Loreley Financing (Jersey) No 30 Limited v Credit Suisse Securities (Europe) Limited & Ors

In late 2007, as clouds gathered in the sub-prime mortgage market, the Claimant (“L30”) paid US$100m to the Defendants (“CS”) for certain Notes (“the Notes”), rated AAA, which formed the basis of what is known as a synthetic CDO transaction. Such transactions are complicated – indeed some of the banking witnesses in this case struggled to explain the concepts themselves.

But the bottom line is this: the financial crisis hit, L30 lost its money in 2010 and seeks to claim that money back. It says that this was more than a bad deal. It says that in selling the transaction CS made representations about the business packaged within the CDO transaction, that the representations were false, made either negligently or deliberately, and that L30 would not have done this deal if the representations had not been made.

L30 has likened the transaction to a Russian doll; that analogy is apt; and it in part explains the complexity of the issues which have been raised and hence fall to be decided. The Notes were not self standing; they were linked via a credit default swap (“CDS”) to the credit of a reference portfolio which comprised 100 residential mortgage-backed securities (“RMBS”). Each RMBS itself packaged a number of underlying individual mortgage transactions, in that it comprised rights to cash-flows arising from pools of underlying mortgage loans. Of those 100 RMBS, 7 RMBS were ones which had been packaged, securitised or underwritten by CS (or its affiliates) (the “CS RMBS (Footnote: ¹)”).

The bare essence of the claim is that:

The misrepresentation case therefore encapsulates a myriad of issues, some factual, some legal: was there any RMBS Misconduct in relation to the CS RMBS? Are there representations in relation to the CDO transaction – and if so what? With what intent and what knowledge were they made? Did they induce L30 to buy the Notes? This latter question includes legal issues as to the right test for what L30 needs to have known or understood or been aware of in relation to such representations.

There are also issues as to limitation. This claim was commenced in late 2018. No legally aware reader will have missed the elephant in the room: on the face of it, all of these claims are long since time barred. There are therefore questions as to whether L30 lacked, as it says it did, the requisite knowledge or information to start the limitation clock running.

Finally there are issues on negligence, as a backup if the fraud case does not work, but the other features of misrepresentation are present; and a rarity – an unlawful means conspiracy case which is a genuine alternative to the primary case.

These issues are dealt with in greater detail below, under the following headings:

As already stated, the Notes were a CDO, with the underlying reference assets being a portfolio of RMBS. The exposure to that portfolio of RMBS was synthetic, created through use of a CDS.

This overview, taken largely from a helpful document agreed by the parties, summarises the general features and characteristics including in respect of the specifics relevant to the Notes themselves, of:

An RMBS is a type of asset-backed security in which, as the name indicates, the securitized assets are residential mortgage loans. As with any asset-backed security, an RMBS is a structured financial instrument.

A single RMBS may be backed by thousands of individual mortgage loans. These loans are originated by companies offering mortgage finance to residential property owners. The originator may be connected with the sponsor entity, or may instead be a third party company which sells on loans it has originated to be securitised by a different financial institution.

The process of securitising an RMBS starts with the pooling of residential mortgages by a “sponsor” entity, which has originated or otherwise acquired those assets in order to pool them. This pooling of assets usually takes place in an entity related to the sponsor known as the depositor. The depositor then typically conveys the pool of loans to a special purpose securitisation trust, which then issues securities that are backed (or collateralised) by the assets and sells them into the market. The RMBS securities entitle investors to a periodic payment of interest, which flows primarily from the monthly revenue generated by the underlying mortgages, and a return of principal upon maturation of the securities.

RMBS securities are split into tranches to be sold to investors with different purchasing requirements. As such, RMBS generally consist of multiple tranches, which are typically differentiated by their risk profiles, and thus credit rating. No particular underlying mortgage is assigned to a particular tranche of the RMBS; rather, the mortgages as a pool collectively back all of the tranches in a particular RMBS. In general, lower risk / higher credit-rated tranches (referred to as “senior” tranches) have lower yields than higher risk / lower credit-rated tranches (“mezzanine” and “junior” tranches). Cash flows in an RMBS are structured such that the senior tranches receive the first “cut” of principal / interest payments when those are made, before mezzanine and junior tranches; and generally any losses of principal or shortfalls of interest from the portfolio of loans as a whole are similarly borne starting from the most junior tranches and moving upwards to the senior tranches.

Purchasers of RMBS receive interest and principal amounts repaid by the original mortgage borrowers of the individual loans that were pooled into the RMBS, but are also exposed to the risk of those borrowers failing to keep up repayments or defaulting. As such, where a sufficient number of those underlying borrowers fail to make repayments, or (for example) default on their loans, this may ultimately lead to the RMBS suffering a loss. Such losses are felt first by the higher risk / lower credit-rated tranches (i.e. the junior tranches), are often referred to as “writedowns”.

A CDO is a financial instrument that is similar to an RMBS insofar as it pools a number of underlying assets (in this case debt instruments) into an investment vehicle that then issues new bonds or notes to investors. Like RMBS, the securities issued by the CDO are typically divided into tranches with varying interest rates and credit parameters. The assets held by a CDO can be any type of debt obligation, including, but not limited to, corporate debt, RMBS or even tranches of other CDOs.

As described in the context of RMBS above, tranches of a CDO will be categorised by risk profile. A CDO is generally divided into multiple tranches, with senior tranches generally the first to receive returns and last to suffer any loss. The point at which a tranche would start to suffer a loss is commonly described as the “attachment point” and expressed as a percentage of the value of the underlying pool of assets. The term “detachment point” is commonly used to describe the point at which a tranche would suffer a total loss.

CDOs may be “managed” or “static” in terms of their pools of underlying assets. “Managed” deals require a CDO manager, who, subject to the specific terms of the CDO and within stipulated investment parameters, “manages” the underlying collateral or assets during the life of the CDO by actively trading in the underlying assets – for example, by substituting out non-performing assets. By contrast, a “static” structure generally means that during the life of the CDO the pool of underlying assets is fixed from the outset. This removes the need for a CDO manager, although the initial selection of the portfolio is often made by, or with, the assistance of a manager.

It is not always necessary for a CDO to acquire the assets that form the basis of the CDO. Instead, it is possible to create exposure to the chosen assets without purchasing the actual assets via a derivative structure, typically by using a CDS. CDOs constructed in this manner are referred to as synthetic CDOs. The way a CDS works to create this exposure is described below.

Put simply, a standard CDS is a contract that transfers credit risk associated with a third party entity or asset from one party (the “Protection Buyer”) to another (the “Protection Seller”). In consideration for the Protection Buyer paying the Protection Seller periodic premiums, the Protection Seller agrees to pay out an agreed sum contingent on the occurrence of certain defined credit events. If a credit event occurs, this will generally trigger payments from the Protection Seller to the Protection Buyer; this is the “swap” element, as the Protection Buyer has effectively swapped the risk of that credit event with the Protection Seller, such that the Seller has taken on that credit risk in respect of the underlying asset. In many ways, a typical CDS functions similarly to insurance, whereby the Protection Buyer pays an agreed upon premium and the Protection Seller provides a payment if the underlying third party entity or asset defaults, i.e. causes losses.

The definitions of credit events are critical parts of CDS documentation as they describe the circumstances in which payments may flow from the Protection Seller to the Protection Buyer and how settlement may occur. Standard RMBS credit events were defined by ISDA, with Protection Sellers making monthly floating payments equal to the incremental losses on the underlying reference assets reported in any given month.

In the context of a synthetic CDO, the CDS element is used to create exposure to an asset without the need to purchase the actual asset. Rather than the issuer of the CDO actually acquiring the asset (for example, the portfolio of RMBS), it becomes a Protection Seller, entering into a CDS with a counterparty (the Protection Buyer), which is usually an affiliate of the broker/dealer arranging the CDO. The Protection Seller then has a contractual exposure to the credit risks of the underlying reference asset without actually owning it, and receives compensation/premiums in return from the Protection Buyer. As such, where the reference asset is performing, the CDS counterparty (the Protection Buyer) will make regular payments to the CDO issuer (which, in turn, funds payments to the CDO investors). Following a credit event, however, the CDO issuer will be required to make payment(s) pursuant to the credit protection it has sold to the CDS counterparty (the Protection Buyer), which it funds from the investment made into the CDO by investors, reducing the principal amount of the bonds / notes held by those investors. At the transaction’s maturity, the proceeds of the investments (if any) are used to repay the remaining principal amount of the bonds / notes to investors.

In order to fund the premiums which it needs to pay to the Issuer under the terms of the CDS, and to hedge its exposure, the counterparty will typically enter into hedging arrangements with other market participants, for example in the form of further CDSs (where the swap counterparty is the Protection Seller). The following section concerning the Notes addresses only the relationship between the Issuer and Noteholders and the Issuer and the swap counterparty and does not address any hedging transactions which took place between the counterparty and other market participants.

The RMBS business and RMBS Misconduct allegations are considered discretely in one of the Appendices to this judgment. The key background information about it can be summarised thus.

The business model of CS’ Structured Products group on the RMBS side (the “RMBS Group”) during the Relevant Period was to acquire loans through various acquisition channels and then either to package these loans into RMBS, which Credit Suisse (USA) LLC (“CSSU”) sold to investors, or to sell them as part of a portfolio of loans to other financial institutions through “whole loan sales”.

CS’ business of acquiring loans onto its own book was known as the Conduit. The Conduit had four separate “channels”, (1) the “Bulk Channel”; (2) the “Mini-Bulk Channel”; (3) the loan-by-loan channel (“LBL Channel”); and (4) the “Wholesale Channel”.

In respect of loan pools in the Bulk and Mini-Bulk Channels, acquisition was processed by CSSU “Trading Desks”. At points during the Relevant Period, the Co-Heads of Non-Agency Trading in the RMBS Group were Mr Daniel and Mr Schoen while Messrs Dodman (Subprime), Schoen (Scratch and Dent), Daniel (Fixed Rate), and Vibert (Adjustable-Rate Mortgage, or “ARM”) all headed a desk.

The Trading Desk produced a bid price for the loan pool which was offered by the seller. If the bid was accepted by the seller, then CSSU carried out due diligence on the loan package. Sometimes sellers imposed stipulations (called “trade stips”) on which the sale of the pool was conditional.

Following the due diligence results, CSSU and the seller agreed a final pool of loans it would purchase.

Through the LBL Channel, approved sellers could access an online CSSU portal whereby an automated process compared the seller’s input loan characteristics with CSSU’s acceptable loan parameters and automatically produced to the seller a denial or a purchase price for the loan.

The Wholesale Channel process was of CSSU originating its own loans. A broker inputted the prospective loan characteristics into an online portal which automatically issued either an initial approval (with or without conditions) or a denial. If the loan was approved, the broker would be instructed to send the full loan file to a Fulfilment Centre - a third party designated to underwrite CSSU’s Wholesale loans subject to criteria given to them.

A critical part of the process for the purposes of this dispute was the due diligence (“DD”) process, to ensure that loans being acquired and securitised were sufficiently good quality .

Other than the Wholesale Channel, where CSSU was the underwriter, CSSU performed pre-acquisition due diligence on loans (sometimes by way of sample in relation to certain categories of loans). CSSU’s due diligence process varied by acquisition channel.

Due diligence in the Bulk Channel was overseen by the Credit and Underwriting Group, headed by Mr Sacco, though it was principally “Bulk” due diligence managers such as Mr Nordyk of that group who supervised it. Due diligence in the LBL and Mini-Bulk channels (and underwriting in the case of the Wholesale Channel) was overseen by the Funding and Fulfilment Group, which supervised and was responsible for the Fulfilment Centers. The Head of this group was Mr Othman.

Due diligence reviews principally involved three main elements: (1) a credit review; (2) a compliance review; and (3) an appraisal review.

In the Bulk Channel, loans were allocated an “event level grade” referred to as EV1, EV2 and EV3 in respect of their credit and compliance review. The import of an EV grade is in issue between the parties.

CS operated a Quality Control (“QC”) process on loans it acquired. This was run by the Credit and Underwriting Group. QC was a monthly process which involved subjecting a sample (usually 3%) of the total loan volume purchased through the Bulk, Mini-Bulk and LBL Channels and originated through the Wholesale Channel to a QC review.

The process of putting together an RMBS deal was a continuum. A part of the process – and generally what one might term an “early stage” part – involved the making of presentations, illustrated by Presentation Slides. To some extent these were put together to educate newcomers to the market or newcomers to business with CS. To some extent of course, since CS was in the business of putting together a successful product the Presentation Slides formed part of a sales effort. The Presentation Slides were not intended as contractual documents. Often it can be seen that points made in the Presentation Slides are summaries, or “jumping off points”. The presentation would be oral and in large measure the Presentation Slides were intended to summarise or navigate what was being said. Any representations which came from a presentation would have to be representations which were the result of the oral and slide presentations. Those might coincide and the result might be that a presentation did result in a representation being made in the form of something which appears on a slide. However what was said/shown needs to be assessed against the full background of the presentations and the other documents which succeeded these early stage communications.

There were then Offering Documents – namely the Prospectus and the ProSupp. They provided information required by law about the offering of securities. The Prospectus (which has been described as covering a whole “shelf”) set out general information, and often indicated a range of possible approaches which could be adopted in the ProSupp for any particular deal. The ProSupp (itself lengthy) “supplemented” that Prospectus with specific information for a particular deal as well as a good deal of material provided for a regulatory purpose in accordance with the SEC’s Regulation AB. It provided detail about the structure and nature of the warranties being provided. For example it might explain that the Seller acquired loans from third parties and that, as part of the securitisation, it “will make certain representations and warranties to the depositor and the trustee regarding the mortgage loans”, and refer to the repurchase/substitution obligation. It would describe the contracts pursuant to which the RMBS were created - the Pooling and Servicing Agreements (“PSAs”) - and summarise representations to be found in that document. It would also contain a number of sections upon which L30 relies indicating that it was a document upon which some reliance was intended to be placed (such as “you should rely on information contained in this document…”) and giving (for example) some information about the mortgage pool indicating that the information in the ProSupp was believed to be representative of the characteristics of the loans to be included in the mortgage pool.

The securitisation of loans and their offer to investors took place pursuant to a detailed contractual framework, with heavy legal involvement from specialist law firms working with market standard documentation. The relevant contractual (and other) documents for these arrangements extended to many hundreds of pages of carefully drafted text.

As just noted, the PSA is the contract pursuant to which the loans were transferred to the trust, and which sets out the terms on which they do so. As L30 submitted, the PSA does not necessarily affect whether representations were made elsewhere. However the PSA must form a part of the context against which the existence of such alleged representations is considered.

Each PSA contains a long list of representations and warranties. For example, taking the PSA for CSMC 2007-3, it is a document of nearly 300 pages (with schedules and exhibits). The following features are worthy of note:

In 2002, a German bank, IKB Deutsche Industriebank AG (“IKB”) established an asset-backed commercial paper conduit which became known as the Rhineland Programme. The Rhineland Programme consisted of a Delaware corporation (the Rhineland Funding Capital Corp or “RFCC”) and a series of special purpose companies, most of which were registered in Jersey (“the Rhineland SPCs”). Bedell Trust Company (later known as Ocorian) provided administrative and operational support to the Rhineland Programme, and (along with a then associated law firm, Bedell Cristin), provided professional directors for the Rhineland SPCs. L30 became one of the Rhineland SPCs following its incorporation on 4 January 2007. RFCC issued short-term secured commercial paper and lent the money so raised on to the Rhineland SPCs for them to invest especially in CDOs.

At first, IKB itself was responsible for identifying potential investments for the Rhineland SPCs and advising those companies which investments they should purchase. However, in 2006 IKB incorporated a subsidiary called IKB Credit Asset Management GmbH (“IKB CAM”) to take over this role, which was returned to IKB upon IKB CAM’s dissolution in 2008.

The procedure that was followed when IKB CAM made a recommendation was documented in a Tradeable Securities Investment Advisory Agreement dated 21 December 2006 (the “TSIAA”). In summary, the TSIAA provided for IKB CAM to make a recommendation to the Rhineland Programme’s Tradeable Securities Investment Committee (the “TSIC”), which would consider and approve the recommendation. The recommendation would then be sent on, together with the TSIC’s approval, to a Rhineland SPC whose directors would consider and approve the acquisition of the investment. IKB CAM would then assist the Rhineland SPC to conclude the acquisition of the investment from the seller. Certain administrative tasks, such as the execution of any funding agreement required between the Rhineland SPC and RFCC, would be arranged by Société Générale, the Rhineland Programme’s administrator.

In addition, IKB was a liquidity facility provider to the Rhineland SPCs. This meant that IKB was obliged to lend the Rhineland SPCs money in certain circumstances, including where a Rhineland SPC lacked the funds to make debt repayments to RFCC.

L30 purchased the Notes from the Third Defendant (“CSSU”) acting through its agent Credit Suisse Securities (Europe) Limited (”CSSEL”) for USD 100m in a transaction which was entered into on or around 13 July 2007.

The Notes were issued by Magnolia Finance II plc (”Magnolia”) and were structured as a synthetic CDO. CSSEL and CSI were the arrangers of the Notes. The underlying reference assets for the CDO were 100 RMBS (described as the “Reference Obligations”). The Reference Obligations were set out in a Reference Obligation Schedule (the “Reference Portfolio”) shown in Annex B of a Series Memorandum dated 25 July 2007 (“the Series Memorandum”). The Series Memorandum is a document 196 pages in length.

The Series Memorandum had been preceded by Termsheet (i.e. documents setting out draft indicative terms, subject to documentation, pricing, and relevant approvals). The Term Sheets for the proposed Notes sent on behalf of CSI and CSSU to IKB CAM were about 30 pages long. They recorded that the selection of the Reference Obligations for inclusion in the Reference Portfolio was the responsibility of Principal Global Investors (“PGI”). The Reference Portfolio was (subject to certain adjustment rights) static, i.e. once fixed prior to the sale of the Notes, the Reference Obligations did not change. As such, L30, as the investor / purchaser of the Notes, was exposed to the performance of those 100 RMBS (and their underlying mortgage loans) for as long as it held the Notes.

The Notes therefore represented “long” exposure on the part of L30 to the Reference Obligations, in that they broadly simulated a scenario in which it was holding itself the Reference Obligations (subject to their weighting in the Reference Portfolio and to the attachment and detachment point provisions). The exposure to each Reference Obligation was weighted up / down (as appropriate) in the Reference Portfolio such that they each had a notional value of USD 10 million, despite the “original principal amount” of the individual Reference Obligations (which, broadly, reflected the value of each underlying RMBS) being of differing values. As a result of this weighting, the Reference Portfolio had a total notional value of USD 1 billion, reflecting the USD 10 million notional value of each of the 100 Reference Obligations.

So, for example, tranche “M5” of the RMBS “JP Morgan Alternative Loan Trust, Series 2006-S4” (row 56 of Annex B of the Series Memorandum) had an “original principal amount” of USD 4,375,000 and therefore was weighted upwards by approximately 228.57% to create the synthetic exposure to the notional amount of USD 10 million in the Reference Portfolio. But in respect of the Reference Obligation in the next row (i.e. row 57), its “original principal amount” of USD 43,279,000 needed to be weighted downwards by approximately 23.11% in order to create the USD 10 million notional value.

The Notes were linked to a tranche of the CDO that had an attachment point of 9% and detachment point of 19%. This meant that the principal value of L30’s investment would not be diminished by any negative performance of the Reference Obligations until notional losses on the Reference Portfolio reached USD 90 million (i.e. 9% of USD 1 billion). L30 would suffer a total loss on its investment if the notional losses reached (and stayed at) USD 190 million or more (i.e. 19% of USD 1 billion).

