Barclays may have said or omitted to say about the Notes. Thus Clause A5 is a complete answer to CRSM’s claims. Each of those claims proposes that CRSM was wrongly persuaded by Barclays to buy the Notes. But the agreed basis for the transaction, and CRSM’s covenant, was that in its decision to buy, CRSM had not been influenced by Barclays at all.
The effect of the second and third sentences was to make it the agreed basis of each purchase transaction, and it was CRSM’s contractual promise, that (a) CRSM was not relying on any communication (written or oral) from Barclays as investment advice or as a recommendation to conclude the contract, or purchase the Notes, in question, and (b) any information or explanation related to the terms and conditions of the Notes would not be considered investment advice or a recommendation to buy (Clause A5 again). This too defeats CRSM’s purchase claims. What Barclays may have said about the Notes was not, and was not to be regarded as, any basis for a decision to buy.
I agree with CRSM that the clause, construed as a whole, is focussing on investment advice and recommendations. It is in this context that the independent decision making referred to in the first sentence falls to be considered. It means independent of any advice or recommendations provided by Barclays. It does not expressly, necessarily or clearly go beyond that. It does not therefore, contrary to Barclays’ submission, apply to “anything which Barclays may have said or omitted to say about the Notes”.
The second sentence precludes reliance upon communications as being investment advice or recommendations In so far as the alleged representations are sought to be so relied upon I agree with Barclays that it has been contractually agreed that they may not be. In such a case CRSM would be contractually estopped from contending that actionable representations had been made thereby, whether under s.2(1) of the Act or for negligent misstatement.
Whether or not the communications are being relied upon as advice or recommendations depends upon the substance of the claim made. On the basis of the misrepresentation case as advanced in Closing I would not characterise any of the representations alleged as constituting investment advice or recommendations or as them being so relied upon. The Purchase Representation involves a statement relating to the risks of the investment. It does not purport to advise or recommend what CRSM should do in the light of that statement nor does CRSM seek to so rely upon it. The Restructuring Representations relate to the criteria to be applied in carrying out the restructuring. Similarly they do not purport to advise or recommend what CRSM should do in the light thereof nor do CRSM seek to so rely upon them. In my judgment the substance of the claim made does not involve reliance upon communications as advice or recommendations.
The third sentence prevents a communication received from being an assurance or guarantee of results. I agree with CRSM that its case does not assert or involve that such an assurance or guarantee was given and accordingly it is not covered thereby.
Clause 6 (“Assessment and Understanding”)
This provided as follows:
“6. Assessment and Understanding. We are capable of assessing the merits of and understanding (on our own behalf or through independent professional advice), and understand and accept, the terms, conditions and risks of entering into this Letter and of purchasing the Notes. We are also capable of assuming and we assume, the financial and other risks of entering into this letter and purchasing the Notes.”
CRSM submitted that:
This clause is typical of a provision designed to underline the point that no advisory duty is undertaken. Accordingly, Barclays had no duty to ensure CRSM understood the risks of the transaction, or to ensure that they did not render the investment unsuitable for CRSM.
The drafting of Clause 6 has regulatory overtones. The FSA’s Conduct of Business rule 5.4 required a firm not to make a personal recommendation to a private customer to invest unless the firm had taken reasonable steps to ensure that the customer understood the nature of the risks involved.
In any event, CRSM’s complaint is not that Barclays sold it a product without ensuring that it understood the risks. It is that, in forming an understanding of the risks, CRSM was actively misled by representations of fact made by Barclays which were untrue. Nothing in this clause purports to negate the making of any such representation or to exclude liability for any such representation. There is no inconsistency between forming one’s own assessment of the risks, and relying (in the process) on specific information provided by the selling as part of that process. A purchaser of an investor is entitled to do this unless clearly told, in terms, that he may not rely on the information provided to him.
Barclays submitted that the effect of this clause was that it was the agreed basis of each purchase transaction, and it was CRSM’s contractual promise, that CRSM (on its own behalf or through independent professional advice) (a) was capable of assessing the merits of and understanding the risks of concluding the contract, and purchasing the Notes, in question, and (b) in fact understood, accepted and assumed those risks. That is fatal to CRSM’s claims. None of those claims could survive a finding that CRSM decided to buy, having been competent to assess and understand, and having in fact understood and accepted, all the risks of doing so. But that was the agreed basis of each purchase, and CRSM promised to hold to that proposition. Or to put it another way, to uphold any of CRSM’s claims would be to allow it to require Barclays to bear (some of) the risks that CRSM took on when it purchased the Notes. But as between Barclays and CRSM, the agreed basis of the purchase, and CRSM’s promise, was that CRSM would bear all of those risks.