The CDO’s synthetic exposure to the Reference Obligations was created through an embedded CDS (described in the Series Memorandum as the “Credit Swap Transaction”). The parties to that Credit Swap Transaction were Magnolia (for the CDO, effectively as the Protection Seller) and the Fourth Defendant (“CSAG”) acting through its Cayman Islands Branch, as the swap counterparty/Protection Buyer. For this reason, the Notes are sometimes referred to as "the Magnolia Notes".

Under the terms of the Credit Swap Transaction, CSAG (acting through its Cayman Islands Branch) would be obliged to make certain fixed payments to Magnolia – and L30 was, in turn, entitled as holder of the Notes to certain payments from Magnolia. In certain circumstances (broadly, those that would amount to a credit event in respect of the Reference Obligations), Magnolia would make payments to CSAG (through its Cayman Islands Branch) to reflect settlement for such credit events in the Reference Portfolio. These processes were complex, and are described in detail in the Series Memorandum – but, in essence, if the RMBS performed, L30 would receive the equivalent of interest payments and repayments of principal, while if the RMBS did not perform, then such payments would not be received (or would be received in reduced amounts), and the value of L30’s principal investment (described in the Series Memorandum as “Aggregated Outstanding Principal Amount”) in the Notes would also fall.

The structure of the Notes can be summarised in graphic form thus (a diagram to which L30 frequently made reference as highlighting the exposure of L30 to the credit risk of the RMBS):

Prior to the CDO Transaction which forms the basis of these proceedings, IKB and IKB CAM had worked frequently with CS in relation to synthetic CDO transactions. Loreley special purpose vehicles purchased more than 150 CDOs before this transaction.

Following a meeting with CS employees in December 2006 during which Mr Volker de Haan of IKB CAM expressed interest in investing in further CDO transactions, Mr Bernard Stroobandt (of CSSEL) provided various indicative proposals during the early months of 2007. IKB CAM in turn provided its feedback on these proposals. To give context, the bundle of documents relating to the CDO runs to 991 documents.

By May 2007, CS had formulated outline terms of a new synthetic CDO to be issued with the involvement of PGI, which would act as manager (if the tranches were to be managed, which was at that stage still to be determined). PGI was also intending to purchase certain junior notes in the proposed CDO for its own account and/or accounts managed by it.

On 16 May 2007, Mr Stroobandt sent Mr Jörg Zimmermann of IKB CAM an email with an initial proposal for a bespoke synthetic CDO, with the pool either selected by PGI (in static format) or managed by PGI (in managed format). The 16 May email explained that the tranche being offered was:

“USD 105,000,000 Aaa/AAA (Moody’s, S&P) tranche. No investment in AA required.

Spread: 1.40% in managed format or at 1.60% in static format

Pool. Min Rating A2”

Attached to the 16 May email was a term sheet (the “16 May Termsheet”) and a spreadsheet with a list of the proposed constituents of the Reference Portfolio entitled “PGI Pool.xls” (the “RMBS Spreadsheet”).

The 16 May Termsheet Stated on its front page that:

It then contained a page entitled “Important Notices” and consisting of various disclaimers (see page 4 of the 16 May Termsheet). This:

The proposed Reference Portfolio set out in the 16 May Termsheet (and attached in the RMBS Spreadsheet) comprised the same 100 RMBS as ultimately made up the Reference Portfolio for the CDO Transaction (i.e. the Reference Obligations). The Reference Obligations included 7 RMBS in the securitisation of which the Defendants or their affiliates had played some role (i.e. the CS RMBS).

The RMBS Spreadsheet provided details of 100 Reference Obligations and indicated that each such Reference Obligation carried a Moody’s credit rating of A2 or higher and a Standard & Poor’s credit rating of A or higher. It also set out information regarding the characteristics of the underlying pools of mortgage loans contained within each of the 100 RMBS, including in relation to: (i) FICO score (a form of credit score), (ii) loan to value ratios (“LTVs”), (iii) occupancy, (iv) property type, and (v) delinquency performance. The first tab of the RMBS Spreadsheet, entitled “Important Notices and Disclaimers”, contained inter alia the words:

“Information and opinions presented in this material have been obtained or derived from sources believed by Credit Suisse to be reliable, but Credit Suisse makes no representation as to their accuracy or completeness. Credit Suisse accepts no liability for loss arising from the use of this material. Nothing in this material constitutes investment, legal, accounting or tax advice, or a representation that any investment or strategy is suitable or appropriate to your individual circumstances”.

On or around 29 May 2007, Mr Zimmermann confirmed that IKB CAM wished to proceed in relation to the “AAA” tranche in the prospective CDO Transaction only. He also said that it preferred a static, rather than managed, format. He asked that the 16 May Termsheet be updated to reflect those changes and to give IKB CAM the right to substitute alternative RMBS for the Reference Obligations, provided that the substitutions did not affect the mark-to-market value of the portfolio. This led to the preparation of a revised termsheet which was sent to Mr Zimmermann by Mr Stroobandt on 30 May 2007 (the “30 May Termsheet”). Amongst other changes, the 30 May Termsheet no longer addressed Series B or C notes (since it was now clear that IKB CAM was not interested in these) and the former Section III (“Portfolio Manager”) was replaced with two new sections entitled “Portfolio Selection” and “Noteholders’ Removal and Substitution Rights” respectively (as the transaction was no longer to be managed). No change was made to the risk factors and disclaimers quoted above.

Meanwhile, following provision of the proposed Reference Obligations with the 16 May Termsheet, IKB CAM had begun its analysis of the proposed transaction. One of L30’s main witnesses, Dr Bauknecht described how this included feeding data provided by CS in relation to the Reference Obligations into the “CDOROM” software provided by Moody’s Investors Service Limited (“Moody’s”). This software was a risk modelling tool, to evaluate the credit risk of the proposed transaction. One of the purposes of this exercise was to determine the likely impact of raising the attachment point from 8.5% to 9%, which IKB CAM was contemplating. CS also provided IKB CAM with access to its proprietary analytics platform called “Locus”, which provided some information as to the characteristics of the underlying pools of mortgage loans for the Reference Obligations, including the distribution of FICO score, LTVs and second liens.

IKB CAM also requested, and CSSEL provided, various additional information regarding the Reference Obligations. In particular:

Mr Zimmermann and his colleague Mr Hubert Langer (who was a credit analyst) then prepared a document entitled “INVESTMENT PROPOSAL for ‘Magnolia II CDO’ to Rhineland Funding Capital Corp.” dated 31 May 2007 (the “Investment Proposal”). The Investment Proposal provided an overview of the proposed transaction and set out the results of IKB CAM’s analysis of the Reference Obligations and its evaluation of the CDO structure. The Investment Proposal set out IKB CAM’s analysis of the expected losses and loss coverage ratios for all 59 bonds for which CS had provided information and statistical data through its “Locus” platform, which included some but not all of the Twelve CS RMBS.

Having considered the risks, IKB CAM recommended an investment of US$100 million in the Notes. The recommendation was said to be based on: “the good risk-return-profile for the Class [A] (Aaa/AAA) yielding up to 150bps” and “the underlying portfolio constraints (minimum rating of A2)”, i.e. the fact that each of the 100 RMBS in the proposed Reference Portfolio was assigned a credit rating of A2 or higher. A strength of the proposed transaction was identified as: “Rating levels of underlying assets (min. A2/A rating)”. Two conditions were identified, namely: “1. Subject to Documentation (Legal Department of IKB CAM). 2. Rating Agency Confirmation for the proposed Ratings on the Notes by Moody’s and S&P”.

On 5 June 2007, the proposed investment was presented to IKB CAM’s Investment Committee, which comprised Messrs Frank Braunsfeld, Frank Lehrbass and Claus-Dieter Wagner, with Messrs Zimmermann and Langer also present. The IKB CAM Investment Committee decided to recommend that (subject to the same two conditions mentioned above) one of the Rhineland SPCs (which was not at that stage identified) should purchase US$100 million of Notes, as recorded in IKB CAM’s Record of Recommendation. The decision was communicated to CSSEL that day by Mr Zimmermann, who informed CS that the purchase had been recommended to a Rhineland SPC.

On or by 11 June 2007, the recommendation of IKB CAM’s Investment Committee was approved by three members of IKB CAM’s Advisory Board (Mr Ortseifen, Mr Braunsfeld and Mr Wagner).

As had been explained in an email from CSSU’s Mr Roland Jawurek on 17 May 2007, which was forwarded to Mr Zimmermann on the same day by Mr Martin Rudolph, CSSU stated that it planned to hedge the short exposure to the CDO tranches under the CDS embedded in the CDO.

On 6 and 7 June 2007 and following IKB CAM’s communication of its recommendation in relation to the CDO Transaction, Mr Jean Francois Dreyfus of CS’s Correlation Trading (US) desk acting on behalf of CS AG Cayman branch entered into delta hedges with Mr David Higgs’ Hedge Trading (EUR/US) desk acting on behalf of CSI in relation to the CDO to reflect the anticipated sale of the Notes.

By 11 June 2007, IKB CAM had identified L30 as the appropriate Rhineland SPC to purchase the Notes. On that date, it wrote to L30, Société Générale (which was responsible for administering the Rhineland SPCs), Wilmington Trust SP Services Ltd and RFCC recommending the CDO Transaction, enclosing a copy of IKB CAM’s Investment Proposal and Record of Recommendation, and asking that the prospective investment be considered by the TSIC.

On 11 June 2007, three officers of the TSIC (Messrs Mark Filer, Wolfgang Güth and Andrew Watt) wrote to L30 confirming that they had considered the recommendation and were satisfied that the CDO Transaction: (1) fell within the applicable credit and investment policy; and (2) would be in L30’s best interests. This letter stated that any decision to invest “shall solely be a decision of [Loreley 30]”.

On 15 June 2007, short notice was given that a board meeting of L30 would be held to consider and, if deemed fit, approve the CDO Transaction. The meeting, attended by Mr Paul Anderson (as alternate for Mr Alan Dart) and Ms Rebecca Bates (as alternate for Mr Edward Buckland), both professional directors provided by Bedell Cristin Trust Company, took place the same day and entry into the CDO Transaction was approved. According to the minutes of this meeting:

On 25 June 2007, CSSU’s Mr Fiachra O’Driscoll sent Mr Zimmermann a draft of the Series Memorandum for the CDO Transaction (“the 25 June Draft”). The 25 June Draft was marked as a draft, “Subject to completion dated […] June 2007 PRELIMINARY SERIES MEMORANDUM” and bore, in red letters, the following warning:

“The information in this preliminary prospectus is not complete and may be changed. This preliminary prospectus is not an offer to sell these securities, and it is not soliciting an offer to buy these securities, nor shall there be any sale of these securities in any jurisdiction where the offer or sale thereof is not permitted, nor has it been approved by the Irish Stock Exchange or the Irish Financial Services Regulatory Authority”.

The 25 June Draft also:

On 5 July 2007 and 7 July 2007, Mr Thomas Schirmer, an in-house lawyer employed by IKB CAM, provided CS with detailed comments on the drafting of the 25 June Draft.

On 9 July 2007, Mr Zimmermann informed Mr Stroobandt that the Rhineland SPC which would purchase the Notes would be L30. That same day Mr Jawurek responded to Mr Schirmer setting out CSSU’s responses to Mr Schirmer’s comments on the 25 June Draft, which included suggestions for changes to that document. Mr Schirmer replied with some further comments on 10 July 2007. These exchanges led CSSU to prepare a further draft of the Series Memorandum, which Mr Jawurek sent to Messrs Schirmer and Zimmermann on 11 July 2007 (“the 11 July Draft”). The 11 July Draft again contained (inter alia) the statements quoted at paragraph ‎78 above, save that the passage corresponding to that quoted from page 2 of the 25 June Draft omitted the words “It is expected that”.

Some limited changes were made to the 11 July Draft on 12 July 2007 and 13 July 2007 following further discussions between Messrs Jawurek, Schirmer and Zimmermann, including to reflect the fact that Standard & Poor’s Rating Services (“S&P”) would rate the Notes “AAA/watch negative” on the issue date rather than “AAA”.

On 13 July 2007, Mr Jawurek sent Mr Zimmermann (and others) what his email described as “the final, black Series Memorandum”.

The definitive version of the Series Memorandum, as filed with the Irish Stock Exchange, was dated 25 July 2007.

Meanwhile, in parallel with its interactions with IKB CAM, CS was in contact with Moody’s and S&P in relation to their provision of ratings for the CDO Transaction.

As far back as 27 April 2007, Mr Yale Chang had individually informed S&P’s Mr Christopher Meyer and Moody’s Ms Min Xu that what would become the CDO Transaction was expected to proceed. His emails to each were in closely similar form, outlining the capital structure of the proposed CDO Transaction, explaining that the expected closing date was 22 May 2007 and attaching some analysis of the transaction in spreadsheet form. However, the spreadsheet attached to the email to Mr Meyer had been generated using S&P’s “CDO Evaluator” software, whilst that attached to the email to Ms Xu made use of Moody’s equivalent, CDOROM.

Ms Xu quickly responded, requesting the CUSIPs (unique identifying numbers assigned to securities) for the Reference Obligations. These were provided by Mr Chang the same day.

Having analysed the portfolio, Mr Meyer provided S&P’s preliminary feedback on the Reference Obligations on 1 May 2007 in the form of five spreadsheets. In later exchanges on 11 June 2007, Mr Meyer confirmed that the results had not changed because there had been no “rating migration” affecting the Reference Obligations since early May.

The next material interaction with the rating agencies took place at the time of the 25 June Draft, when Mr Meyer and Ms Xu provided input into the drafting of the Series Memorandum. In particular, on 25 June 2007 Mr Meyer asked that a notice of any “Adjustment” (as defined) should be required to be sent to a specified S&P email address, whilst Ms Xu provided various comments on rating-related aspects of the drafting. These comments were reflected in the 11 July Draft. Ms Xu then provided further drafting comments on the 11 July Draft on 12 July 2007, which were again reflected in subsequent versions of the Series Memorandum.

On 12 July 2007, Moody’s provided a signed rating letter to be held in escrow whilst S&P sent CS a draft rating letter for the CDO Transaction.

The CDO Transaction was entered into during a telephone call between Mr Jawurek and Ms Inna Ivanova of Société Générale, which took place at approximately 2pm UTC on 12 July 2007, with a settlement date of 13 July 2007. A written confirmation of the trade was faxed to IKB and L30 at 3.02pm on 12 July 2007. As reflected in the trade confirmation, L30 purchased the Notes from CSSU, which acted through its agent, CSSEL.

Finally, on 13 July 2007:

In light of S&P’s decision to impose the “watch negative” on its rating, Mr Jawurek agreed with Mr Zimmermann immediately prior to closing that the attachment point of the Notes would be increased if, following S&P’s review, that action was necessary to raise the S&P credit rating to “AAA”. However, S&P’s review resulted in the removal of the “watch negative” and the affirmation of its “AAA” rating on 24 July 2007 without the need for any such step.

From 27 to 29 July 2007, less than a fortnight after the conclusion of the CDO Transaction, IKB experienced what it later described as a “crisis weekend”. There were doubts about IKB’s liquidity situation and, to stabilise the situation, the German public sector banking institution Kreditanstalt für Wiederaufbau (“KfW”) assumed (inter alia) IKB’s liquidity commitments to the Rhineland Programme, amounting to €8.1 billion. Four of IKB’s board members left the bank as a consequence of this crisis. PricewaterhouseCoopers was instructed to undertake a special review of IKB’s business and made findings critical inter alia of the Rhineland Programme’s risk management, portfolio monitoring and excessive concentration on subprime investments.

As a result of the rescue and its assumption of IKB’s obligations, KfW became a liquidity facility provider to the Rhineland SPCs, including L30, and their largest creditor.

Following a subsequent restructuring of the Rhineland Programme in 2012, KfW became the sole liquidity provider and sole substantial creditor to the Rhineland SPCs.

According to notices issued to the Noteholders by Magnolia, between about the beginning of October 2008 and the end of April 2009 CSAG delivered to Magnolia a series of Floating Amount (“Writedown”) Notices relating to (inter alia) certain of the Reference Obligations.

Between L30’s purchase of the Notes and 28 April 2010, certain Writedowns (i.e. realised losses in view of the non-performance of the Reference Obligations, as defined in page 144 of the Series Memorandum dated 25 July 2007) occurred in respect of Reference Obligations. These resulted in the triggering of CDS credit events. Consequently, Magnolia, as Protection Seller, was obliged to make payments to CSAG (through its Cayman Islands Branch) as Protection Buyer, with the result that the Aggregated Outstanding Principal Amount of the Notes held by L30 was reduced. As a result of the performance failures amongst the 100 Reference Obligations, that Aggregated Outstanding Principal Amount reached zero in March 2009, at which point the original USD 100 million principal value had been paid to CSAG under the CDS.

By the end of April 2009, therefore the “Aggregate Outstanding Principal Amount” of the Notes (as defined in Condition 7(d) of the Notes) had been reduced to zero, at which it remained thereafter. On 26 April 2010, CSAG delivered to Magnolia a Notice of Redemption, notifying it that the Notes had been redeemed without payment. The Notes were delisted from the Irish Stock Exchange with effect from 28 April 2010, as it was a mandatory redemption condition of the Notes that they would be redeemed without payment in these circumstances (see Condition 7(b)(4) of the Notes in the Series Memorandum).

In 2008, KfW put in place a press monitoring and/or alert system “to capture information on developments relating to the financial crisis”. As will be explained further below, from 2008 onwards it also instructed both UK and US lawyers to, as Mr Bulgrin put it, “provide advice on potential legal avenues in both jurisdictions including what potential claims may be available at a general level in connection with SPC assets”, and later to provide updates on litigation relating to the financial crisis.

From about 2009 the US Department of Justice (“DoJ”) began to take an interest in deals involving RMBS, setting up an RMBS Working Group. In due course the investigations encompassed CS.

In January 2017 the DoJ announced a $5.28 billion settlement with CS, related to CS's conduct in the packaging, securitization, issuance, marketing and sale of RMBS between 2005 and 2007.

That settlement and the press release which accompanied it (“the 2017 Press Release”), formed an important plank of L30's case. It might indeed be said to be a foundational document for L30, with its case in closing starting with another detailed look at the document. L30 says that the Statement of Facts (“SoF”) which accompanied the settlement contains admissions of fraud. It takes the DoJ Settlement and those alleged admissions as its starting point. It says that “CS’ protestations of innocence ring hollow in light of that settlement.”. The contents of those documents are therefore of some moment.

The 2017 Press Release, dated 18 January 2017, said this:

“Credit Suisse Agrees to Pay $5.28 Billion in Connection with its Sale of Residential Mortgage-Backed Securities

The Justice Department announced today a $5.28 billion settlement with CS related to CS’s conduct in the packaging, securitization, issuance, marketing and sale of residential mortgage-backed securities (“RMBS”) between 2005 and 2007. The resolution announced today requires CS to pay $2.48 billion as a civil penalty under the Financial Institutions Reform, Recovery and Enforcement Act (“FIRREA”).

Today’s settlement underscores that the Department of Justice will hold accountable the institutions responsible for the financial crisis of 2008” said Attorney General Loretta E Lynch. “Credit Suisse made false and irresponsible representations about residential mortgage-backed securities, which resulted in the loss of billions of dollars of wealth and took a painful toll on the lives of ordinary Americans.

Credit Suisse claimed its mortgaged backed securities were sound, but in the settlement announced today the bank concedes that it knew it was peddling investments containing loans that were unlikely to fail” said Principal Deputy Associate Attorney General Bill Baer. “That behavior is unacceptable. Today’s $5.3 billion resolution is another step towards holding financial institutions accountable for misleading investors and the American public.

This settlement includes a statement of facts which CS has agreed. That statement of facts describes how CS made false and misleading representations to prospective investors about the characteristics of the mortgage loans it securitized”.

The SoF accompanying the announcement, agreed by CS, contains the following passages (Footnote: ²):

“[1] From May 2005 through 2007, Credit Suisse Securities (USA) LLC, itself and through certain of its U.S affiliates (collectively “Credit Suisse”, securitized hundreds of thousands of residential mortgage loans into residential mortgage-backed securities (“RMBS”). It sold these RMBS for tens of billions of dollars to investors. In marketing and selling its RMBS, Credit Suisse made numerous representations about the quality and characteristics of the underlying loans. These representations were made to RMBS investors and potential investors, ratings agencies and others. As described below, in the diligence process, Credit Suisse repeatedly received information indicating that many of the loans reviewed did not conform to the representation that would be made by Credit Suisse to investors about the loans to be securitized. …

II. Credit Suisse’s Representation to Investors

[4] In connection with its RMBS offerings on principal transactions, Credit Suisse made representations about the loans it securitized. Those representations often varied from securitization to securitization. Credit Suisse made representations that, among other things:

1.

The loans were originated generally in accordance with applicable underwriting guidelines, with exceptions to those guidelines being made when sufficient compensating factors were demonstrated by a prospective borrower.

2.

For each loan, a determination had been made by the originator that the borrower had the ability to repay the monthly obligations on the loan and other debts.

3.

Beginning in early 2006, Credit Suisse modified its representation in certain offering documents to state that each Correspondent loan was “in fact” originated in accordance with Credit Suisse’s underwriting guidelines or guidelines that did not vary materially from such guidelines and that Credit Suisse “employed…certain quality assurance procedures designed” to ensure that such loans were originated in accordance with the underwriting guidelines.

4.

Each loan had been originated in compliance with all federal, state and local laws and regulations, including all predatory and abusive lending laws.

5.

For each loan, the adequacy of the mortgaged property as security for repayment generally had been determined by an appraisal in accordance with pre-established appraisal procedure guidelines for appraisals established by or acceptable by the originator.

6.

The loans had various characteristics, including certain loan-to-value (LTV) ratios disclosed to the trustee in Mortgage Loan Schedules for each loan. The distribution of these LTVs in five-to-ten percent bands was also included in prospectus supplements. Other characteristics, such as borrower FICO scores, were disclosed as well. Credit Suisse represented to the trustee that these characteristics about the loans were complete, true and correct for the loans as of the cut-off date for the securitization.

7.

None of the loans had a loan-to value ratio or, with respect to second lien mortgages, a combined loan-to-value ratio, in excess of 100 (i.e., none of the loans were “underwater”).

8.

Credit Suisse ‘will not include any mortgage loan in [the RMBS] if anything had come to [its] attention that would cause it to believe that the representations and warranties of a seller will not be accurate and complete in all material respects in respect of the related mortgage loan’.

[5] Credit Suisse also represented that if it later learned that any loan RMBS breached its representations in a material respect, it would cure the breach or repurchase or substitute that loan from the RMBS.

[6] In various marketing materials, Credit Suisse made additional representations about the loans it securitized into RMBS, and its process for reviewing loans before it selected them to securitization. In these marketing materials, Credit Suisse told investors and others, among other things:

1.

Credit Suisse conducted a ‘rigorous due diligence process’ for reviewing loans;

2.

All final loan decisions were made by Credit Suisse senior underwriters, not third-party contractors.

3.

Credit Suisse conducted ‘quality control’ reviews on its loans after purchasing or originating them.

[7] As Credit Suisse employees discussed internally, Credit Suisse made these representations in order to ‘convince investors and insurers that [its] disciplined approach to underwriting [its] multi-layered origination process, and an eligibility matrix that utilize[d] a more layered risk of performance will result in an acceptable level of delinquencies for [its] products’. …

III Credit Suisse's RMBS Business

[8] …The senior traders reported that a number of the loans had gone delinquent or into foreclosure within a few months after Credit Suisse sold the RMBS to investors. The Co-Head of the Structured Products Group responded, ‘of course we would like higher quality loans. That’s never been the identity of our conduit, and we’re becoming less and less competitive in that space”. Around the same time, one of the senior traders described his situation in a Bloomberg message: “I figure I could just capitulate… and keep buying all [these] crappy loans, but that would be suicide. And we still have almost $2.5B of conduit garbage to still distribute.

[9] In October 2005, Credit Suisse’s Head of Credit and Underwriting wrote to two senior Credit Suisse traders and others, ‘we are selling and securitizing loans with missing docs all the time through the other desks’ (Missing documents included borrower credit reports, verifications of borrower income and assets, property appraisals and legal compliance documents such as HUD-1 statements and Truth in Lending Act disclosures.) The Head of Credit and Underwriting continued: ‘it only becomes an issue when we are asked to repurchase the loan or we receive a complaint from a customer’.

[10] In April 2007, Credit Suisse’s Co-Head of Transaction Management wrote to the Head of Credit and Underwriting and two senior traders that loans with potential defects ‘pile up in inventory and they won’t be put back to originators (we are disproportionately the originator) So my theory is: we own the risk, 1 way or another, I don’t want to securitize the loans that are obviously Section 32 cost, Georgia Fair Lending disasters with unlimited assignee liability and criminal sanctions. But I am inclined to securitize loans that are close calls or marginally non-compliant and take the risk that we’ll have to repurchase if we can’t put them back, rather than adding to sludge in inventory. I think when we review the first month report, we’ll see how it shakes out in actual practice’. One of the senior traders responded, ‘agree’. …

IV. Credit Suisse’s Credit and Compliance Review of Bulk Loans

…

[15] Credit Suisse told investors and rating agencies that it used a propriety risk analysis model in its sampling process to identify high-risk loans in Bulk pools. The purpose of this model was to flag these high-risk loans for inclusion in due diligence samples. In many instances, however, Credit Suisse excluded loans that its own model has flagged as high risk from its samples. Through this practice, Credit Suisse approved some of these flagged loans without any credit or compliance due diligence review. Credit Suisse did not disclose this practice to investors.

Examples:

[16] 1. In December 2006, Credit Suisse bid on a Bulk pool of approximately 10,000 loans originated by Countrywide Home Loans (“Countrywide”). Before the bid, the senior Credit Suisse trader bidding on the pool wrote in an email, ‘amazing that 8% of the pool is already delinquent after 10 months, when a large part of it is Jumbo A!!!’. He stated that the pool was ‘obviously plugged with the worst paper’. The senior trader then bid to purchase this pool and, when he was informed that Credit Suisse was the winning bidder, wrote in an email, ‘I’m ecstatic’ with this trade.

[17] Credit Suisse selected fewer than 10 percent of these loans for due diligence review. Reports from Credit Suisse’s due diligence vendors showed that approximately 85 percent of the loans in this sample violated Countrywide’s underwriting guidelines and/or applicable law. These loans included, for example, loans whose borrowers had debt-to-income ratios that were more than 10 percent higher than the amount allowed by underwriting guidelines; loans that were missing appraisal documents and loans flagged for ‘potential fraud’. The Credit Suisse due diligence manager overseeing this review reported that Credit Suisse’s vendor ‘has found a number of guideline exceptions (fico score exceptions LTV [loan-to-value] exceptions, Program exceptions) [.] However, I am not rejecting for these items’. The Co- Head of Credit Suisse’s Transaction Management department wrote to the due diligence manager ‘thanks for working thru this mess. If it helps, it looks like we will make a killing on this trade’.

[18] Credit Suisse securitized over half of these loans into various RMBS it then sold to investors. Credit Suisse additionally securitized an additional $1.5 billion worth of unsampled – and therefore unreviewed – loans from this pool into various RMBS it then sold to investors.

[19] In October 2005, Credit Suisse bid on a Bulk pool of approximately 4000 loans from Wells Fargo Bank. Credit Suisse’s risk analysis model flagged approximately 2900 loans in the pool for heightened due diligence due to their risks. Among other things, Credit Suisse’s risk analysis model estimated that approximately 2000 loans in this pool had a greater-than-5 percent likelihood that the borrowers would miss three or more monthly payments in a row within their first twelve months and flagged these loans for heightened due diligence. Credit Suisse represented to ratings agencies. ‘To choose the sample from the Wells pool, the loans first were run through our propriety due diligence sampling model, which identifies loans for review based on credit characteristics, compliance (via location of subject loan) and soft market/fraud areas. Those chosen for full credit and compliance review consisted generally of those loans which the model found displaying the most risky characteristics’.

[20] But Credit Suisse wrote to Wells Fargo that ‘We are working to reduce this due diligence sample” and “We have worked to reduce this sample further’ and sent its vendor a list of over 1,500 loans – over 58 percent of its sample – with the instructions ‘REMOVE from sample’. Over 2,900 of the loans in this pool were flagged by its model, but Credit Suisse excluded approximately 60 percent from the due diligence sample it sent to its vendor. Credit Suisse’s Head of Credit and Underwriting wrote in an email that these removed loans ‘would not receive a credit, compliance, and in-file appraisal review [.]’….

V. Credit Suisse’s Credit and Compliance Review of Conduit Loans

[22] Credit Suisse did not review Conduit loans in the same way that it reviewed Bulk loans. Instead of using Bulk due diligence vendors, Credit Suisse contracted with different third-party vendors known as “fulfillment centres” to conduct this review. CS designed the Conduit such that these fulfilment centres had the ability and authority to approve loans for purchase. When fulfillment centres approved loans, Credit or Underwriting employees rarely reviewed the loans or the approvals.

Credit Suisse represented to investors, ratings agencies and others that ‘Credit Suisse senior underwriters make loan decisions, not contracted due diligence firms’. For Conduit loans, these representations were false. As Credit Suisse employees discussed in internal emails, the loan review and approval process for Conduit loans was “virtually unmonitored”. Fulfillment centres often cleaned loans that were not subsequently reviewed by Credit Suisse senior underwriters prior to Credit Suisse purchasing the loans.

A March 2006 evaluation by Credit Suisse of one of its fulfillment centres similarly reported that ‘There are serious concerns as to compliance[.]’. A senior Credit Suisse trader described the Conduit in a November 2006 email: ‘we make these underwritings exceptions and then we have liability down the road when the loans go bad and people point out that we violated our own guidelines…. The fulfillment process is a joke’. …

VI. Credit Suisse’s Valuation Review of Loans

[29] Credit Suisse reported property values in its offering documents, as part of a ratio called the loan-to-value (LTV) ratio. This ratio reflected the size of a loan compared to the ‘value’ of the underlying property. For purchase transactions, the “value” was established by using the lower of the appraisal or the purchase price at the time of origination. For refinance transactions, the value was established by an appraisal.

…

[32] During this period, Credit Suisse employees were aware that the LTVs may have been calculated using appraisals with values that were inflated, and that Credit Suisse’s approach could lead to the acceptance of inflated appraisals. As Credit Suisse’s Head of Credit and Underwriting wrote in a November 2006 email ‘originators can get away with potential appraisal inflation as long as it stays within 15%’.

[33] Credit Suisse employees expressed concern that loans with inflated appraisals were being approved as within its tolerance. For example, in August 2006, Credit Suisse’s Head of Credit and Underwriting emailed two CS senior traders in connection with loans Credit Suisse was buying from Accredited Home Lenders. He wrote, ‘20+% of their loans have value issues >20% off – that is unheard of. [Accredited] is acknowledging that their values are inflated. There should be a 0% variance from an originator[’s] standpoint. [Accredited] is saying that [it] knows and is okay with [its] values being off by up to 15%. Some would say this is predatory and criminal. How would investors react if we say that 20% of the pool have values off by 15%? If we are comfortable buying these loans, we should be comfortable telling investors.’

[34] From May 2005 through 2007, Credit Suisse received reports from vendors that it might have been acquiring and securitizing loans with inflated appraisals. During this period, despite those reports, Credit Suisse chose to continue to apply for a 10-15 percent tolerance. In early 2007, Credit Suisse’s Head of Credit and Underwriting wrote an email proposing to tighten Credit Suisse’s variance ‘tolerances’ to 10 percent for loans with LTVs below 75, and to 5 percent for loans with LTVs above 75, to ‘address the issues of inflated appraisal values and one of the issues that create the high severity numbers we are seeing today’. His boss, the Head of the Conduit group, responded in an email by writing, ‘I’d like to see what % of our current production would be effected [sic] by the new guidelines versus the old guidelines. Also would like to see statistics on BPO variances over the last 6 months and values we ultimately used. I’m trying to gauge the volume impact and level of customer interaction that will be required with these changes’. Credit Suisse’s internal analysis showed that 30-40 percent of the loans it reviewed and approved in the prior six months would have been rejected if it had used a 5 percent tolerance. Credit Suisse then did not implement any tolerance changes for two months. It then reduced tolerances only for loans with LTVs over 90 to 10 percent, but still left other tolerances at 15 percent.

VII. Credit Suisse’s Whole Loan Business

[35] On occasion, Credit Suisse sold pools of mortgage loans to other banks. Unlike RMBS investors, which generally were, as a practical matter, not able to conduct due diligence on the loans being securitized, these other banks conducted their own due diligence on the Credit Suisse loans before buying them. Those other banks found that loans that Credit Suisse marketed for sale violated underwriting guidelines or applicable law and refused to buy them.

[36] Credit Suisse employees discussed these rejections by other banks. For example, in October 2006, a senior CS trader, wrote to the Co-Head of Credit Suisse’s Structured Products Group; ‘ Our fulfilment process is a major problem. [The other bank] again came back with an embarrassing number of diligence kicks this month. You’ll remember that last month they pointed out the loan file with 3 different applications, with 3 different stated incomes (amazingly, not the first time this has happened). This month’s kicks include an employment verification kickout where the borrower never had a job, but rather had a letter stating his intent to get a job. Another employment verification kick where the employment letter stated the borrowers’ employment would be terminated in July. Another kick where the borrower had large deposits to meet reserve requirements but no one could verify the source of funds. You get the idea. The point is that [the bank] is (a) our best customer (b) only scratches the surface of our conduit via a 20% sample. If their results are in any way representative of our compliance with our reps and warrants, we have major problems’. But rather than holding those loans in its own inventory, Credit Suisse securitized certain of these loans into its RMBS.

…

VIII. Credit Suisse’s Quality Control

[39] Credit Suisse informed various investors and ratings agencies that it conducted a “quality control” review on a 3-5 percent sample of the loans it acquired for securitization. Credit Suisse contracted with vendors to provide it with quality control (QC) reports on a monthly basis. From May 2005 throughout 2007, Credit Suisse’s vendors reported, overall, that more than 25 percent of the loans that they had reviewed for quality control were designed “ineligible” because of credit, compliance and/or property defects.

[40] Credit Suisse employees observed in emails that quality control results showed problems with the loans. A senior Credit Suisse trader wrote in a January 2007 email to one of the Co-Head of the Structured Products Group, regarding Correspondent channel loans; when we have a competent QC firm do an underwriting review, they flag all kinds of errors that our fulfilment [sic] centers did not catch. Moreover, our very own underwriting group agrees with the QC firm rather than our fullfillment [sic] center. I think a lot of the problems stem from the fact that our conduit and underwriting group send mixed messages. One [sic] on the one hand they profess horror that our deals are defaulting like banana republics, but on the other hand every time we try to tighten up our underwriting processes they push back claiming it makes us uncompetitive. Here are our takeaways: 1) Our fullfillment [sic] process is broken. Either they don’t know how to do their jobs, or they don’t report defects because the conduit doesn’t want them to create waves. 2) Our underwriting group needs independence. 3) Our conduit needs to spend less time marketing to sales, and more time looking at the gross operation defects of our business’.

[41] In emails Credit Suisse’s Co-Head of Transaction Management expressed concern that the quality control results could serve as written record of defects, and sought to avoid documented confirmation of these defects. As Credit Suisse’s Co-Head of Transaction wrote in a January 2007 email: ‘I am concerned that we have so many piecemeal QC/DD processes, internally and externally, that some of these reports could serve as no more than a written record of some potential defect that would require a repurchase [of loans] from a [RMBS] deal, but we do not have sufficient resources to verify the information or complete the repurchase. This is one of the concerns we’ve always had w/[or vendor]’s role in the deals – we want to leverage your resources to improve the performance of our deals and improve our business, but not receive so much haphazard information that we do no more than create a record of potential problems’.”

I have quoted the SoF at very great length because it was a central plank of L30’s case (see in opening: “L30’s case in relation to the RMBS Misconduct takes as its starting-point the admissions made in the SOF”, or in oral closing its description of the SoF as being of “foundational significance”) and because its status was hotly in issue. L30 says that it follows that the RMBS Misconduct which it alleges “plainly did not happen accidentally”. It says that in the light of what it sees as admissions it is remarkable that CS continues to deny the making of actionable representations or falsity. CS says that the suggestion that there were admissions is misconceived.

Despite the characterisation of CS's position on the SoF as “remarkable” L30's written submissions did not press the status of the SoF with any vigour, and rightly so. Two points need to be made. The first is that there was a tendency to elide the SoF with the press release which accompanied it. The 2017 Press Release was not a document authored by CS, nor agreed to by CS. Its contents cannot lie at CS's door. The 2017 Press Release has to be treated with some caution. While emanating from the US Department of Justice it is designed to report success in a digestible and easily reportable form. It is commentary from one side; it is not fact. It is also perhaps worthy of note that however scathing The 2017 Press Release was about CS's business it did not use the language of fraud. In this an interesting distinction can be noted between this press release and the one which the DoJ issued the day before in relation to the business of Deutsche Bank which said in terms:

“’In the Statement of Facts accompanying this settlement, Deutsche Bank admits making false representations and omitting material information from disclosures to investors about the loans included in RMBS securities sold by the Bank. This misconduct, combined with that of the other banks we have already settled with, hurt our economy and threatened the banking system,…

Deutsche Bank knowingly and intentionally securitized loans originated based on unsupported and fraudulent appraisals. Deutsche Bank knew that mortgage originators were ‘giving' appraisers the value they want[ed]’ and expecting the resulting appraisals to meet the originators' desired value, regardless of the actual value of the property. Deutsche Bank concealed its knowledge of pervasive and consistent appraisal fraud, instead representing to investors home valuation metrics based on appraisals it knew to be fraudulent. Deutsche Bank misrepresented to investors the value of the properties securing the loans securitized in its RMBS and concealed from investors that it knew that the value of the properties securing the loans was far below the value reflected by the originator's appraisal….

Deutsche Bank will also provide $4.1 billion in the form of relief to aid consumers harmed by its unlawful conduct.”

The second point is that the SoF is also not a CS production and that CS's agreement goes only to the facts in the SoF. So when emails are quoted, CS is not admitting the truth of the contents of the email, it is admitting the truth of the fact of the sending of an email which said that. Nor are the facts admitted to have English law legal consequences; because of course a word may mean one thing in a US context, but mean something entirely different in an English law context. So just because the SoF says that representations were made does not mean that as a matter of English law actionable representations were made.

As one scrolls through the SoF it is possible to see that what CS has admitted via the SoF is – save in one respect – not anything which amounts in law to an actionable representation. That one respect is that CS has agreed that there is an admission that a representation was made that all final loan decisions were made by CS senior underwriters, not third party contractors (the “Approval Representation”) and that that representation was false.