Barclays submitted that the fact sheets in the present case are the equivalent to the oral description provided in Peekay and that clause 6, which is similar terms to the Risk Disclosure Statement in Peekay, precludes any claim for misrepresentation arising out of the fact sheets.
In my judgment this puts the matter too broadly. The misrepresentation alleged in Peekay was as to the nature of the transaction which was being entered into, it was the nature of the transaction which it was being contractually agreed was fully understood, and the estoppel prevented Peekay from asserting that it entered into the contract by a misunderstanding of the nature of the transaction. The specific misunderstanding was as to the specific matter which it had been contractually agreed was fully understood.
CRSM on the other hand stressed that amaterial fact in Peekay was that the investment product was fully and accurately described in the documentation provided to the investor by the bank, including the contractual documents which Mr Pawani agreed, by signing, that he had read and understood (though in fact he had not). By contrast, in the present case the misrepresentations complained of were contained in the “fact sheet” provided by Barclays to CRSM and were never corrected.
It forcefully submitted that it is one thing to ask an investor to confirm that he has read and understood an accurate description of the transaction into which he is agreeing to enter and to treat him as precluded by such confirmation from asserting that he has read and understood it. It is quite another and is plainly unreasonable to seek to treat such a confirmation as precluding the investor from subsequently complaining that the description which he was asked to confirm that he had read and understood turns out in fact to have been inaccurate.
Whilst I agree with CRSM that the facts in Peekay were as they describe, in my judgment the crucial issue is the proper construction of what has been agreed and a sufficiently widely and clearly drawn agreement may well have the effect which CRSM submits would be unreasonable.
In the present case by clause 6 CRSM was contractually agreeing that it understood and accepted the risks of entering the transaction and purchasing the Notes. In my judgment if the substance of the claim for misrepresentation is that representations were made which led it to misunderstand the risks of entering the transaction and purchasing the Notes then such a claim would be precluded. It is contractually estopped from asserting that it was induced to enter into the contract by a misunderstanding of the nature of the risks entering the transaction and purchasing the Notes. As in Peekay, the specific misunderstanding would be as to the specific matter which it had been contractually agreed was fully understood.
In my judgment the substance of the Purchase Representation claim is that representations were made which led CRSM to misunderstand the risks of entering the transaction and purchasing the Notes. The Purchase Representation is in terms a representation as to the risks of purchasing the Notes – it is a statement that the Notes have a very low risk of default. The consequence is alleged to be that CRSM understood the Notes to have a lower risk of default than they in fact had. Moreover, that is a risk of a kind specifically identified in the fact sheets which CRSM had signed and acknowledged. I accordingly hold that (absent proof of fraud) it is contractually estopped from making its Purchase Representation claim.
I would not, however, regard the substance of the Restructuring Representation claim as being that representations were made which led CRSM to misunderstand the risks of entering the restructuring transaction. They were statements as to the criteria to be applied in carrying out the restructuring. Although that may have had an impact on the risks involved they are not statements as to such risks nor is the claim based on a misunderstanding of such risks. I accordingly hold that no contractual estoppel arises in respect of such claims under clause 6 (or its equivalent in the restructuring contract).
Clause 8 (“Terms and Conditions”)
This provided as follows:
“8. Terms and Conditions. We acknowledge that the operative terms and conditions of the Notes will be exclusively those set forth in the Offering Documentation and that we are not entitled to rely on any description of the terms and conditions of the Notes or any undertaking by any other party with respect to the Notes that is not set forth in the Offering Documentation, including, without limitation, any such description or undertaking communicated orally or set forth in any pitchbooks or other marketing materials. For purposes of this letter, "Offering Documentation" shall mean the Programme Memorandum and any supplement thereto in respect of the Notes.”
CRSM submitted that Clause 8 (“Terms and Conditions”) is designed to prevent an investor from alleging collateral contractual terms, or from relying on “any description of the terms and conditions of the Notes” or “any undertaking” (i.e. collateral promise) other than as set out in the formal contract documents. It has no bearing on a claim for misrepresentation.
Barclays submitted that it was the agreed basis of each purchase transaction, and it was CRSM’s contractual promise, that the operative terms and conditions of the Notes would be exclusively those set out in the Offering Documentation and that CRSM was not entitled to rely on any description of the terms and conditions of the Notes not set forth in the Offering Documentation, including (without limitation) any such description communicated orally or set forth in any pitchbooks or other marketing materials (Clause A8). Barclays accepted that that does not, without more, preclude CRSM’s claims. However, it was submitted that it does mean, at all events when read with Clause A6, that CRSM cannot succeed if its claims are inconsistent with the provisions of the Offering Documentation, and all of CRSM’s purchase claims are in fact inconsistent with the Offering Documentation. Clause A8 also reinforces, were reinforcement needed, the effectiveness of Clause A6 to defeat CRSM’s claims. Read with Clause A8, CRSM promised by Clause A6 that it had agreed to buy, and that on completion it bought, having been competent to assess and understand, and having in fact understood and accepted, all the risks created by the definitive final terms set out in the Offering Documentation.
I agree with CRSM that this clause is concerned with identifying the terms and conditions applicable to the Notes and making it clear that those are the only relevant terms and conditions. It does not address the effect of such terms and conditions. I accordingly hold that it does not preclude the misrepresentation claims, either in itself or when read with other clauses.
Finally, CRSM had originally contended that if and to the extent that the contract clauses precluded claims for misrepresentation they were rendered ineffective by section 3 of the Misrepresentation Act 1967. This argument was not pursued in Closing.
The counterclaim
Barclays has pleaded a counterclaim on the basis of the following provision of the purchase letters (and the restructuring letter):
“Indemnification. We understand that the representations, warranties and covenants herein ... are being relied on by you and we agree to indemnify and hold you harmless from and against all losses, damages, liabilities, claims, costs, charges and expenses which you may incur by reason of any breach of our representations, warranties and covenants herein, or by reason of any claims or legal proceedings arising out of or in connection with the subject matter of any of our representations, warranties and covenants herein. ...
Barclays alleges that it is a breach of this provision for CRSM to bring the present claim “because CRSM’s misrepresentation claims are based on matters which in the Purchase Letters and Restructuring Contract it represented and warranted it would not claim in respect of”.
If this claim is to be pursued I would wish to hear further argument on it in the light of my conclusions on the issue of contractual estoppel.
X. Implied Term
In relation to the initial sale of the Notes, CRSM made an alternative claim for breach of the contracts for the sale of the Notes. CRSM submitted that it was an implied term of each of those contracts that, in structuring the CDO2 incorporated in the relevant Note, Barclays would not do so with the deliberate intention that its risk of default would be materially different from that indicated by its anticipated AAA rating. Barclays breached that term by deliberately engaging in credit ratings arbitrage in the manner described earlier. CRSM claims damages on a reliance loss basis, i.e. the same measure as it claims as damages for misrepresentation in relation to the sale and purchase of the Notes.
CRSM relied upon the principle of business efficacy - that the Courts are willing to imply a term on the ground that without it the contract will not work (Chitty on Contracts, 30th ed., para 13-005).
It drew attention to the House of Lords decision in Equitable Life Assurance Society v Hyman [2002] 1 AC 408 and submitted that a number of significant points emerge from that decision:-
The implied term fills a gap left by the express terms. It is not a bar that the express terms cannot be construed so as to give effect to the putative implied term: the only limitation derived from the express language of the instrument is that the terms must not conflict.
The implied term is derived by a process of interpretation having regard to the “particular commercial setting” of the contract, the parties' “actual circumstances” or the contract’s “objective setting”.
That setting may include the fact that a facet of the investment would have been a “good selling point” or a “significant attraction for purchasers” of the investment.
The implication must be strictly necessary in order to give effect to the reasonable expectation of the parties.
CRSM submitted that the relevant circumstances in the light of which the contract must be construed included in particular:
The AAA rating was understood to be an essential feature of the transaction;
It was also mutually understood that the reason why this rating was essential was because CRSM required a secure investment with a minimal risk of default;
Barclays knew that the reason why CRSM was relying on the rating was because it did not have the expertise to carry out its own analysis of the risk of default; and
The Notes were structured by Barclays, which did have the relevant expertise.