Therefore save in that one respect the SoF, though interesting, has no legal consequences. L30 cannot make good a misrepresentation case by reliance on it. Nor (even if falsity were demonstrated by the examples) can it be inferred that because there are examples of certain things within the SoF there are further instances which support L30’s case of systemic or serial falsity. The SoF shows what it on analysis shows and no more.

At the same time it is clear from the text of the SoF, from the documents quoted within it – and from the events from 2008 onwards – that there were problems with CS's RMBS business. The issues as to RMBS Misconduct hinge on the seriousness of those problems and CS’s treatment of them and are considered further in due course.

The SoF was published in 2017. It was Mr Bulgrin’s evidence that it was alerted to L30 as a part of the ongoing monitoring by its lawyers. It does not appear that it was considered a matter of huge urgency. Indeed the fact that Mr Bulgrin and Mr Hollywood both gave evidence that they only read it a year or so later suggests that there were no meetings or discussions of it for some time.

The evidence was that they were prompted to read it by an advance which was made by either or both of David Doble of David Doble Law and Reynolds Porter Chamberlain LLP (“RPC”).

It may or may not be of some significance that on 2 March 2018 the Court of Appeal handed down its judgment in Property Alliance Group v Royal Bank of Scotland [2018] 1 WLR 3529. That judgment upheld the possibility of a limited implied representation in the context of complex banking offer documents, albeit in the context of swaps, not CDOs.

By July 2018 RPC and David Doble Law had prepared a case overview for a claim by L30. That was provided by Mr Doble to Mr Bulgrin and was later referred to the L30 Board. The document was prepared in contemplation of litigation and privilege was claimed over it on that basis. It formed the basis for the claim originally brought. At this stage RPC were not instructed; it was prepared “as a preliminary step which led to the retainer of RPC”. RPC were not paid for it.

Following discussions by KfW the possibility of a claim and the funding of a claim, the advice memo was referred to the L30 Board in November 2018. There was an unminuted oral discussion on 5 November, which was alluded to in the minutes. The result was a decision either to enter into a standstill or to issue proceedings. There appears to have been a draft claim form available as Mr Hollywood signed off on the statement of truth.

On 12 November 2018 RPC were engaged by L30. A claim form was issued on 15 November 2018.

On 21 December 2018 L30 sent a Letter before Action. It stated (inter alia):

“On 18 January 2017, the US Department of Justice (“DOJ”) and CSUSA entered into a settlement (the “DOJ Settlement” - enclosed), in which CSUSA acknowledged the facts set out in a statement annexed to the DOJ Settlement at Annex 1 (the “Statement of Facts”) 

 The Statement of Facts sets out representations made by CSUSA to its RMBS investors in respect of the Covered Conduct, and facts which demonstrate that CSUSA knew at the time those representations were made that they were false, and therefore fraudulent (the “RMBS Fraud”). Put simply, CSUSA was (i) packaging residential mortgages in RMBS and selling those securities to investors, and (ii) knowingly making false representations about the quality and characteristics of those mortgage loans to RMBS investors. This was a systematic and serious fraud conducted by the bank.

The relationship between the DOJ Settlement and the Magnolia Notes

Seven of the Credit Suisse originated RMBS included in Annex 3 to the DOJ Settlement as the subject of RMBS Fraud admissions were subsequently included in the Magnolia Notes’ reference portfolio. The Magnolia Notes therefore included as part of its reference portfolio the very RMBSs to which Credit Suisse’s RMBS Fraud in part related.  

 The fact that these seven RMBS were known by CSUSA to have been the subject of its now admitted fraudulent activities was not disclosed in the Series Memorandum for the Magnolia Notes, nor was it otherwise disclosed to Loreley 30 or its investment advisor, IKB Credit Asset Management GmbH (“IKB-CAM”) which dealt with Credit Suisse in relation to the Magnolia Notes on behalf of Loreley 30. 

 The admission by CSUSA of the facts set out in the Statement of Facts is incontrovertible evidence that the Defendants knew at the time the Magnolia Notes were being arranged and marketed that it was not conducting itself honestly in relation to the underlying subject matter of the transactions (namely the pooling, structuring, arranging, formation, packaging, marketing, underwriting, sale or issuance of the RMBS listed in Annex 3 and included within the Magnolia Notes reference portfolio).

There can be no question that the RMBS Fraud (which CSUSA has admitted was knowingly dishonest) was known to the [Credit Suisse Structured Credit Group] and to the corporate entities used to carry out the business of the [Credit Suisse Structured Credit Group]. 

 Moreover, the Defendants could only have procured the necessary AAA rating for the Magnolia Notes from S&P and Moody’s by concealing its RMBS Fraud from the credit ratings agencies and without that, Loreley 30 would not have purchased the Magnolia Notes. 

Loreley 30’s Claim

In brief outline, we summarise below the principal basis on which these claims arise.

Fraudulent misrepresentation

Loreley 30 purchased the Magnolia Notes in reliance upon representations made by CS Europe and/or CSI and/or CSUSA relating to the quality and characteristics of the RMBS securities that comprised the Magnolia Notes’ reference portfolio. In particular, in putting forward the transaction that was structured and arranged the relevant Credit Suisse Defendants made implied representations including (among other matters and in broad summary) that: 

(i)

They had no reason to believe that the reference portfolio would contain any RMBS that was known by Credit Suisse to contain underlying loans that failed to conform with the relevant offer documents or underwriting criteria; 

(ii)

In respect of any RMBS contained in the reference portfolio that Credit Suisse had itself structured or distributed, Credit Suisse had acted honestly in structuring and distributing the RMBS and was not engaged in the practice of knowingly securitising loans that failed to conform with offer documents or underwriting criteria;  

(iii)

In respect of the ratings of both the Magnolia Notes, and the RMBS contained in the reference portfolio for which Credit Suisse had itself procured ratings, that Credit Suisse had acted honestly in procuring the relevant ratings from the credit rating agencies”.  

Ultimately at trial all of this played out into a claim which was not entirely straightforward. To flesh out slightly the skeletal account given in the Introduction to the judgment:

L30 seeks to establish one or more elements of the RMBS Misconduct. That misconduct is then said to falsify one or more of the CDO Representations.

Some impression of the scale of the trial is perhaps worth giving. The Trial Bundles were electronic. They ran between Bundle AA and Bundles V (V1-6). About 120 English authorities were cited. There were 991 tabs of CDO documents (Bundle H) and over 2000 RMBS documents (Bundle J). There were 28 bundles of deposition documents (P1-28).

In terms of written submissions, I was assisted by hugely detailed written submissions, to which I cannot begin to do justice in this judgment. Written openings ran to 400 pages. Written Closings ran to 300 pages. All were liberally footnoted, and included detailed summaries of materials which could easily, if expanded, have added more hundreds of pages to the material before me. This judgment is therefore very much a process of concentrating on the determinative issues and key facts.

While the trial was not always conducted in the most co-operative of spirits inter se there was key and useful co-operation to assist the court in a number of places, for which I am duly grateful. I have already mentioned the factual summaries provided by the parties. There was also very sensible co-operation on the subject of Irish Law, leading first to the agreement that such issues of Irish Law as remained could be decided on the basis of submissions without the need for the experts to be called (in line with Commercial Court Guide paragraph H3), and later to a further narrowing of those issues for decision.

I am particularly pleased to say that (following a certain amount of judicial encouragement at the PTR) the parties found it to be possible to divide the oral advocacy so that nearly all members of the counsel teams on both sides conducted meaningful advocacy. For L30 Mr Hobson dealt with limitation and Mr Woolgar with conspiracy and Irish Law. Mr Rudzitis was acknowledged to have provided key help in the written documents. For CS Mr Sher cross-examined a number of the witnesses and dealt with understanding/reliance in relation to misrepresentations, Mr Brock dealt with RMBS Misconduct 5 and 6 and Mr Field with negligence and unlawful means conspiracy.

As will be apparent from the comments below on those topics it was clear that a number of these areas very much benefitted from the full and detailed consideration and laser focus which the junior counsel responsible were able to give them.

Dr Bauknecht, the Chief Economist of IKB, spoke predominantly to understanding of the alleged representations and reliance. He was, in his written evidence, realistic about his recollections, though he tended to slip into reconstruction in places. Orally he attempted to be more optimistic about his recollections. But he was generally a reliable witness, who although careful was prepared to make concessions where appropriate.

There were problems which were not of his making with his statement. His statement was one of those where (because it was not originally produced as a trial witness statement) the PD57AC process was not followed and it rapidly became apparent that his statement had been produced by sending him material which he had not previously seen. The result was that his witness statement did not record his “unrefreshed” evidence, and also did not record evidence based on his own knowledge. In addition it was apparent that at least at some points the draft was “lawyer-side” drafting, with an identical passage appearing in another witness’s evidence. It became apparent that he had not given unrefreshed evidence in preparation of the drafting of his witness statement and it was established in evidence that there was no conversation with his legal team about his recollection of the transaction, before he had seen and reviewed around 100-150 pages of documentation in relation to it. As a result there was a significant challenge to his recollection, which I will consider further below.

Mr Shane Hollywood, a qualified Jersey lawyer, was a director of L30 as well as a number of other Loreley companies. He spoke to the management of L30, its relationship with KfW and the way in which L30 made decisions. He also dealt with the question of knowledge for the purposes of limitation. There was an issue as to whether, in dealing with the issue whether L30 reasonably could have discovered sufficient information to start the limitation clock, his statement waived privilege in two paragraphs of the served statement. It fairly plainly did do so and an agreement was reached that his evidence as tendered should omit those portions of his evidence. Mr Hollywood had a calm and thoughtful demeanour. As is probably to be expected of a lawyer he was careful in his answers. In broad terms earnest and helpful, he avoided overstating his knowledge and he frequently gave evidence that he had no recall of matters which were put to him. He was also prepared to make concessions after due thought. At the same time, while never unreasonable, he was reticent to make concessions which were not in L30’s interests. An example was in relation to a passage of his witness statement which neglected to mention a fact he knew and which put a rather different complexion on his knowledge. He required considerable pressure to be applied before he accepted (even with a degree of qualification) that his evidence was in that respect misleading.

Mr Thomas Schirmer, a German qualified lawyer was IKB CAM's Legal Counsel at the time when the transaction was entered into. He therefore spoke to reliance, in the period 15 June to 12 July 2007, particularly in relation to the Category C and D representations. His statement did not comply with PD57AC because of the stage at which it was submitted (it having been drafted at an interlocutory stage of the dispute) and like that of Dr Bauknecht strayed into reconstruction. While at that stage it was not necessary for it to comply with PD57AC it is regrettable that the evidence of a witness who spoke to reliance should have been polluted by that process, which involved showing him various documents he had not seen at the time. Mr Schirmer was a frank and believable witness and while allowance fell to be made for him giving evidence in English, he was so clear that this was barely necessary. While (like the other witnesses) being clear about the limits of his actual recollection, he was very diligent in trying to make clear what he did recall or understand.

Mr Bulgrin, the Senior Legal Counsel at KfW was involved in the Rhineland programme from 2007. He has been responsible for the legal aspects stemming from KfW’s role as main creditor to the Rhineland Programme. His evidence was directed towards the roles of IKB and KfW and the steps taken post 2007. As such he gave evidence relevant to the limitation issues. Mr Bulgrin was a very serious and careful witness, who took time to consider questions before giving answers and making scrupulous distinctions (perhaps based in the caution coming from one giving evidence in his second language) where he saw them as relevant, which on occasion made his answers somewhat lengthy. It was clear that he took his duty towards KfW and the German government very seriously. Again his was a statement not prepared in accordance with PD57AC, with the consequent issues about what was recollection and what was “refreshment”.

Mrs Jean Lawrence was a Trust Director at Bedell Trust and an alternate director of L30 between November 2010 and 4 April 2012. Like Mr Anderson she emphasised the fact that Mr Hollywood was the day to day person dealing with L30. Her evidence addressed limitation issues. She was cross-examined by Mr Sher. As a witness she was quietly spoken and clear about the limits of her recollection and also about the fact that she may not have read everything that she was sent, in circumstances where a somewhat maximalist approach to the copying in of personnel seems to have been taken.

Mr Alasdair Hunter, an English qualified solicitor who was a partner at Bedell Cristin and director of L30 from November 2010 until August 2016. Like other witnesses he emphasised that the lead person at Bedell Cristin was Mr Hollywood, and said that his role was limited to attending Board Meetings. His evidence was primarily directed to the question of knowledge for limitation purposes. He was cross-examined by Mr Sher.

Mr Paul Anderson was also a director of L30. He spoke to reliance in particular as regards rating, as well as giving some general evidence about L30's processes for approvals and board meetings. He was frank about his lack of recollection of specific events and his evidence tended to suggest that his memory for the routines was also less than perfect.

Mr Andrew Nowotny, Legal Counsel at IKB is an US qualified attorney who manages IKB's litigation in the US and UK including in relation to RMBS. As such he gave evidence as to knowledge for the purposes of limitation as well as the relationship between IKB, the Loreley companies and KfW – in particular as to the lack of sharing of information.

Mr Alan Dart, formerly an English qualified barrister, but later a partner at Bedell Cristin was a director of L30 from January 2007 to February 2012. He gave a statement as to how L30 was run – consistent with that of the other witnesses in identifying Mr Hollywood as the prime mover. Otherwise his evidence was directed to the question of knowledge. His evidence essentially replicated other evidence being given and so he was not called. Similarly a witness statement was provided from Mr Butel.

Mr O’Driscoll worked on the CDO side of the business from 2003 to 2009 as Co-Head of CDO Structuring, Americas. He put together this transaction. He reported to Mr Kareem Serageldin (of CSI, based primarily in London), who was head of the CDO Group but had no direct involvement in the transaction, although he was kept periodically informed of its progress. Mr O’Driscoll and Mr Serageldin are both accused of dishonesty, with L30 claiming that they knew the alleged CDO Representations were false due to the RMBS Misconduct. Mr O’Driscoll gave evidence as to the structure of the CDO Group, what it did and its strategy in the relevant period, as well as his specific interactions with the RMBS Group in the relevant period, structuring of CDOs generally and the transaction in this case. Mr O’Driscoll was a firm, intelligent and fluent witness; while his evidence was given with a dash of warmth and humour he did not hesitate to take the fight back to Mr Lord or to criticise some of CS’s own practices. He was an impressive witness, whose evidence I accept. L30's case that he was dishonest was realistically not actively pursued in closing.

Mr Kaiserman was one of the individuals against whom L30 alleged fraud (RMBS Loan Compliance and QC representations). He was until recently a Managing Director in Non-Agency RMBS Trading at CS. It appears that, shortly after signing his witness statement, Mr Kaiserman left CS for Apollo Global, as part of Apollo’s acquisition of CS’ securitised products business. Mr Kaiserman was at the relevant time the Co-Head (with Mr Peter Sack) of the “Transaction Management Group” (“TMG”), which was a largely administrative group within the RMBS business, which co-ordinated different parts of the business. Mr Kaiserman was more focussed on the acquisition side of the RMBS business than any other relevant aspect; in particular he had little knowledge of CDO business generally. He gave evidence as to the business of the RMBS Group and its processes, as well as his dealings with investors and rating agencies. Mr Kaiserman was another impressive witness, demonstrating keen concentration and great clarity in his evidence. His evidence was for the most part given with a quiet and calm confidence; the odd occasion when his delivery was more hesitant was noticeable for being a variation from this.

Mr Sacco was the third live witness against whom fraud was alleged and I have been asked to find that he was an untruthful witness. He was Head of the Credit Policy and Underwriting Group within the RMBS Group between 2003 and 2008. This included managing due diligence in respect of loans from the Bulk Channel, compliance (although compliance stopped reporting to Mr Sacco in 2006-7), and addressing post-closing quality control. He gave evidence as to bulk DD, quality control and interactions with rating agencies as well as the process of presenting risks via the presentations.

Mr Sacco was initially a more hesitant and defensive witness though he gained confidence as the process continued. His recollection was less fluent than that of the earlier witnesses evincing a lack of familiarity with a number of the underlying documents and on occasion with details in his statement. However his recollection of systems and what his team actually did prompted him to greater fluency and while he engaged good-temperedly with a very long (2.5 days) cross-examination, he resisted answering questions which he felt were put unfairly. L30 said it was incredible that he would not have read the SoF (for which he was deposed); I cannot agree with that. It was entirely credible that having been deposed, and knowing the outcome he would rather not wallow in the detail. I was entirely unpersuaded that he was dishonest.

Mr Sacco’s evidence gave rise to one of the most contentious passages of the trial. There was an issue raised as to whether Mr Sacco had spoken to his own lawyer during the course of his evidence, which ran over three days. A member of the Claimant’s team was sure that they had heard Mr Sacco’s lawyer telling him that a particular answer had been wrong. That would obviously have been in breach of the purdah warning given by me before the break and potentially a contempt of court.

The matter was dealt with, following discussion with counsel, by my asking a series of questions, the substance of which had been agreed. Mr Sacco denied making any such statement and stated that his conversation with his lawyer had been limited to the kinds of domestic/logistical matters as are entirely proper. His lawyer, Mr Zelenko, also (separately) denied any discussion of evidence both in oral questioning and in a witness statement which I asked him to lodge. That evidence was not accepted by L30, who submitted that Mr Zelenko’s account accepts that he did have a discussion with Mr Sacco outside Court and that I should treat the denial of any relevant or improper conversation with scepticism.

I have carefully considered the submissions made and the submission that Mr Zelenko’s evidence that he didn’t even understand what the phrase said to have been overheard meant was not credible. I do however find it credible as the answer was a matter of considerable detail whose relevance was by no means apparent. On the balance of probabilities (into which the implausibility of such a breach committed (i) by two people one of them a lawyer and former Federal prosecutor, warned not just by me but also by the CS team (ii) actually within the court building and effectively under the nose of the Claimant’s team, when there would be ample opportunities to talk privately outside of the building) I have concluded that there was no impropriety or breach of the purdah rule and this concern probably arose out of a mishearing informed by complete focus on and immersion in the case. I do not for one instant consider that the member of the L30 team who identified the concern was doing anything other than reflecting what they honestly believed they saw and heard. I do however conclude that they were mistaken.

What this episode did however throw into clear perspective is that Mr Sacco – and many of the other CS witnesses – has considered the events of this period repeatedly both via depositions and in discussion with his lawyer. He too, therefore has been “refreshed” on much of the material in play. I do therefore have to treat the evidence of Mr Sacco and the other deposed CS witnesses as being likely to have been affected by the process of recollection and discussion.

Mr Nordyk was another of the individuals against whom fraud was alleged by L30 (in respect of all but one of RMBS Representations). He was a Vice President at CS (he now works elsewhere). Mr Nordyk was in the Credit Policy & Underwriting Group. He was responsible for due diligence in the Bulk Channel, and his evidence is focussed on that issue. He was also in charge of (and gave evidence on) quality control for some time during the Relevant Period. He reported to Mr Sacco, of whom he spoke very highly as a manager and a teacher in the course of his evidence. Mr Nordyk was a firm and fluent witness. He was however also disciplined in his approach to evidence – perhaps as a result of the numerous depositions he had undergone over the years. Thus he was unwilling to answer questions he regarded as unclear or ambiguous, meaning that on occasion he appeared a little defensive. When within his comfort zone he was extremely earnest and fluent, speaking quickly and decisively and being prepared to offer considerable detail. His earnestness and fluency was compelling; subject to the point made about the effect of the deposition process.