Having regard to these circumstances, it was submitted that it is necessary to imply a term that Barclays would not structure the CDO2 with the intention that its risk of default would be materially different from that indicated by the anticipated AAA rating, thereby defeating the purpose of the rating requirement. Such an implied term is essential in order to give business efficacy to the contract and to give effect to the parties’ reasonable expectations (objectively viewed).
In considering CRSM’s argument on implied term regard should be had to the recent Privy Council case of Attorney General of Belize v Belize Telecom Limited [2009] 1 WLR 1988. Giving the judgment of the Privy Council Lord Hoffman said that the essential question is “what that instrument, read as a whole against the relevant background, would reasonably be understood to mean” (at paragraph 21) and explained that the various different formulations which the courts have used all come back to this question and are different ways of saying “although the instrument does not expressly say so, this is what a reasonable person would understand it to mean..” (at paragraph 25).
The Belize Telecom case was considered in the Court of Appeal decision Mediterranean Salvage & Towage Limited v Seamar Trading & Commerce Inc [2009] 2 Lloyd’s Rep. 639. In that case it was emphasized that the touchstone remains necessity rather than reasonableness.
It is also relevant to have regard to the considerations set out by Lord Simon of Glaisdale in BP Refinery (Westernport) Pty Ltd v Shire of Hastings (1977) 180 CLR 266 , 282–283 in support of the arguments for/against implication. In that case Lord Simon identified the following considerations:
“(1) it must be reasonable and equitable; (2) it must be necessary to give business efficacy to the contract, so that no term will be implied if the contract is effective without it; (3) it must be so obvious that ‘it goes without saying’ (4) it must be capable of clear expression; (5) it must not contradict any express term of the contract.”
Whatever test one adopts I am satisfied that the alleged term cannot be implied. In particular:
It is not capable of clear expression. As I have already found, there is no such thing as an actual risk of default. There are simply different means of estimating that risk. The term does not identify by what means this is to be done. It is also wholly unclear what is meant by a materially different risk of default. What expected loss is that meant to cover and on what basis is that to be determined?
It is not necessary to make the contract work. Contracts between banks for the sale and purchase of complicated structured products work perfectly well on the basis of the principle of caveat emptor. Further, as already found, structuring of the Notes to increase the spread implied probability of default does not have any necessary impact on the real world or actual probability of default.
It is not reasonable in that its inclusion would potentially undermine the practices of the banking market. The evidence showed that it was the very basis of the CDO business to make an arbitrage, buying protection at spreads which meant that Barclays would make a MTM profit on the CDO. Further, the evidence was that it was common for banks to make profits thereby of the order of more than 10%.
It contradicts the express terms of the contract or at least is inconsistent with the contractual allocation of risk and responsibility agreed thereby. In particular CRSM expressly agreed that it understood, accepted and assumed the risks of purchasing the Notes, which risks included the risk of default, as the fact sheet qualifications and disclaimers made clear.
This is not what a reasonable person would understand the contract to mean. On the contrary it would render the contract uncertain, self-contradictory and unreasonable in its effect.
I accordingly reject CRSM’s implied term case. In the light of the findings I have made, it is in any event not made out on the facts.
XI. Quantum
In the light of my conclusions on liability I shall address this issue relatively briefly and shall only deal with the issues of principle which arise.
The Purchase Claim
The basic object of an award of damages in tort is to put the injured party in the same financial position as he would have been in if he had not sustained the wrong for which he is being paid compensation: Livingstone v Rawyards Coal Co (1880) 5 App Cas 25, 39.
The general rule in tort as in contract is that damages are assessed as at the date of breach: see e.g. Miliangos v George Frank (Textiles) Ltd [1976] AC 443, 468; Dodd Properties (Kent) Ltd v Canterbury City Council [1980] 1 WLR 433, 451; Future International Ltd v Sealand Housing Corpn [2002] EWHC 2454 (Ch), [2002] All ER (D) 28 (Dec) (where shares obtained by fraud, the value should be calculated as of their date of purchase).
Damages may be assessed by reference to another date if the court considers that to do so would more fairly and appropriately give effect to the basic compensatory principle: see e.g. Smith New Court Securities Ltd v Scrimgeour Vickers (Asset Management) Ltd [1997] AC 254, HL; Golden Strait Corpn v Nippon Yusen Kubishika Kaisha [2007] 2 AC 353, 381, 397.
In Smith New Court, as the result of fraudulent misrepresentation the representee bought a large parcel of shares, the true value of which was greatly reduced by a then undiscovered separate fraud practised on the company. It was held that the representee could recover the whole of the difference between what it paid for the shares and the amount actually realised on the resale of the shares. Lord Browne-Wilkinson stated at p266-7:
“In sum, in my judgment the following principles apply in assessing the damages payable where the plaintiff has been induced by a fraudulent misrepresentation to buy property: (1) the defendant is bound to make reparation for all the damage directly flowing from the transaction; (2) although such damage need not have been foreseeable, it must have been directly caused by the transaction; (3) in assessing such damage, the plaintiff is entitled to recover by way of damages the full price paid by him, but he must give credit for any benefits which he has received as a result of the transaction; (4) as a general rule, the benefits received by him include the market value of the property acquired as at the date of acquisition; but such general rule is not to be inflexibly applied where to do so would prevent him obtaining full compensation for the wrong suffered; (5) although the circumstances in which the general rule should not apply cannot be comprehensively stated, it will normally not apply where either (a) the misrepresentation has continued to operate after the date of the acquisition of the asset so as to induce the plaintiff to retain the asset or (b) the circumstances of the case are such that the plaintiff is, by reason of the fraud, locked into the property. (6) In addition, the plaintiff is entitled to recover consequential losses caused by the transaction; (7) the plaintiff must take all reasonable steps to mitigate his loss once he has discovered the fraud”.
In this case CRSM’s primary claim was for damages in accordance with the general rule assessed at the dates of the relevant transactions, being the dates when the CDO2s were purchased and the date of the restructuring. Alternatively, damages were claimed on the basis of the net sums actually lost.
In addition, CRSM claimed its lost return since that date by reference to its return on assets (not, as currently pleaded in the Particulars of Loss, return on equity), in accordance with the principles set out in Sempra Metals v Inland Revenue Commissioners [2007] UKHL 34, [2008] 1 AC 561 at paras 94-95.
Barclays submitted that this was a case in which the general rule of assessment at the date of the transaction should be displaced in favour of the selection of another date of assessment, or another measure, in order to give a true measure of the claimant’s loss. Two matters in particular were relied upon:
where the claimant has actually re-sold the financial product, producing an actual re-sale loss, it may well be more appropriate to assess the loss by reference to that loss, or to put the same point another way, the claimant must give credit for the proceeds of the re-sale: see Kennedy v Van Emden (1997) 74 P&CR 19; Smith New Court v Scrimgeour Vickers [1992] BCLC 1102 (Chadwick J), 1142-1143, para 2 (overturned on appeal, but not on this point).
Where the claimant was buying to hold and continues to hold under the continuing influence of the misrepresentation, it may well be inappropriate to assess damages at the date of the transaction – see Downs v Chappell [1997] 1 WLR 427, 441E-443F; Naughton v O’Callaghan; Smith New Court (HL), 267B-C.
I agree that this is a case in which date of transaction damages would not appropriately reflect CRSM’s loss. In particular: (1) these were investments which CRSM was always intending to buy and hold rather than resell; (2) these were not readily marketable investments; (3) the alleged misrepresentations did continue to operate after the date of acquisition; (4) CRSM were to a significant extent locked into the investments – the only available buyer was Barclays itself; (5) this is a case involving actual re-sales, and (6) the logic of taking date of transaction damages is that CRSM would still be entitled to its damages claim of some €92 million even if the Notes had been repaid and a full coupon received, a remarkable windfall.
On the basis that this is not an appropriate case for date of transaction damages it was common ground that the damages claim in relation to Umbria II and Como I should be based on the net proceeds following their resale to Barclays.
In relation to Como II it was CRSM’s case that the damages should be based on the full amount of the Note since there has been no recovery of any part of the principal amount of €70 million paid by CRSM to Barclays, as defaults have reduced the coupons and principal payable under the CDO2 to nil.