Mr Sack was the final of the witnesses against whom L30 alleges fraud (RMBS Loan Compliance and QC Representations) and another witness whose evidence L30 urged me to reject as dishonest. He also remained at CS until very recently. He was (with Mr Kaiserman) Co-Head of TMG. Mr Sack was more focused on the issuing and underwriting of securitisations. He dealt in particular with the Quality Control/due diligence process and the Mortgage Loan Purchase Agreement (“MLPA”) process, as well as the culture of the RMBS Group.

It is fair to say that his evidence was not as clear and focused or as enjoyable to listen to as some. But I cannot accept the submission that he had no regard for the importance of telling the truth. He was a careful and considered witness – verging on the overcareful in terms of giving detail when he was not clear where a question was going or pointing out ambiguities in questions. He was thinking ahead about where questioning was going and gave some answers which suggested a familiarity with the deposition and/or trial process to which he alluded in his statement; and very possibly a desire to needle Mr Lord. His recollection was (understandably) imperfect as he frankly acknowledged both in writing and orally. My strong impression of him from his live evidence was as a witness of candour and I therefore substantially accept his evidence, subject to the rider about the effect of previous encounters with the legal process on these matters.

The starting point is that it is common ground that L30’s claims in deceit and negligence are prima facie time-barred. L30 says however that it is not barred from suing because of the effect of sections 14A (negligence) and 32 (fraud) of the Limitation Act 1980. It is common ground that the cut off date for these purposes is 15 November 2012 for the claim in fraud and 15 November 2015 for the negligence claim.

It is section 32 which is principally in focus because it is this provision upon which L30 relies primarily (and solely with respect to the deceit claim). In addition there is a close connection between section 32 and section 14A, so the answer to one question will tend to drive the answer to the other. That connection was noted in FII Group Test Claimants v HMRC [2012] UKSC 19 [2012] 2 AC 337 (“FII 1”) where at [63] Lord Walker emphasised the interrelationship of the two when noting that an expansion of section 32 would require the limits and indeed the rationale of ss. 11 and 14A of the Limitation Act 1980 to be revisited. That analysis was endorsed in FII Group Test Claimants v HRMC [2022] AC 1 (“FII”) at [194].

Section 32 provides materially as follows:

“Postponement of limitation period in case of fraud, concealment or mistake

(1)

…where in the case of any action for which a period of limitation is prescribed by this Act, either -

(a)

the action is based upon the fraud of the defendant; or

(b)

any fact relevant to the [claimant’s] right of action has been deliberately concealed from him by the defendant; or

(c)

the action is for relief from the consequences of a mistake;

the period of limitation shall not begin to run until the [claimant] has discovered the fraud, concealment or mistake (as the case may be) or could with reasonable diligence have discovered it.”

There was little dispute as to the principles which come out of the many authorities on this section. They may be summarised as follows:

Discovery can in some cases create a debate. Two different “tests” or points in time have been discussed in the authorities, namely the point at which the claimant had (or would acting with reasonable diligence have had): (a) sufficient knowledge to plead its claim (“the Statement of Claim Test”); and (b) sufficient knowledge “to justify embarking on the preliminaries to the issue of a writ” or “in the sense of recognising that a worthwhile claim arises” (“the Worthwhile Claim Test”).

The cases in which there is a difference between the two will be rare: In Gemalto Holdings BV v Infineon Technologies [2022] EWCA Civ 782 Sir Geoffrey Vos MR observed at [45]: “there is unlikely in most cases, as in this case, to be a real difference between the application of the statement of claim test and the [worthwhile claim] test. Indeed, the statement of claim test is, perhaps, little more than a gloss on the [worthwhile claim] test”.

The Statement of Claim Test which the parties have agreed that I should apply in this case requires that the claimant is in a position (i.e. has sufficient actual or constructive knowledge) to plead a complete cause of action, which in fraud cases entails the critical allegations that a representation has been made, that it was false and that the representor knew it to be false: see Barnstaple Boat Co. v Jones [2007] EWCA Civ 727 at [34]; Bilta (UK) Ltd v SVS Securities plc [2022] BCC 833 at [31(7)(h)].

A final aspect of the legal backdrop is the question of which version of the case must be known in order to trigger the running of time. In this context the question is whether it is the version originally put forward, or the version finally pursued (to the extent that they are different). On this the submission was made by L30 that it is the fraud as found by the Court, not (if different) as originally pleaded by the claimant which must be reasonably discoverable. This argument was advanced by reference to the recent Court of Appeal decision in Seedo v Gamal and Salfiti and others [2023] EWCA Civ 330 in particular [53] per Nugee LJ:

“I accept [the] submission that there is nothing in the authorities which requires the Court … to confine its attention to the case as pleaded by the claimant. Normally of course one would expect the claim as found to reflect the claimant’s pleaded allegations, but one can envisage a case where the pleading contains allegations that are not pursued or turn out not to be well-founded, but where the judge nevertheless finds there to have been a fraud. It seems wrong in principle, and a distraction, to ask when the claimant discovered allegations which in the end went nowhere: the question is when he discovered the essential facts of the fraud found proved by the Court.”

While this point was clearly stated by Nugee LJ, it is not ratio and I conclude that I am not bound by it in circumstances where;

The point which was really in issue in Seedo was the following point, isolated at [4] of the judgment: “Where the defendant has deceived the claimant into entering into a transaction by telling more than one lie, what happens if the claimant discovers that one lie is untrue but does not then bring a claim, and then some years later discovers that a second lie is untrue?”

That was the true nature of the problem in Seedo; it was a case where the claimant had been deceived by two lies told by the defendant Mr Salfiti. The question was whether he was time barred six years after he could have discovered the first lie – or, in respect of a claim brought in relation to the second lie, only six years after he could have discovered that lie. The first lie was the one pleaded: the fraud in respect of the alleged loan and Mr Seedo’s financial contribution, the Trust Deed being alleged to be invalid and so of no effect. The second lie, that which the judge found, was that Mr Salfiti led Mr Seedo to believe that he would be 100% owner of the property. It can be seen that these are essentially two different points.

I conclude that the question of which version of the case needs to have been known is not as cut and dried as L30 suggests. Where there are two distinct lies, as in Seedo, a party will not be timebarred if one (bad) claim is timebarred, if an entirely different claim, ultimately successful, is only discoverable later. However the position may be different where one is looking at two versions of the same argument, or arguments which have some relation to each other. The argument is likely to be fact sensitive – as was the decision in Seedo.

L30, relying on Seedo, suggested that it would be wrong to take as the starting point the question of the point at which L30 could with reasonable diligence have pleaded the “precise deceit” set out in the 2018 letter from RPC enclosing the Claim Form. It suggested that the claim as ultimately argued is the correct starting point.

As indicated above I take the view that Seedo does not justify this conclusion, and that it is necessary to analyse whether the case as run is essentially the same or essentially distinct from the case set out in the letter before action.

The Letter Before Action sets out three implied CDO representations relating to the characteristics and quality of the underlying business – specifically the RMBS – in summary:

The falsity alleged in the Letter Before Action was skeletal: “The representations above are demonstrably false in the light of the admissions contained in the Statement of Facts and they were made dishonestly (alternatively negligently) thereby entitling L30 to rescission and/or damages”.

The alleged falsity was expanded in the draft Particulars of Claim (“DPoC”) produced on behalf of L30 (on KfW’s instructions) in May 2019 to be, in essence, as follows:

No particulars of falsity or knowledge were set out in RPC’s letter of 21 December 2018, but some guidance as to L30’s analysis at the time of issue of these proceedings may be gleaned from the DPoC. In particular, at [41] of the DPoC L30 set out the facts derived from the SoF which it claimed (at [46]) falsified the representations. These included:

It was also said that the representations were made fraudulently. On this it was said in the DPoC that the Relevant Persons (Messrs Kimura, Marriott, Serageldin and O’Driscoll) knew that the representations were false or were reckless where:

Then there is the question of the case as finally run. The CDO representations as pleaded and run were in essence:

The falsity alleged is similar but not identical to that alleged in the DPoC. It hinged on:

The fraud case was by trial directed at a wide variety of people.

It follows that the case advanced at trial was in some respects slightly broader and in some respects slightly narrower or more detailed than that which was outlined in the DPoC. But it was essentially similar to that earlier iteration. It was not a case of the DPoC case being abandoned and a different case pursued. In those circumstances I do not consider that the analysis in Seedo is applicable. However given the resemblance between the two cases run, nor would the Seedo analysis be likely to make a material difference here.

The question is therefore what L30 could have discovered and whether it could have used that material to plead (or to embark upon the preliminaries to pleading) this claim against CS.

The central issue for limitation purposes is as to L30’s awareness of the falsity of these representations. There can be no issue, given the nature of the representations and the case L30 runs that the representations were discernible (and if we get so far) perceived/understood. The only thing which holds up the ability to make the claim must therefore be the knowledge of falsity.

The issue then becomes whether there is any practical difference involved in the argument as to whether L30 had to be aware of the entirety of the case that it made either at the outset or later. L30 says both that the case as advanced at trial was different and also that there were additional grounds of falsity relied on.

I conclude that there is in fact no difference. This is because, as summarised in Appendix 2B of this judgment (reflecting L30’s careful summary in written closing), all the CDO Representations are said to be falsified by all and any aspects of the RMBS Misconduct. Thus any part of the RMBS Misconduct, if established, is said to justify a conclusion of falsity (see for example in closing “Representations C13-C14 are falsified by any strand of the RMBS Misconduct” or, as Mr Lord KC put it on Day 1 of the trial: “A little bit of fraud goes a long way. So we don't need very much fraud in and about the underlying representations, we say, to establish the falsity at the CDO level.”).

It follows that:

Before moving on a point should be noted about knowledge. Reliance was principally placed by CS upon the actual knowledge of KfW, the Loreley SPVs and their (joint) attorneys (and with such assistance from IKB as was required); this was said to be demonstrated by the actual complaints which were brought which were claims in fraud against numerous banks, specifically: Calyon, Morgan Stanley, Countrywide, Wells Fargo, Merrill Lynch, Citi or Deutsche Bank (“DB”). L30’s original case was that only its own directors’ knowledge could be counted against it. It appeared that the question would be hotly contested and that it would be necessary to analyse just what L30's directors knew, as opposed to KfW/IKB.

In closing that position shifted: what was said was:

“L30 is content to proceed on the assumption that L30 either knew or could with reasonable diligence have known of (i) public material in all proceedings in the United States which CS relies upon and (ii) the contents of the FCIC and Levin reports (and their exhibited materials). This renders redundant (and will avoid the Court needing to embark on) any factual enquiry into whether any such claim/report in the public domain actually came to the attention of a particular Loreley director...

L30 is content to proceed on the basis that, at all material times, KfW and L30 were reasonably attentive as to whether the Loreley companies may have potential claims against the banks which sold CDOs in respect of the Rhineland portfolio, including the Magnolia Notes. Therefore, if there had been material in the public domain from the US claims/reports which would have enabled L30 to bring its fraud claim, it can be assumed that such material would have come to the attention of L30 for the purpose of considering a potential claim against CS.”

The net result is that the question of attribution of knowledge largely disappeared. The issue as to knowledge remained only in relation to such matters as contract specific investigations performed by IKB in relation to other contracts.

The question of trigger was therefore less in focus in closing than it was in opening. L30 says that because it has on reflection decided that it is prepared to proceed on the basis of certain assumptions, it is not necessary to decide the trigger issue. However CS pointed out in closing that the effect of the assumptions which L30 is prepared to make is to focus attention on only one part of the evidence and take attention away from the actual subjective knowledge of KfW. That is a valid point.

Further it can often be the case that assumptions or concessions can lead to tensions within the other parts of the argument. It is therefore preferable to consider the trigger issue on its merits, via the emergence of knowledge and the making of other complaints.

In looking at what happened in those complaints it is necessary to bear in mind the division highlighted in argument, which is one between what were described in argument as “special knowledge” claims (a phrase which denoted that the banks which had arranged the CDOs had “special knowledge” about the assets selected for the CDO and which involved allegations of “non-disclosure of information to investors because the banks wanted to remove poor quality assets from their books”) and those described as “Magnetar-type claims” – a description which covers an adverse selection type claim which will be explained further via the timeline below. The claims in this case are special knowledge claims: L30 sought to argue that the knowledge it could have had related only to Magnetar type claims.

The story proceeds as follows. The Financial Crisis of course impacted in 2008. It did not however come from a cloudless sky. As later inquiries would report, concerns about a residential mortgage bubble had been the subject of widespread discussion since at least 2005. The link to the securitisation and the repackaging of securities into CDOs stayed for some time under the radar but began to emerge in 2007.

By July 2007 Federal Reserve Chairman Ben Bernanke was forecasting that subprime losses could hit US$100 billion. By the end of July a deal to rescue IKB via a buyout by Lonestar was announced with stories suggesting that it had already bought more than US$30 billion of CDOs from Merrill Lynch at 22 cents on the dollar.

In 2008, KfW put in place a press monitoring and/or alert system “to capture information on developments relating to the financial crisis”. It also instructed lawyers to provide advice on potential legal avenues in both jurisdictions including what potential claims may be available, and later asked both U.K. and U.S. lawyers to provide “updates on litigation relating to the financial crisis” and advice on potential claims that might be available “in connection with [the Loreley SPVs’] assets”.

In parallel from mid 2009 the Financial Crisis Inquiry Commission was taking evidence.

One resource available to those reviewers were the “All Clayton Trending Reports” (“the Clayton Report”). This report was produced by Clayton Risk Insight Solutions, a provider of due diligence, underwriting and risk mitigation solutions for the residential and commercial real estate markets. The Financial Crisis Inquiry Commission Report (“FCIC Report”) noted that it was “a major provider of third-party due diligence services” and that “Because of the volume of loans examined by Clayton during the housing boom, the firm had a unique inside view of the underwriting standards that originators were actually applying—and that securitizers were willing to accept”.

Its reports were publicly available on its website. The Clayton Report for the period from Q1 2006 to Q2 2007 covered nearly 1 million loans – and set out data showing the rates at which loans reviewed by Clayton were rejected (i.e. marked EV3) and how many of those loans were nevertheless “waived” in by the client financial institution, all broken down by financial institution. It showed for the 18 months to June 2007 39% of loans Clayton found deficient were “waived in” and recorded Clayton’s view that clients waived in loans to preserve relationships with a business originator. The summary – set out clearly in the FCIC Report - stated for CS that 32% of loans were rejected by Clayton (graded EV3) yet CS waived in 33% of those rejected loans.

The Clayton Report provided a basis for any careful reader to say that:

The publicly available testimony of Clayton’s employees given to the FCIC gave more – for example the testimony of Ms Vicki Beal that banks would often use the information obtained from Clayton to negotiate a discount rather than to reject or put back the loans. This is a point which was not specifically reported in the final report, but was made in oral evidence; which evidence was picked up in some of the actions subsequently brought by Loreley companies (against Citi, DB, UBS and Merrill Lynch/Morgan Stanley). All of these show that the transcripts were readily available and read by KfW’s lawyers. But they also show that the testimony was reported and interesting issues “flagged” at the time. So for example the Citi complaint cross refers to an article at the time Ms Beal’s evidence was given: Jonathan R. Laing, Banks Face Another Mortgage Crisis, Barron’s, Nov. 20, 2010.

I note here for completeness that in 2009 MBIA sued CS (MBIA Insurance Corporation v Credit Suisse Securities (USA), LLC (“the MBIA Proceedings”)). While the contents of this action became relevant at a later stage, I do not consider that the profile of the original claim was such that it should be counted amongst the publicly available material which could constitute or contribute to a trigger.

In 2010 Magnetar emerged. In mid-2010, there was publicity surrounding a hedge fund called Magnetar (reputedly named for the super-magnetic field created by the last moments of a dying star) which had taken a short position in CDO transactions which it had helped create. The particular feature of the allegations relating to Magnetar was that it had selected deliberately risky assets in the Reference Portfolio such that the CDO was “designed to fail”. It was therefore an adverse selection claim. But at the same time, a part of what was said in that claim was that those involved had known of the lack of quality involved and had nonetheless sold the CDOs.

A similar adverse selection allegation was made in relation to Goldman Sachs’ involvement with the Abacus CDOs, a series of 24 of which were sold from 2004. In the 2004 CDO Goldman was the short investor for the entire US$2 billion deal: it purchased credit default swap protection on these reference securities from the CDO. The funded investors, including IKB, put up a total of US$195 million to purchase mezzanine tranches of the deal. These investors would receive scheduled principal and interest payments if the referenced assets performed. If the referenced assets did not perform, Goldman, as the short investor, would receive the US$195 million. Thus in this deal IKB and other investors were “long” investors, betting that the referenced assets would perform well, and Goldman was a “short” investor, betting that they would fail. But in April 2010, the SEC charged Goldman Sachs with fraud for telling investors that an independent CDO manager, ACA Management, had picked the underlying assets in a 2007 Abacus CDO when in fact a short investor, the Paulson & Co. hedge fund, had played a “significant role” in the selection. The SEC alleged that those misrepresentations were in Goldman’s marketing materials for one of Goldman’s Abacus deals. In July 2010, Goldman Sachs settled the case, paying a record US$550 million fine. Goldman acknowledged that the marketing materials for the transaction contained incomplete information.

These developments prompted KfW to identify the 161 CDOs owned by the Loreley companies and to investigate whether they might give rise to claims. Primarily in the picture were “Magnetar-type” claims, i.e. claims arising out of a “net short deal” or adverse portfolio selection, but (in line with the initial instructions given to IKB’s legal advisers in 2008) the consideration extended to any issues which were thought to be worthy of investigation.

This review (the “IKB Review”) was based on public information, together with IKB’s own information and documents stemming from its due diligence in connection with the investment recommendations. The IKB Review was completed by the end of 2010. The review exercise included a review of the Notes and other CS CDOs. It was conducted by three firms engaged by IKB. The agreement by which this review was conducted and the outputs have not been before the court, on the basis that privilege was claimed over them, and therefore what was said is not known.

However, in around August 2010, KfW instructed US counsel, Meister Seelig & Fein (“MSF”) and Kasowitz Benson Torres LLP (“Kasowitz”), to advise it in relation to “potential Rhineland Programme asset recoveries.” Kasowitz also set up a monitoring system to alert KfW to developments regarding RMBS and CDO litigation, particularly in New York. Again, the ambit was “anything that would be relevant” to the 161 CDOs, in particular to identify possible claims and material to support them. Mr Bulgrin was sent and read documents as part of the monitoring exercise.

At more or less the same time the FCIC Report emerged. That was published in January 2011. It concluded (amongst other damning conclusions) that “collapsing mortgage-lending standards and the mortgage securitization pipeline lit and spread the flame of contagion and crisis.” It is notable that some of the evidence submitted to the FCIC was freely reported earlier. So, as one of the L30 US complaints notes:

“the New York Times noted in a September 26, 2010, article, ―[b]ecause these loan samples were provided to the Wall Street investment banks that commissioned them, they could see throughout 2006 and into 2007 that the mortgages they were financing and selling to investors were becoming increasingly sketchy.”

A few months later, in October 2011, came the Levin report (the Report of the Permanent Subcommittee on Investigations of United States Senate entitled “Wall Street and the Financial Crisis: Anatomy of a Financial Collapse”). That similarly offered the damning conclusion: “a host of financial institutions that originated, sold, and securitized billions of dollars in high risk, poor quality home loans that inundated U.S. financial markets. Many of the resulting securities ultimately plummeted in value, leaving banks and investors with huge losses that helped send the economy into a downward spiral. These lenders were not the victims of the financial crisis; the high risk loans they issued were the fuel that ignited the financial crisis…. Two case studies, … illustrate a variety of troubling practices that raise conflicts of interest and other concerns involving RMBS, CDO, CDS, and ABX related financial instruments that contributed to the financial crisis.”