It was Barclays’ case that the appropriate date of assessment is the date when the claimant was no longer under the influence of the defendant’s misrepresentation. In such cases, provided that the claimant was freely in a position to sell the asset, it follows that from that date (or such date thereafter by when a sale could reasonably be expected to have occurred, given the practicalities), the claimant was in a position to assess its position and sell if it so chose, and there is no reason to continue to “postpone” a loss assessment that on the prima facie rule would have been performed at the date of the original purchase. Subsequent declines in value of the instrument are caused by the free choice to hold onto the asset, or by subsequent market events, and not by the misrepresentation: Twycross v Grant 2 CPD 469, 544-545; Downs v Chappell [1997] 1 WLR 427, 441E-443F, 444H-445A. However, if the factual consequence of the misrepresentation is that the claimant is “locked in” to the asset, even after the moment of discovery, then damages after that moment may be recovered, and the relevant date for assessment will be when the relevant constraint is removed: Smith New Court (HL), 267C.
Barclays submitted that any date after February 2007 (at the latest) is not appropriate for assessment, and that it would be wrong to assess damages in respect of Como II on the basis of the difference between the purchase price and its current nil value, because there have been multiple breaches of the chain of causation between the misrepresentation and such dates and losses in that:
Mr Morolli’s evidence was clear: from March-November 2006 onwards, CRSM were engaged in an independent trading speculation. They were no longer influenced by the misrepresentations in any way and were assessing matters for themselves. They took a view on market movements and decided the risk of holding onto Como II was acceptable. This was a calculation based on their views of the risks of the relevant numbers of defaults arising before the end of the life of the note, compared to the immediate mark to market loss they would have to accept if they re-sold it. From this moment, the logic of the market measure cases applies in full force.
CRSM was not locked in: they were entirely free to sell. There was no financial constraint which meant that realising a loss on Como II would be problematic. None was alleged, and Mr Morolli accepted that they could have sold Como II if they had chosen. They had ample reserves to absorb the “hit”. A desire not to realise a loss does not mean you are locked in. Further, the mere fact that CRSM was a buy-to-hold investor does not mean it was locked in. In fact, as its own evidence showed, although not a speculator, it was an investor that would exit assets which seemed unpromising as and when appropriate; and the whole context is that it had bought to hold these investments on a false basis, but the scales had now fallen from their eyes.
Further, the loss post November 2006 was caused by (i) the subsequent defaults of corporate debtors, so far as the actual loss of principal is concerned; (ii) declines in MTM value, due to market movements. The defaults which caused this note to default occurred after Mr Morolli and CRSM decided that it was worth taking the risk of defaults to hang on to the notes. That is not a consequence of the breach. The case is therefore analogous to the example given by Cockburn CJ in Twycross v Grant: if a purchaser is induced to buy a racehorse by fraud, he is not liable for the entire price of the racehorse because the racehorse subsequently catches a disease and dies (at 544-545). In effect, the financial crisis of 2008-2009, and the 10 defaults that occurred during and after it, are the “disease” that Como II subsequently caught.
CRSM submitted that it would not be right to deduct from the amount paid by CRSM to Barclays and which it has not recovered a notional amount for which CRSM would have been able to sell Como II back to Barclays either in March 2007 or on any other date. In particular:
The loss of CRSM’s investment in Como II flowed from the acquisition of the Notes by CRSM in reliance on Barclays’ misrepresentations. Nothing happened between the date of acquisition and the date on which Como II became a total loss which can be said to have broken the chain of causation.
It cannot and in any event is not said that CRSM acted unreasonably in retaining Como II and thus failed to mitigate its loss.
Even if Barclays was right that the defaults which wiped out the value of Como II were not foreseeable, that would not assist Barclays as it is clear law that consequential losses directly caused by the transaction are recoverable even if not foreseeable.