The IKB Review exercise was followed by careful consideration of the outputs with KfW’s lawyers (MSF and Kasowitz). Where the view was then formed that there may be a viable claim in respect of a CDO, that then led to discussions between KfW and the Loreley company as to progressing that potential claim. Potential claims were identified against 23 potential defendants; across 95 CDOs, including CS and the Notes. A Common Interest Agreement was executed in September/October 2011. That Agreement stated:

“Whereas the Parties anticipate litigation… against the Arranging Banks, Managers, Magnetar and others … in connection with the CDOs… (the “CDO Litigation Project”)…

Whereas this agreement confirms that the parties and the undersigned counsel have already been in communication regarding the Anticipated Actions and the CDO-Litigation Project …”

One of the CDOs identified in the Common Interest Agreement was the Notes.

While the centre of gravity of the investigations was with KfW and the lawyers, L30’s directors did get updates about what was happening. There were regular substantial (about an hour long) telephone calls every week or so. These included explanations of the progress of the investigations into potential claims. There were also email updates regarding new litigation filings, press reports or public reports. Those actual documents were again not before the court, having been withheld on the grounds of litigation privilege. However summary documents show (to give but one example) that in September 2011 Mr Hollywood of L30 (as well as the main players at KfW) was sent an email entitled “SEC Widens CDO Probe”. In June 2011 he was also sent a copy of the MBIA decision. Mr Hollywood accepted that he would have read this material and that he expected the lawyers to familiarise themselves with the docket in the MBIA Proceedings.

Various claims were then brought in the United States by certain of the Loreley companies in 2011, against e.g. DB, Merrill Lynch, Wells Fargo, JP Morgan and UBS.

This is an appropriate point to ask whether there was a “trigger” which would reasonably prompt L30 to investigate. I conclude without any real hesitation that there was.

As for the contention by L30 that nothing thus far indicated wrongdoing, still less fraud on the part of CS and that all that had emerged was generalised concerns regarding the practices of banks in general which were insufficient to amount to a trigger, I reject that submission.

Firstly, the focus here is not on precision but on a trigger – being on notice that there is something to investigate: a something which would have caused a reasonably attentive and diligent person to ask questions (cf. OT Computers [47]). What all of this did was not to highlight a case against CS alone nor indeed to point a finger at specific other banks. Rather it indicated the potential for claims in multiple directions. Even absent specific material relating to CS I would have been minded to conclude that the fact that numerous banks were mentioned would cause a reasonably attentive person who had bought similar (and similarly unsuccessful) products from other banks to ask whether there might not be something wrong with its own purchase.

This conclusion is supported by the fact that IKB in fact did ask precisely such questions, as the narrative above reveals.

But in fact the indications went further than this. L30’s case in opening that a conclusion had been reached that there was not a viable claim against CS, or that CS was not at that time seen as a potential defendant, is contradicted by the evidence: by September 2011 the CDOs in respect of which Mr Bulgrin anticipated litigation were the 95 in Exhibit A to the Common Interest Agreement, including the Notes; “At least that point in time, it may have been sooner” as Mr Bulgrin said. This means that it is established on the documents, supported by a relevant witness, that at this point CS was seen as a potential defendant in relation to this transaction.

Accordingly I conclude that there was a trigger by the date of execution of the Common Interest Agreement in September/October 2011. I also note that in the light of the evidence which has emerged L30’s disputing of this point all the way up until the service of closing submissions is somewhat surprising. That evidence included not just Mr Bulgrin’s recollection noted above, but also Mr Hollywood’s frank admission that:

Even if that were not so, I conclude that this trigger would have been reached by late 2011, when KfW discovered that IKB had sued CS for fraud in respect of certain RMBS which IKB had purchased. That litigation was the result of the monitoring processes for publicly available documents it had implemented.

The main point at issue on the question of whether L30 could with reasonable diligence have known of the central RMBS allegations – including the dishonesty aspects – before 15 November 2012 concerns what one can properly take from the claims which were made.

Here I shall consider three possible sources, on top of the fully public materials to which I have already referred;

The relevant claims are:

The Rhineland claims were all approved by KfW (Mr Bulgrin and the steering committee), the US law firms retained by the Loreley companies, Mr Hollywood and the relevant Loreley companies’ directors. They therefore reflect L30’s knowledge.

L30 says that the claims it made were primarily “Magnetar-type” claims and “the fact that one has a claim against Peter for some wrongdoing does not mean that one has a claim against Paul for some other wrongdoing. …. Moreover, the ‘Magnetar-type’ claims were wholly different in kind from the RMBS Misconduct alleged against CS in these proceedings.”

There has therefore been a lot of concentration in submissions and the evidence of what was alleged in those complaints and what knowledge can be taken from them. Before entering into this it is worth reiterating what the essence of the “worthwhile case” or “statement of claim” case here is: knowledge that the underlying business did not match up with the offer documentation – and most specifically knowledge as to falsity of ratings.

The essence of that is the “riskier than you told us” element. This broad allegation was central to essentially all of the complaints and can be seen to have been the malaise underpinning much of the publicity. So on this one sees for example:

Then if one looks to the RMBS elements and knowledge of that, there is also considerable specific knowledge alleged.

Eerily similar language showing a familiarity with Clayton and the underwriting guidelines issues appeared in other complaints;

The main thrust of the argument put forward by L30 was two fold. The first is that these complaints were Magnetar type complaints. The second is that even if the relevant knowledge was capable of being gathered, L30 was not capable of doing that.

However, I did not find either of these points persuasive. The relevant kinds of “special knowledge” were deployed in a number of the other claims. For example:

One thing which is worth noting at this stage is that the material used by all these claimants in the Rhineland claims was essentially publicly available material. The sources were:

All of this suggests that the kind of complaint which one sees in the other litigation could have been put together by L30.

The 2011 FHFA CS Action was a claim brought in the New York courts by the FHFA against CS arising out of CS’ sale of RMBS to Fannie Mae and Freddie Mac. This was one of 17 actions filed against different banks. The claim was issued on 2 September 2011. The claim concerned 43 RMBS, including five of the CS RMBS. In broad summary, FHFA contended that CS had made misrepresentations e.g. as to whether the mortgage loans complied with underwriting guidelines. This was not however a fraud claim – a point emphasised by L30. The causes of action were breach of US securities laws and negligent misrepresentation.

The complaint did however have some notable overlap with the case run here. In particular:

The FHFA complaint also provided a reworking of allegations in proceedings brought by MBIA against CS in 2009 (“the MBIA CS Proceedings”). This was apparently done via a review of documents obtained by FHFA from the publicly available court file in that action. Such documents were equally obtainable by KfW/L30 and its lawyers. FHFA relied upon MBIA’s review of original loan files (to which it had access, though they were not publicly available) and also MBIA’s reliance on CS’s QC reports. The material pleaded based on the MBIA allegations included:

The FHFA then in addition conducted its own “review of loan-level data”, using thousands of randomly selected loans, which FHFA claimed supported claims of material representations based on owner-occupancy and LTV ratios. As for the LTV ratios, these allegations were based on the use of AVMs (which required the list of addresses of the underlying properties). FHFA also appears to have obtained nearly 2,000 loan files which it then used to conduct an alleged “forensic review” of compliance with underwriting guidelines. While the work was, as noted above, not based on publicly available material, it was used by the Loreley entities as part of some of the Rhineland claims.

It is then sensible to look at the complaint which IKB did bring against CS in the US. IKB commenced an action in the New York courts on 10 November 2011 by way of a short Summons. A detailed Complaint was then filed on 2 July 2012. The claim related to 14 RMBS which IKB had acquired outside the Rhineland Programme which included two of the CS RMBS. The gist of IKB’s claim is that CS had made misrepresentations regarding the loan quality (e.g. LTV ratios, occupancy status) and the loans’ compliance with underwriting guidelines. IKB alleged that CS was aware that the representations were false. IKB’s claim in this action did therefore involve a plea of fraud.

On 31 August 2012, CS issued a motion to dismiss the claim. In its Memorandum of Law in support of its motion to dismiss, CS set out its position that the claims failed in their entirety and fell to be dismissed on a summary basis for six reasons. The second of these reasons was that IKB “fails adequately to plead fraud because it fails to plead an actionable misstatement or omission of material fact, scienter, reliance or loss causation”. CS explained that all of IKB’s allegations failed because the offering documents contained no actionable representations. CS explained that IKB had failed adequately to plead “scienter”, that being the required element of knowledge of falsity to support a fraud claim and that IKB’s allegations had failed to identify any proper basis for a plea of scienter.

It is between this point and the ruling in this action that the key limitation date of 15 November 2012 passes.

CS’ motion to dismiss was ruled upon on 3 March 2014. The Court rejected CS’ motion to dismiss. CS filed its Answer to the Complaint on 15 May 2014, which consisted of comprehensive denials of any wrongdoing. The case then moved to discovery. The proceedings concluded in 2022 prior to trial.

L30 contends that there was nothing in the 2011 IKB CS Action which enabled L30 to bring a fraud claim in respect of the RMBS Misconduct. Nor was there anything which ought to have put L30 on inquiry.

The first point made by L30 is that as at the time of the limitation “cut-off” (i.e. November 2012), CS’ position was that IKB’s claim fell to be struck out, including on the basis that the claim disclosed no proper basis for alleging fraud. L30 says that there is a fundamental inconsistency in CS saying today that, as at November 2012, the 2011 IKB CS Action was sufficient to enable L30 to plead fraud, in circumstances where CS was itself contending as at November 2012 that the claim disclosed no viable basis for a fraud claim.

This is not L30’s best point. The question of what CS was arguing and what L30 knew (to a standard of being able to allege it) is fundamentally different. There is no “fundamental inconsistency”. There was an active dispute about whether the facts pleaded by L30 passed the pleadability hurdle as a matter of US Law. That does not affect L30’s ability to rely on those facts. The same point arises in relation to the assessment of the Clayton Report material. L30 says that it could not have pleaded fraud because the FCIC Report recorded testimony which indicated that there could be innocent explanations for this. In order to be able to allege fraud a party does not need to know something to the point of certainty; it needs only to have a proper basis to allege it (cf. LIA [34]).

Second, L30 says that the fact that IKB considered that it could plead fraud in New York does not mean that L30 could have pleaded fraud in England. This position led into a vibrant debate, with CS maintaining in closing that L30 was debarred from denying the pleadability of fraud. I will revert to this point later.

As a primary route of analysis I regard it as preferable to consider whether what was alleged in the US contained the same building blocks as L30 did in the event use to advance its case here; in other words whether (regardless of pleading standards or ingredients for the local law or procedure) it knew and deployed substantive elements which it could have used to advance the same case in these courts.

It was L30’s case that there is very limited overlap between the allegations made in the 2011 IKB CS Action and the allegations of RMBS Misconduct made in these proceedings. This was addressed in Schedule 2 to the L30 opening skeleton the main thrust of the point being that:

This is the critical point at the heart of the limitation case. It is therefore necessary to analyse the allegations in the two sets of proceedings, and ascertain to what extent the IKB CS proceedings did give all or part of what was needed to found a claim in fraud in this jurisdiction.

The 2011 IKB CS Action focused on 11 CS securitizations, including two of the CS RMBS. In that case IKB’s case on fraudulent misrepresentations was placed under three headings;

The last of these does not overlap with the case advanced in this litigation. However the first two (plainly) do. As regards the quality issues, while there was focus on LTV ratios by reference to AVM valuations and occupancy status, these were heavily concerned with the underwriting standards and process. Thus IKB’s case theory was that “In the Offering Documents, the Defendants knowingly misrepresented the credit quality and characteristics of the pools of loans…” and that “The Offering Documents also contained material representations…regarding the purportedly rigorous loan underwriting standards employed in the loan origination process, the characteristics of the mortgage loans purportedly underlying the Securitizations…” in that “…there was a widespread and systematic departure from the originators’ underwriting guidelines, and the Defendants were fully aware of that fact at the time they made the false representations in Offering Documents…”

Pausing here, a central point in this litigation concerns ratings. This is an area where the US litigation was not explicit. In the 2011 IKB CS Action, while there was no headline allegation about ratings, the quality issues also did engage allegations regarding ratings. Thus at [169] of the 2011 IKB CS Action it was pleaded that: “In order to meet Plaintiffs’ investment criteria, Certificates had to have an investment rating of at least BBB– and at least 77% (or at certain times, more) of Plaintiffs’ investment portfolio was required to have a rating of A or higher. As explained above, the investment ratings of the Certificates were a direct product of Defendants’ misrepresentations concerning the loan pool.”. Again at [172] it was said: “All of Plaintiffs’ Certificates received investment grade ratings at issuance as a result of the Issuer Defendants’ misrepresentations concerning the credit risk of the underlying loan pools. The vast majority of the Certificates have since been downgraded to junk status.” Those pleadings, while not central to the US case, encapsulate allegations which are essentially similar to those run in this case.

There is also very considerable overlap regarding falsity. On falsity regarding substandard loans the complaint alleged:

On falsity regarding absence of due diligence on unsampled loans the Complaint pleaded:

“even though the reports from Clayton gave notice to Defendants that on average 32% of the sampled loans did not comply with underwriting guidelines or possess compensating factors, Defendants failed to conduct any additional review of the loans not yet sampled. In other words, even though they knew that the un-sampled set would contain approximately the same proportion of bad loans (as such is the purpose of sampling), Defendants ignored this obvious defect, and instead, placed all of the un-sampled loans into the Securitizations as well.”

On falsity regarding unreliable LTV ratios and inflated appraisals it was pleaded that:

On knowledge and fraud no specific allegations were made against the Defendants. The sole basis for alleging dishonesty came back to the Clayton material, as Mr Nowotny made clear in his evidence. However at [52] the pleading did effectively make allegations against individuals: “because of the substantial overlap in high-level personnel, the knowledge of each entity with an overlap can be attributed to the others with which it overlaps: a director, officer, or high-level employee does not forget what he knows about the operations of an entity when he changes hats to work for its affiliates.” It then set out individuals who cross deployed across various CS entities. Among these were Messrs Marriott and Kimura. The overlap here with the present case is striking. The draft POC pleaded knowledge against “at least” Mr Kimura, Mr Marriott, Mr Serageldin and Mr O’Driscoll. The final RRRRAPOC pleaded actual knowledge on the part of Mr Kimura and Mr Marriott and inferred knowledge on the part of the others, as well as knowledge on the part of the various others named in Schedule 7 to the RRRRAPOC.

Taking all of the above into consideration I reach the conclusion that, having had a trigger, there was ample material available from which L30 could have advanced the case pursued in 2018 at some point during the course of 2012.

It is not necessary in my judgment (based on the analysis of the authorities) for the precise case now run to have been available. Nor need the exact details have been available. What is necessary is that a case of the same nature (fraudulent misrepresentation) based on essentially the same representations could have been run. Here the alleged falsity of the CDO Representations is grounded in the alleged RMBS Misconduct and specifically those aspects I have highlighted above.

There are three aspects here. First was Clayton enough – L30's case being that this was all it had prior to the SoF, and that it was not enough. Secondly was there more available than Clayton? Thirdly, there is the question about the significance of the New York proceedings.

The headline (and easy) point here is that the allegation that CS used sampling and that Clayton marked many loans “EV3”, and yet CS still purchased them – were matters which could have been pleaded on the basis of the Clayton Report. It follows that on the face of it those parts of the L30 case could have been advanced simply on the basis of the Clayton materials. L30 says (citing the FCIC Report) that the EV3 point was not enough because EV3 was a designation which did not necessarily mean that a loan was defective. This is a somewhat ironic point, because this adopts CS's own case on the substance of the allegation. However the reality is that L30 did advance this point, both here and elsewhere on the basis of this material.

L30's answer here is via Seedo; it says that the case originally indicated, which did fully take this point was not pursued to trial. However that is to ignore (i) the fact that the RRRRAPOC still pleads that EV3 means “that the loans were not originated in compliance with applicable laws and regulations, the loans did not comply with applicable underwriting guidelines and lacked sufficient compensating factors, or the loan file was missing a key piece of documentation” and that the fact that 39% of loans in the due diligence pool were graded EV3 meant that the Loan Compliance Representations were falsified and (ii) the point was still deployed, if not separately pleaded: in closing L30 submitted that “The obvious inference is that most EV3-graded loans were in fact defective.”

The second point is that the Clayton Report addressed only loans which were acquired; not which were securitised or what if any further due diligence was undertaken post-acquisition. That is a perfectly valid point, but L30 has said in these proceedings that I should infer that loans were securitised: “the whole purpose of CS acquiring loans was to securitise them”.

The third point which L30 takes is that the Clayton Report provided no link between the sample of loans addressed and the 7 CS RMBS. Again logically this has force, but it need not (and did not) prevent the point being pleaded here and in the DB complaint in the US.

L30 then contends that falsity and dishonesty could not properly be alleged, because on “waiving in” the FCIC Report indicated the possibility of an innocent explanation. To this the answer is similar to that given earlier on regarding the US Proceedings: the fact that it is disputed does not make it unpleadable. And indeed L30 did allege fraud, albeit not specifically in connection with this aspect, and it did so while the same explanation as that given in the FCIC Report was still being tendered by CS. Therefore if the case was not pleadable for this reason, it was unpleadable throughout the entirety of the case. The same point can be made as regards inferring dishonesty from CS's knowledge of the extent of non-conforming loans; dishonesty was pleaded in the US on this basis and this has been pleaded here as a basis for dishonesty. It therefore ill becomes L30 to say that this could not have been pleaded when it was first known of.

Finally I should consider the submission that L30’s case did rest on more than the substance of Clayton – for example the emails cited in the SoF. However none of the materials relied on appear to be significant in terms of providing evidence or any meaningful further material in addition to what was provided by Clayton.

If there were to be enough doubt on the point via Clayton, the question arises as to whether there was other material – linking directly to the CS RMBS – which would carry L30 over the line. One answer to this would be to say (as CS does) that when L30 issued these proceedings it had no material to link the misconduct to the CS RMBS and therefore the point does not arise.

However in my judgment it is appropriate – and very helpful – to pan out and examine whether the point is really as binary as L30 suggests.

Here I conclude that L30 relentlessly undersells the other material available and in particular the combination of Clayton, the FCIC Report and the FCIC witness testimony. It argues that the public material was “The Clayton Reports as addressed in the FCIC report”; but that is wrong. In truth L30 (and any other concerned party) had access to:

All of this together, even before one adds in the other materials which can be seen being deployed in the Rhineland claims, makes a resource which demonstrates that “Is Clayton enough?” is actually a false point. Aside from the Clayton Report, the other publicly available material was not inconsequential.

I have also outlined above the detailed matters which were pleaded in the Rhineland claims. That demonstrates that knowledge that underlying loans in the CS RMBS failed to conform with offering documents underwriting criteria was plainly being alleged and pleaded in documents which were available to L30 and which it could very easily find.

For example, the allegation that the reported LTV ratios were unreliable in the case of many loans since the valuation appraisal was inflated, could have been pleaded on the basis of the same exercise which FHFA and IKB had performed (i.e. obtaining AVMs), or simply by adopting those parties’ claims pending disclosure based on publicly available material. While Mr Hobson submitted that L30 could not rely on other litigants’ work and would need to see the underlying documents, this is not consistent with what IKB/L30 did. As I have noted, the Rhineland claims show L30 adopting the analysis in the MBIA case.