In arguing that the write-down of the amount repayable under Como II to zero should be regarded as a consequence of an “independent trading speculation” made by CRSM to retain it, Barclays placed misplaced reliance upon Mr Morolli’s acceptance in cross-examination that, in holding on to Como II, CRSM was “making a trading judgment” and decided to “run the risk” that loss might occur as result of subsequent defaults; and also that CRSM made an entirely “independent decision” as to what to do about the Notes in the sense of deciding independently of Barclays what to do about them. The same could equally be said of the plaintiff in Smith New Court and did not prevent it from recovering damages based on the sums which it actually received from the sale of the shares when it decided in the exercise of its trading judgment to sell them. It does not follow from the fact that the claimant makes its own commercial decision as to whether to sell or hold onto an asset which it has acquired as result of misrepresentation that subsequent losses (or gains) are to be treated as a consequence solely of that decision and not of the misrepresentation.
Even if there did come a time when it could no longer be said that CRSM was under any remaining misapprehension as to the nature and risks of the CDO2s, CRSM’s subsequent losses should be regarded as a continuing consequence of the representations which caused it to buy the CDO2s, in circumstances where the evidence has established that CRSM was “locked into” its investment, in at least as strong a sense as the plaintiff in the Smith New Court case was regarded by the House of Lords as “locked into” holding the shares which it had bought.
I find that CRSM was not in a position to take a truly independent trading decision. It remained to a significant extent “locked into” the investment. In particular:
The CDO2s were not readily marketable assets and it is unlikely that anyone other than Barclays would have been willing to buy them at a realistic price.
The prices indicated by Barclays for Como II were always significantly (at least 15% and often considerably more) below par, so that CRSM could not have sold the Notes back to Barclays without realising a substantial loss.
Unlike shares (or a case involving the sale or shipment of goods), the asset was not one which could only be realised by selling it at its present value in any available market: if CRSM retained the Notes, it would continue to receive the full coupon and would automatically recover the full principal amount of its investment at maturity unless sufficient defaults occurred to reach the attachment point of the CDO2.
It has not been and could not be suggested that CRSM’s decision to retain Como II was unreasonable.
I accept CRSM’s case on this issue essentially for the reasons it gives. I am not satisfied that it has been shown that the loss flowed not from the acquisition of the Note but from CRSM’s decision to retain it – see Smith New Court at p268A. This is a case in which the claimant acted reasonably in retaining the asset – see Smith New Court at p265B. There is no allegation of failure to mitigate and it is accepted that CRSM’s conduct was not unreasonable. Nor do I consider, if relevant, that it has been established on the evidence that it was the financial crisis that caused the Como II note to default. I find that there was no new intervening cause or break in the chain of causation and that the loss claimed does directly flow from the purchase of the Note.
If CRSM had made good its claim I would accordingly have held that it was entitled to recover damages on the basis of the full value of the Como II Note.
The restructuring claim
CRSM advanced a claim for date of transaction damages. If, however, I found that that was not the appropriate basis for the claim they advanced no alternative claim but submitted that any such claim should be held over pending a consideration of the court’s findings. Barclays objected to this. It stressed that this was the trial of both liability and quantum. If CRSM had a claim it had to plead and prove it, which it admittedly had not. I agree with Barclays. It would not be satisfactory to hold this aspect over absent an order or agreement to that effect.
The Sempra metals claim
CRSM claimed interest on the principal amounts in the form of its lost return, in reliance upon Sempra Metals v Inland Revenue Commissioners [2008] 1 AC 561, 601 at paras 94-95. The rates of return were calculated by Mr Morolli and allegedly represented CRSM’s average rate of return in each relevant year on the rest of its investment portfolio (excluding the CDO2s).
CRSM’s case as ultimately put forward was not pleaded until after the start of the trial and was supported by evidence produced during the course of the trial. This is not a satisfactory way to advance a claim of this nature. Further, the evidence was of a very generalised nature. There was no evidence that a specific investment opportunity had been lost. I am not satisfied that a case based on the use to which the monies would otherwise have been put is established simply by evidencing the return achieved on a general portfolio. As Barclays pointed out, what would not have happened is that CRSM would have invested in the portfolios that Mr Morolli used as the basis for his calculation. Mr Morolli’s portfolios were not portfolios of the assets available for purchase at the times when the notes were purchased. As became clear in cross-examination, they were instead the portfolios which CRSM was holding in the relevant year, including assets purchased long before at higher rates of interest. I reject this late claim.
XII. Conclusion
I am grateful to Counsel and solicitors on both sides for their assistance. Despite the considerable skill with which the Claimant’s case was constructed and presented I have reached the clear conclusion that the claims fail on various grounds and must be dismissed.