The “Clayton or nothing” case is thoroughly wrong, therefore.

The point then taken by L30 is that the primary facts known were insufficient to support a plea of fraud in England and in particular that a case on individual dishonesty could not have been run – relying on (i) the fact that the 2011 IKB CS Action did not identify individuals as dishonest and (ii) the high requirements for pleading fraud in this jurisdiction.

As to the latter, more general point, advanced by reference to Three Rivers DC v Bank of England (No.3) [2003] 2 AC 1 at [186] and JSC Bank of Moscow v Kekhman [2015] EWHC 3073 (Comm) at [20], L30 neglects to take into account the authorities which establish that the pleadable case (in circumstances where ex hypothesi all may not have been open and above board) is somewhat lower than the standard at trial which is what L30's authorities address.

L30’s case cannot stand in the face of a consideration of these other authorities which focus on the requirements at the time that the plea of fraud is first deployed. Indeed at an earlier stage L30 itself argued that it need not name individuals, praying in aid the “generous approach to pleadings” in such situations referenced in Portland Stone Firms v Barclays Bank [2018] EWHC 2341 at [27]-[28] per Stuart-Smith J Nokia Corporation v AU Optronics Corporation [2012] EWHC 731 (Ch) per Sales J at [62]-[67] and Bord Na Mona Horticultural Ltd & Anr v British Polythene Industries Plc [2012] EWHC 3346 (Comm) per Flaux J at [29].

Indeed Kekhman itself does not really assist L30 because Flaux J there made clear that:

“At the interlocutory stage … the court is not concerned with whether the evidence at trial will or will not establish fraud but only with whether facts are pleaded which would justify the plea of fraud. If the plea is justified, then the case must go forward to trial and assessment of whether the evidence justifies the inference is a matter for the trial judge”.

Further even if these authorities did not exist, it would seem rather strange for a limitation defence to be defeated by reference to an assertion of pleading standards which was not observed when L30 did commence its action: no specific allegations were made against individuals in the original Particulars of Claim . It was only on the emergence of the RRRRAPOC in 2022 that underlying representations were pleaded and names named in relation to the RMBS Misconduct allegations. Although L30 says it was the SoF which enabled it to plead fraud, the link between the SoF and individuals has never been explained.

But in any event the nature of the case on fraud as regards individuals has always been inferential and little has really changed. The 2011 IKB CS Action did name names: two of the same names named in the pleading, namely Mr Marriott and Mr Kimura. Although the universe of names was ultimately considerably expanded, in closing it was notable that when pressed by the court as to the “main targets” the answer went thus:

“Messrs Marriott and Kimura, as your Ladyship knows, we say are, putting it rather colloquially, bang to rights here, …in view of … what those two gentlemen likely knew about the CDO operation within Credit Suisse, given that Messrs Marriott and Kimura were co−heads of effectively the RMBS Group and we’ve seen reports back to them of some very serious misconduct matters …

Mr Serageldin … performed a very central role…Mr Healy … the chief risk officer …The other four... who really stand out are the four traders …Daniel, Dodman, Schoen and Vibert”.

It follows that based in the matters known in 2012 L30 could perfectly well have pleaded a case at least against Messrs Marriott and Kimura.

The other aspect relied upon is the one which becomes obvious when one looks at the Rhineland claims: whether L30 could properly plead matters pleaded by others in the US. L30 says, rightly that a claimant cannot “adopt” allegations made by another litigant, unless it has an independent basis for believing the truth of those allegations. That however begs the question of whether it has an independent basis for believing the truth of those allegations. A broad unparticularised assertion may well be something which a party will not feel able to adopt (unless for example it dovetails with other material which it has to similar effect).

But many of the allegations in the non IKB/KfW complaints were very far from unparticularised. Where, for example, one is looking at the pleaded results of a statistical exercise performed on original documents pursuant to a contractual right of inspection by reputable lawyers, that very pedigree provides its own independent basis for belief. This is an argument which applies to the MBIA and FHFA Complaints. L30 knew that a serious “deep dive” had been done by these litigants and that they had pleaded the results. I see no reason why this could not have been relied upon in the same way as (say) reports of conversations in the DoJ SoF. This is an even stronger argument when one comes to an exercise commissioned by IKB – even if underlying material was not made available to L30 or KfW, those acting for L30 would be able to have confidence in the robustness of that exercise.

Accordingly and despite Mr Hobson’s very skilful submissions on this issue, I conclude that prior to November 2012 L30 could with reasonable diligence have pleaded all the elements of the cause of action that it ultimately pleaded in 2018. This is not a case of knowledge of some other wrongdoing or wrongdoing of other banks; it is knowledge of the key facts relating to CS which would enable the CDO Representations to be pleaded by reference to falsity created by one or more facets of the RMBS Misconduct.

It follows that the protection of section 32 (and 14A) of the Limitation Act is not available to l30, and the claims advanced are time barred.

Of course the question of whether the knowledge L30 had is sufficient is one for me. However it is nonetheless interesting to note what the L30 witnesses had to say about their actual subjective knowledge of the RMBS Misconduct. As I have noted above, the L30 witnesses were honest and doing their best to assist. They were also clear that L30 did have pieces of information which the above analysis identifies as key and their deduction from that was that L30 had knowledge to bring a claim based on RMBS Misconduct by 2012.

In particular:

This was inevitable against a background where the L30 witnesses accepted that they:

This is also consistent with the fact that when the DOJ SoF emerged it was not apparently regarded as being of burning relevance to L30: Mr Bulgrin knew of the settlement in January 2017, but it was not until 2018 that he read it, or was prompted to read it by the KfW lawyers.

I add for completeness that in the light of the evidence it cannot matter that Mr O’Driscoll’s evidence was that he knew nothing of the RMBS Misconduct until the DoJ SoF. As there is direct evidence from L30’s witnesses one is not driven to inference from Mr O’Driscoll’s evidence. Moreover the suggestion that “it could not realistically be suggested that an outside investor in 2012 ought to have cottoned on to a fraud about which the bank’s co-head of CDO structuring knew nothing” is based on a false premise. The out of pocket outside investor was looking for information which could found claims; Mr O’Driscoll, a member of a team separate from the RMBS team, was not and was not directly concerned. This is an “apples and oranges” comparison.

Having reached that conclusion, it is unnecessary for me to deal with the argument as to whether L30 was precluded from disputing that the pleading standard in New York was lower than in England (which in practical terms approximated to an argument that it was precluded from pleading fraud).

The way the argument goes is this. L30 says that the fact that IKB considered that it could plead fraud in New York does not mean that L30 could have pleaded fraud in England. It says that it may have been wrong about its ability to plead fraud, or it may be that the requirements in England were different, or it may be that the facts pleaded in New York were not within L30's knowledge. L30 submits that regardless of whether IKB had proper grounds to pursue a fraud claim in New York, the position remains that L30 did not know, and could not with reasonable diligence have known, of facts and matters prior to November 2012 which would have enabled a fraud claim to be brought in England.

CS then says that L30, having satisfied itself it could plead fraud in the US, and having in correspondence disclaimed any difference between English law and US law, cannot now be heard to say that the material on which they did in fact rely was insufficient – referring back to AB v Ministry of Defence (“or AB”) and FII. Those authorities say (see above) that a party cannot say that a stricter approach be adopted than the one which was in fact adopted by them; here fraud being pleaded by other Loreley SPVs (on the instructions of the very same people) in 2011/2012 against other banks, L30 cannot say more was needed to plead fraud here. This is said to be particularly so given the history of the point.

In closing Mr Goodall took me in detail through the pleadings and the correspondence which had been exchanged. In essence:

L30's position was that its confirmation that it would not run a positive case that a lesser pleading standard applied in New York did not amount to an acceptance that the same standard applied. But it took two not entirely compatible stances. At times it said that the deletion was reflective of a decision that the averment was unnecessary; at others it said that it was simply saying that it advanced no position one way or the other. The only way these two cases could join up would be if one ignored the correspondence. But the reality was that this was not an unprompted decision; it was a decision taken against this considerable background of debate (and pressure) as to the significance of the point.

As has been made clear above, I would have been reluctant to decide this important issue on such an arid point. However had I had to do so I would have concluded that:

In all the circumstances therefore I would if necessary conclude that even without an explicit admission L30 in essence unequivocally indicated that this was not a point it would take.

Accordingly if it were necessary to say so I would conclude that it was not open to L30 to dispute that IKB had knowledge sufficient to plead a case in fraud from 2011.

This was a new argument, emerging in closing. It was touched on fairly lightly – and rightly so. It is plainly wrong; in that it is premised on the idea that there are separate claims which accrue at different times. However there are no separate claims: the claim is to unwind the Notes, which were entered into at a single point. The limitation argument arises from the fact that L30 (as noted above) needed only to be able to plead one aspect of the RMBS Misconduct giving rise to the fraud on which they relied. Limitation is therefore an all or nothing argument; it is not a case where part of the claim can be saved by unpicking the strands of RMBS Misconduct.

I should not leave this subject without dealing at least briefly with the “reality checks” on which L30 placed so much emphasis: why would it be that L30/KfW had all the knowledge necessary to bring this claim, yet sat back and let a claim slip by? Or, putting it another way: if it took the US DoJ (with “close to unlimited resources and extensive coercive and evidence gathering powers”) until 2017 to piece the fraud together how can it be expected that L30 should have done so much better?

These are both misconceived points, which on further analysis only serve to underline the conclusion to which I have come. I will deal with this second question first. Even assuming that the DoJ SoF did set out the details of a fraud (which, for reasons given elsewhere, it did not) the DoJ investigation covered not just investigation, but also negotiations and settlement discussions and the drafting of documents for public consumption, all done with no particular deadline in mind. Even if it were a question of “uncovering”, the DoJ did not have a limitation clock ticking against it. Nor in many respects did the DoJ report “uncover”: it is in fact replete with footnotes to other sources, such as the FCIC Report.

As for the first question: why would KfW sit back, against the knowledge which I have outlined above? The answer was relatively clear from Mr Bulgrin’s evidence that: “It did not occur to me that there was any sufficient linkage to merit any further follow-up between (i) IKB's action concerned with RMBS and (ii) any claim that the SPCs might have in relation to CDOs.” The knowledge of the RMBS Misconduct therefore did not naturally link itself in the minds of those at L30 or KfW with the CDO transaction. It was only when approached by RPC in 2018 with a memo which apparently posited a means whereby the link could be drawn that the possibility of a claim was considered. This is a point to which I shall return below.

Accordingly I conclude that L30’s fraudulent misrepresentation claim in this action is time-barred. It follows that the claim based in negligence is also time barred. That conclusion follows from an analysis of what L30 knew or could with reasonable diligence have discovered. It coheres entirely with the timeline of events. It is not contradicted by L30’s “reality checks”.

Consequently the issues on the making of the CDO Representations, whether they were understood and relied on and whether some aspect of the vast amount of material on the RMBS Misconduct case falsified them are not determinative. I do however deal with them below for completeness and against the contingency that this first conclusion might be sought to be challenged elsewhere.

In this case L30 needs to show that in relation to the CDO transaction: (i) the defendant made a false representation; (ii) the defendant knew the representation to be false, or had no belief in its truth, or was reckless as to whether it was true or false; (iii) the defendant intended the representation should induce the claimant to enter into the contract; and (iv) the claimant was induced to enter into the contract by the representation. In this latter context there arises a debate about the necessity of awareness/understanding of the representation.

The first step is the making of the representations.

Before turning to consider the facts a brief recap is useful.

The starting point is that a representation is:

Express representations are the most straightforward form. While these are less in issue in this case, it is worth noting (for the purposes of read across into implied representations) the uncontroversial propositions that:

Here one moves into the context of representations arising out of the representor's conduct rather than their words, or the implications of their words and conduct combined.

One of the passages on which reliance was particularly placed by L30 is IFE [50]:

“In determining whether there has been an express representation, and to what effect, the court has to consider what a reasonable person would have understood from the words used in the context in which they were used. In determining what, if any, implied representation has been made, the court has to perform a similar task, except that it has to consider what a reasonable person would have inferred was being implicitly represented by the representor’s words and conduct in their context.”

In other words the exercise is in essence the same as that conducted for express representations - except the question is what a reasonable person would have inferred was being implicitly represented by the representor’s words and conduct in their context.

Generally, but in particular when it comes to implied representations, determining whether any, and if so what, representation was made by a statement requires (a) construing the statement in the context in which it was made, and (b) 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.

That context will include, importantly, the contractual and other documentation passing between the parties. This is a point which has been noted repeatedly in other cases dealing with complex banking transactions:

That context may have a significant impact when it comes to the question assessing the impact it might be expected to have on a reasonable representee in the position and with the known characteristics of the actual representee. It is said to be so here. There will be cases where statements are accompanied by qualifications or explanations which indicate that the putative representor was not assuming responsibility for the accuracy or completeness of the statement or was saying that no reliance can be placed on it: see Hamblen J in CRSM [222] (citing Christopher Clarke J in Raiffeisen [86]).

“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.”

It is worth noting that this is a question which tends to come into focus in the banking context. The CRSM case was a CDO case; CRSM claimed that it was induced to buy four sets of structured notes from the defendant (Barclays), and to agree to the subsequent restructuring of the notes, by misrepresentations as to the risk of default which in each case were fraudulent, or which in any event Barclays did not believe or have reasonable grounds to believe were true. In concluding that there were no such representations, Hamblen J placed weight on the qualifications and disclaimers in the fact sheets which he concluded made it plain that CRSM was to be responsible for its own independent appraisal of the risks in respect of the notes and could not rely on Barclays in connection with the matters referred to. Raiffeisen concerned allegations of misrepresentations in the context of a syndicated loan, again in the context of complex documentation with disclaimers. In concluding that there were not actionable representations the judge observed rather tartly that “The court may regard a sophisticated commercial party who is told that no representations are being made to him quite differently than it would a consumer.”

Central to the debate before me was the significance of the so called “helpful test” enunciated by Colman J in Geest plc v Fyffes [1999] 1 All ER (Comm) 672 at 683B. L30 contended that it was highly significant in three places. First it contended that it per se justified the conclusion that some of the implied representations had been made (C13 and C14) and that it was an element (alongside prior authority) in finding representations C10-12. Secondly it contended that the test had significance at the reliance stage. Thirdly it was said to have relevance at the stage of finding fraud; L30 argued that knowledge that the CDO group was selling CDOs linked to the credit risk of CS-securitised RMBS, plus the helpful test, was enough to lead to a conclusion that implied representations C10-14 and D20 had been made.

The relevant passage of Colman J’s judgment is this:

“Where there is no express misrepresentation, the first question to ask is whether there has been any implied misrepresentation at all and, as with any other type of contract, the essential issue is whether in all the circumstances relating to the entering into of the contract of guarantee or indemnity, including in particular (a) the nature of the contract between the beneficiary and the principal debtor, (b) the conduct of the beneficiary and (c) express representations made by him to the surety, it has been impliedly represented to the surety that there exists some state of facts different from the truth. In evaluating the extent of the beneficiary’s conduct a helpful test is whether, having regard to the beneficiary’s conduct in such circumstances, a reasonable potential surety would naturally assume that the true state of facts did not exist and that, had it existed, he would in all the circumstances necessarily have been informed of it.”

That test was considered and expressly approved in the case of Property Alliance Group v Royal Bank of Scotland (“PAG”) [2018] 1 WLR 3529 at [132], having been followed on numerous occasions at first instance, most notably perhaps in IFE and Raiffeisen. The Court of Appeal gave that approval in these terms:

“The present case appears to be the first in which Colman J's test has been considered by the Court of Appeal. We do think it is a helpful test, in relation to the existence of an implied representation, to consider whether a reasonable representee would naturally assume that the true state of facts did not exist and that, if it did, he would necessarily have been informed of it. To that extent we would approve the dictum …. but that is not to water down the requirement that there must be clear words or clear conduct of the representor from which the relevant representation can be implied.”

Both these citations, when given in full, make clear the problems which face L30 in its approach to the “helpful test”. The first problem is that, as is clear from these citations, the test has thus far been used simply at the stage of deciding what if any representations have been made. The second problem is that even in that context it has not been given (by Colman J, whose use of language is always scrupulous) or adopted as “the test” for determining the existence of an implied representation:

The necessity to keep this latter part of the equation (the need for clear words or conduct) in mind is illustrated by the cautions in the other authorities. Those as to context have already been given. But in addition:

That need for balance is required by the underlying position as a matter of law – there is generally no obligation to disclose information, as explained in Banque Keyser Ullmann SA v Skandia (UK) Insurance Co Ltd [1990] 1 QB 665 at 798-799:

“The general principle that there is no obligation to speak within the context of negotiations for an ordinary commercial contract (though qualified by the well known special principles relating to contracts uberrimae fidei, fraud, undue influence, fiduciary duty, etc.) is one of the foundations of our law of contract, and must have been the basis of many decisions over the years. There are countless cases in which one party to a contract has in the course of negotiations failed to disclose a fact known to him which the other party would have regarded as highly material, if it had been revealed. However, ordinarily in the absence of misrepresentation, our law leaves that other party entirely without remedy.”

As Christopher Clarke J makes clear in Raiffeisen, the boundary between pure omission and misrepresentation must be carefully distinguished. The helpful test, stripped of its proper context erodes that distinction.

Therefore to the extent that L30 contends that representations can be found on the basis of the “helpful test” alone, I reject that argument. The test is that stated in IFE: what a reasonable person would have inferred was being implicitly represented by the representor’s words and conduct in the context in which they were used. The “helpful test” is no more than a “helpful test” in relation to one of the inputs into the overall decision.

L30 relies for the second part of the basis for finding an implied representation on what it asserts are authorities in its favour. It relies on the cases listed in Schedule 1 to its Opening as establishing a sufficient basis in law for some of the representations it seeks to establish – principally those as to honesty. It contends that where parties enter into a transaction, the courts will readily imply a representation that each party is acting honestly in relation to the transaction. This core representation of counterparty honesty is said to be found in:

These authorities and what they do and do not establish will be considered below. However what can immediately be seen is that these cases cover a range of disparate factual situations and none of them does more than say that in the circumstances of that case there was an implied representation (generally bearing some relation to honesty). None of them is authority for the proposition that an implied representation of honesty can be implied generally or in a wide range of cases.

Against that background it is necessary to look at the representations alleged against the relevant context. The background context (on which CS placed emphasis) is not really controversial.

There is also essentially no controversy about the structural and contractual context (the former of which was the context on which L30 placed emphasis). As to this:

A controversial point as to context was the “black box” allegation. L30 contended that an investor in the position of IKB CAM/L30 was reliant on the offering documents in respect of the CS RMBS having provided a faithful presentation of the credit risk; and that the securitisation of the CS RMBS was, in essence, a “black box” from the viewpoint of an investor in the position of IKB CAM/L30. It contended that other than the RMBS ratings/CDO rating, the pool characteristics as stated in the offering documents were the only source of information available to a CDO investor as to the quality of the loan pools that comprised the RMBS and that the CDO investor was entirely reliant on the bank which securitised the RMBS having done so in accordance with the representations/statements made to investors and rating agencies.

That is so – as Mr O’Driscoll stated the offering documents were “the public side information that every investor had” and such investors would have no knowledge of the pool characteristics other than as stated in the offering documents. But the “black box” argument implied an information imbalance which did not exist. This aspect of the argument can be seen in the formulation of the submission in closing “it doesn’t allow Credit Suisse to make representations and get away with it because to disabuse IKB would involve revealing essentially market−sensitive information…”

The absence of such an imbalance is because in reality the CS team structuring the Notes was part of the CS CDO Group and did not deal with RMBS. This was, as Mr O’Driscoll explained, a facet of the “general requirement within any investment bank, Credit Suisse being no exception, to separate people in possession of non-public information from people who are not in possession of that information, and that was the uniform rule that applied to the entire firm. There wasn’t a distinct separate policy for it as between the RMBS Group and the CDO Group.”

The result is that the members of the CDO team had no better access to underlying information than did IKB. The CDO team were in no position to give more information; there was therefore no asymmetry of information – as Mr O'Driscoll put it “they [buyers] had exactly the same information that I had and there was nothing I knew that they didn’t”. When asked for information on the Reference Portfolio the RMBS Spreadsheet provided information sourced from public sources such as Bloomberg, which L30 could equally access.

This information barrier was derided by L30 as a “complete red herring” which had no bearing on whether representations were made. But that information barrier is part of the context for ascertaining whether any representation was made; and the unthinkableness of this barrier being breached was vividly illustrated by Mr O’Driscoll’s obviously genuine astonishment in the following passage of evidence:

“Q. … there was nothing in that information barrier that would have prevented somebody in the RMBS Group of Credit Suisse from discussing the sorts of problems, the types of problems, that were being encountered with someone from the CDO Group, would there?

A.

You’ve got to be kidding me, sir, that would be a tip.

… Barred by insider trading rules in the United Kingdom and in the United States for, you know, a generation now.”

That evidence – addressed to the actual issue – is unaffected by the other points on which L30 relied, such as the very general and broad hypothetical indication by Mr Kaiserman that an RMBS problem that might affect what Credit Suisse CDO Group were doing would probably be flagged to CDO Group.

L30's approach to the “black box” which indicated a position of dependency on L30's part, was also in tension with L30's own witnesses' evidence:

During the opening of the case I indicated that I found the range of representations relied on difficult to comprehend and somewhat inconsistent with the way the case on representation was summarised in the L30 opening, which said:

“The essence of these representations is that CS was unaware of any conduct on its part which tainted the credit quality of the Notes or the .. CS RMBS or which otherwise undermined the reliance which could be placed by an investor in L30’s position on the credit ratings ascribed both to the Notes and the .. CS RMBS.”

Implicit in the concern I expressed was a query about how such numbers of different implied representations could all conceivably be made, effectively at the same time.

Mindful of that implicit warning L30 then rearranged their case on representations; not explicitly withdrawing any parts of the case but rearranging the representations into buckets or groups which more closely aligned with the case as opened and effectively pursuing only 9 representations arranged in three groups. Those rearranged representations are annexed to the judgment as Appendix 2B. What became apparent was that (doubtless with an eye to the authorities) core representations of honesty were the “hero” point, followed by representations regarding credit rating. I will deal with the representations as finally relied on in closing, and not as originally advanced.

The honesty representations were as follows:

L30's case in summary was that they connote that CS acted honestly in respect of the securitisations of the CS RMBS and had no knowledge of matters which tainted those securitisations.

These representations are, on the authorities, the apparently soft target. However, the authorities stop some considerable way short of saying that there is an implied obligation of honesty generally. The correlate is that it is necessary to consider carefully whether the honesty-related implied representations are justifiable here. It is not, as L30 submits, “perfectly straightforward”. L30 relies on two things (aside from the “helpful test”) to justify the conclusion that these representations were made. The first is the offering of the transaction against the relevant context. The second is the authorities.

Looking first at the question separately from the authorities, simply looking at the written down form of the representations certain things are apparent:

From that, base questions immediately arise: How is it said that they are made in these terms? How is it said that they not only co-exist as implied representations, but that they are each, distinctly made?

That naturally leads to a question as to how other aspects of the exercise – the determining of intention to be relied upon and understanding/reliance – can be discerned. In the end it seems to me that these secondary concerns do not require specific answers because the argument founders in any event when analysing a sub-set of the very first question, which is how representations relating to just some of the underlying reference obligations are spelled out.

The key point here is that the honesty representations, which L30 says are implied as a result of CS's conduct in proposing the transaction, via the 16 May email attaching the Termsheet and the RMBS spreadsheet derive from material which plainly covers all of the Reference Obligations; and yet the representations sought to be implied (and which, on the authorities, must be clearly discernible) relate only to a small subset of the Reference Obligations. Those are the Reference Obligations which CS itself had “structured, arranged, underwritten, issued, marketed or sold” – i.e. they distinguish between the CS RMBS (in relation to which CS played such roles) and the other 93 RMBS where it did not. That is therefore a subset of seven out of the one hundred Reference Obligations.

That is not a distinction which seems capable of founding an implied representation. An implied representation needs to be clear and capable of being spelled out from the surrounding circumstances. No representation relating to these seven underlying RMBS's can be clear – or even discerned at all - in circumstances where (i) the documentation which accompanies the conduct relied on relates to all Reference Obligations and (ii) that distinction has no foundation in the factual interactions between CS and IKB-CAM. There is no evidence that the CS RMBS were ever referred to by CS, or L30; or that there was any discussion about them, or about CS’s role (if any) in dealing with ratings for those deals. In fact, the CS RMBS were treated completely indistinguishably from the other 93 RMBS. There is no allegation to the effect that there ever was any distinction of treatment. CS put it this way “There is therefore no ‘hook’ – in the form of express statements or conduct – upon which these implied representations are said to hang.” Another way of putting it is to say that the allegation that there was an implied representation on these points lacks its basic foundation.

This is a point which L30 ignores when praying in aid Colman J's “helpful test”: it simply cannot be said that IKB CAM/L30 would naturally have assumed the state of affairs reflected in representations C10-C12 in a situation where it did not know (or apparently care) that there were CS RMBS within the Reference Obligations. This feeds into CS's submission (which I consider well founded) that this set of representations has been reverse engineered. There is simply no way when looking at the relevant documentation that anyone would take representations specific to a subset of underlying reference obligations as being manifest. The conclusion seems inescapable that these representations have been constructed to dovetail with the specific facts sought to be captured.

It follows that the allegation that there were honesty related representations falls at the first fence.

I should add that I would in any event conclude that the Group 1 (Honesty) representations fail for another reason. And that is that the second base relied on by L30 (precedent) gives way; the representations here are not analogous to those which have been found in other cases. In addition they are not ones which would have a material impact.

These two points elide and are therefore best considered together. The factual centre of this issue is that the relevant representations concern the existence of RMBS Representations; rather than being representations directly about the underlying subject matter of the RMBS Representations. There is a distance between the subject matter of the two. While L30 says (as it did in closing) that “the mischief was the bank selling a CDO product which packaged credit risk of underlying RMBS which the bank knew to be misdescribed” or focuses on Dr Bauknecht’s consideration of the stated pool characteristics of the underlying RMBSs because their credit risk drove the credit risk of the CDO, that description or concern does not map onto the representations it alleges. For a misrepresentation case there has to be a representation. The representations are (for example) that CS had no reason to believe that the RMBS offering documents made false representations regarding credit quality (a slightly contracted version of representation C10). In focussing on the mischief of the credit risk of the RMBS L30 seeks to move the subject matter of the RMBS representations (in the past, to others) into the CDO Representations when in fact the two are distinct.

That distance is not merely a conceptual distance, but a practical distance, which can be seen when one considers the question of what effect those representations could have on the performance of the Notes. The misrepresentations to third parties did not and could not affect the RMBS itself, or its credit quality or performance. Each RMBS (good or bad) was what it was; the underlying misrepresentations alleged affected who might buy those RMBS. They did not affect their performance. If CS had never made any representations to any investors, the RMBS would perform exactly the same and so would the CDOs.

Against that background one can then look at the cases where a representation has been found to have been made. A good starting point is PAG. In PAG the claimant was buying from the bank interest rate swaps referenced to LIBOR. What happened to LIBOR therefore affected directly what the claimant had to pay to the bank. As the Court of Appeal said at [133]: “RBS was undoubtedly proposing the swap transactions with their reference to LIBOR as transactions which PAG could and should consider as fulfillment of the obligations contained in the loan contracts”. Against that background (with the addition of some lengthy discussions), the claimant succeeded in persuading the court that RBS was saying that it was not itself manipulating and did not intend to manipulate LIBOR. The representation, the transaction and the bottom line were all intimately connected. In PAG the representation mattered because if RBS manipulated GBP LIBOR, that would very directly affect the published LIBOR rate and therefore a counterparty’s payments under an interest rate swap referenced to that rate.

But PAG did not succeed in persuading the Court of Appeal to find some other representations. PAG's submission (advanced there as here by Mr Lord KC) went wider than this and it was argued that the following representations were made:

“(a)

On any given date up to and including the date of each of the swaps: LIBOR represented the interest rate as defined by the BBA, being the average rate at which an individual contributor panel bank could borrow funds by asking for and accepting interbank offers in reasonable market size just prior to 11 am on that date …

(b)

RBS had no reason to believe that on any given date LIBOR represented anything other than the interest rate defined by the BBA, being the average rate at which an individual contributory panel bank could borrow funds by asking for and accepting interbank offers in reasonable market size just prior to 11 am on that date …

(c)

RBS had not made false or misleading LIBOR submissions to the BBA and/or had not engaged in the practice of attempting to manipulate LIBOR such that it represented a different rate from that defined by the BBA …”

Those submissions were rejected at [122] as being “unnecessarily complex”. As in this case it appears that there was a need during the course of argument “to reduce their complexity to something simpler”. The Court then went on to find the much more limited representation, and noted at [137] that “any implied representation cannot legitimately extend further than the particular transactions allegedly induced by the representations”. The Court of Appeal therefore rejected a submission:

The representations relied on here of course (i) extend to past conduct and (ii) are dependent upon representations made to others. Unlike PAG the representation is not as to the present and it is not as to the particular transaction. Nor does it have, as the PAG representation did, a very direct and clear correlation to, and effect on, the payment obligations under the contract between the parties.

In closing Mr Goodall sought to systematise the distinction between the PAG representation and the representations here via a confection of layers (hence Mr Lord’s description of it as “the layer cake analogy”) which stretched from Layer 1 of a CS managed CDO with a representation that CS would not act dishonestly in how it managed the Reference Portfolio, to Layer 6 (said to be analogous to these representations) where the representation was that CS had been honest when telling investors about how it diligenced loans which might end up in an RMBS in the Reference Portfolio.

The cake was, in truth, somewhat indigestible, but it had value in forcing attention on the different elements rolled up into the CDO Representations and reminding one of the rigour involved – and demonstrated by the Court of Appeal in PAG.

I therefore conclude that L30 is wrong to say that there is no material distinction between the facts here and what the Court of Appeal has implied in PAG. There is also nothing in any of the other authorities relied on by L30 which enables it to say either that life has moved on since PAG or that there is a different line of authority which assists with these problems. In each of those cases there was a clear and direct linkage which is lacking here. Here there is a discernible link, but it is a link which is indirect and at some distance.

It is not therefore fair to say as L30 does, that the honesty representations here are straightforward or grounded in the authorities. L30 is not seeking to capture dishonesty in connection with the CDO Transaction, but rather dishonesty in relation to something completely different (representations made about a small subset of the RMBS within the Reference Portfolio). If one were to search for a directly equivalent representation to PAG it would have to be something along the lines of: CS would not act dishonestly in how it manages the CDO. Consequently L30's case falls not within the ambit of previous authority, but well outside it.

This was to an extent conceded by L30 in closing, when Mr Lord said that

“it’s not an altogether helpful endeavour to seek to fit our CDO representations into the glove of the LIBOR reps any more than it’s helpful to seek to fit the LIBOR transaction into the Spice Girls transaction. The Goldman Sachs v IFE test, and helpful test, fall to be applied to the particular facts of the transaction at hand.”

As I have already indicated the classic IFE approach using the “helpful test” only as it properly applies, produces a negative result for L30. It was only if the authorities provided a route through that the case could succeed. But they do not provide that route.

In addition the representations (particularly C10 and C11), even if they navigated these hurdles, are at least in a degree of tension with the contractual context. Representations C10 and C11 (said to arise from inter alia the Termsheets) directly concern the completeness and accuracy of the publicly available prospectus documents for the Reference Obligations. Both the Termsheet Important Notices and the Series Memorandum make clear that no representation was made by CS as to the completeness of the information provided and no responsibility was accepted by CS for any omissions, and that L30 had to make its own, independent assessment of that credit quality.

The authorities indicate that this presents a difficulty for the party seeking implication:

For this reason too the honesty representations alleged would fail.

I should mention for completeness the very considerable reliance placed by L30 on evidence from the CS witnesses as to what investors would have assumed. L30 submitted that “CS own witnesses accepted the core underpinning of these implied representations” i.e. that IKB would naturally have assumed that CS had acted honestly in relation to the securitisation of the CS RMBS, and had no reason to believe that the information in relation to those RMBS was false or misleading in a way relevant to an assessment of the RMBS. L30 says that “while the legal test remains an objective one, the evidence of CS’ witnesses is highly relevant as evidence of what a counterparty in the position of L30/IKB CAM would reasonably understand” and that it helps to inform the assessment of whether to imply a representation.

As noted during the course of argument the question of whether a representation was made is one for the court based on its assessment of the context. The subjective opinions of witnesses as to what investors would have assumed is both irrelevant and inadmissible for these purposes. L30 relied upon Males J in UBS v KWL [2014] EWHC 3615 (Comm) [740]-[743] as establishing the proposition that such evidence was relevant and admissible. However that passage does not support the submission made. While clearly in that case on some very striking facts (the bank's knowledge that an intermediary had proposed the fabrication of some documents) some quasi-expert evidence as to banking practice and the expectations of a reasonable and honest banker was given, no principle or ruling is given. Indeed, in the (very particular) context of that case, it does not seem to have been contentious.

That factual occurrence in one case does not however mean that the evidence of any banking witness will automatically be relevant or admissible even as to banking practice or the understanding of a reasonable and honest banker in relation to a particular type of transaction. Further here we are (as already noted) not looking at simple basic propositions for which all the building blocks are well established; but more complex ones where (for example) the building block of knowledge of the inclusion of the CS RMBS was not established.

As for the proposition that this evidence should be taken as important material information to inform the assessment of whether to imply a representation, it is just possible that if appropriately put against the full relevant background there might be some assistance to be gained from answers from those involved. But the questions which were put and which founded the answers upon which reliance was placed were not so put, but rather were based on taking documents out of context

These are representations B7, A2, C13 and C14 namely that:

As regards representations C13 and C14 these can be dismissed at once, essentially for the same reasons as the honesty representations in that:

The body of C14 deserves separate treatment. This fails for the reasons given, but itself encapsulates what appears to be a representation as to the Notes’ own credit rating, which need not be logically dependent on the “In respect of …” rubric. Were it to stand alone it would not therefore suffer from either the problem of reliance on part only of the Reference Portfolio or on the distance issues. It is much closer to the PAG formulation. There are however still the problems of:

So far as B7 and A2 are concerned the position is slightly different. In respect of Representation B7 (anomalously said to be derived from the now effectively abandoned (express) Representation A1) L30 admits that the express representation and indeed B7 as worded was (literally) true. However L30 contends that that true statement carried with it a further implied representation effectively along the lines of C13 and C14. CS's characterisation of this argument (“a convoluted mess which bears no relationship to reality”) is perhaps a little harsh. However it certainly (again) falls foul of the requirement that an implied representation be clear. The rating of the Notes was a condition precedent to the purchase of the Notes; that represented the parties' contractual bargain. There is no logic in the requirement for and no apparent basis for a further representation as to CS’s belief in the rating.

Substantially the same point arises in relation to Representation A2. The rating of the Reference Obligations was a matter of objective fact which was capable of being (and was in fact) checked by IKB CAM. There is no logic to or basis for a further implied representation as to the non-existence of the points which cannot themselves stand as representations for the reasons already canvassed.

It follows therefore that the claim based on these misrepresentations also fails at this first stage.

The sole representation in this category is Representation B8: “CS impliedly represented that L30, as a prospective investor, would be able to make an informed determination of the value and credit quality of the Reference Obligations.”

L30’s case is that by stating that a potential investor was required to make its own independent determination of the value and credit quality of the RMBS, this necessarily implied that it would be possible for potential investors to make an informed determination of the value and the credit quality of the RMBS. If CS was aware of something which disabled an investor, however sophisticated and with full access to publicly available information, from making an informed determination of the value and credit quality of the RMBS, CS could not act honestly in saying that it fell to each investor to make its own assessment. That statement carried with it an implicit representation that CS was not aware of anything which would disable an investor, however sophisticated, from making an informed assessment of the credit risk of the RMBS.

It is tempting to think that such a representation does, as L30 suggests, flow from the contractual documents. However when more closely examined it is possible to see that the representation contended for is not a simple correlate, but a considerable extension, which justifies the submission made by CS in opening that L30 is seeking to convert an express warning that the Notes could only be suitable for sophisticated investors into a positive representation that the Notes were suitable for sophisticated investors and then (in substance) a representation that if the facts relied on by L30 existed, they would be disclosed. In other words, it “flips” a disclaimer, by trying to make it into a term encompassing a positive duty of disclosure (which is prima facie itself inconsistent with the other express terms).

The starting point here is the passage which is used as the springboard for the implied representation. It is not itself a warning, or a warranty. What it is, is part of Section 6 (Certain Risk Factors) which outlines 14 risk factors, ranging from “Risky Investment” through “no other Recourse” via “Potential Conflicts of Interest” to “Complex Tax Issues”. Part 6 is a 3 page section which comes after Parts 1-5 (General Terms, Transaction Description and Capital Structure, Portfolio Manager, Portfolio Information and Credit Default Swap Provisions.) The text in question is at page 20 of the document:

“Asset Backed Securities Market: The Reference Obligations will consist of asset backed securities, which are subject to a variety of risks that may adversely affect creditworthiness and/or performance of the Issuer and that could adversely affect demand for the Notes generally. Each prospective investor must make its own independent determination of the value and credit quality of the Reference Obligations.”

All of Parts 1-6 come after the section called “Important Notices” which contains the warning about the accuracy and completeness of the information and the statement that no representations are being made.

Further the first warning within Part 6 says this:

“Risky Investment: Investing in 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 will not be principal protected and investors in the Notes are exposed to full loss of principal. Only prospective investors who can withstand the loss of their entire investment should buy the Notes.”

Thus even within Part 6 the warning actually addressed the need to evaluate the risks is directed to “the risks and the merits of an investment in the Notes”. It is not a wording which is directed to “the value and credit quality of the Reference Obligations”.

Thus when one looks at the documentation at all holistically it becomes impossible to see how the wording relied on can conceivably spell out, essentially by way of a logical correlate, the representation relied on. If this one passage were to give rise to a correlative representation it would be hard to see why all the other parts of the text should not do likewise. In the context of such a lengthy document if that were the case, the result would be that the Termsheet would metamorphose into something positively bristling with add-on implied representations (much like, as I suggested in argument, a covid germ); and yet at the same time in the important notices at the start of the document the making of representations within the document is expressly disowned. That cannot be right.

The single representation here, D20, is that: “In preparing, approving and supplying the draft Series Memoranda, CS impliedly represented that it was acting in good faith, that is, not knowingly putting forward information in the draft Series Memoranda that was likely to mislead.”

D20 is the sole representation from Category D which is pursued. It suffers from three considerable difficulties:

These points present a formidable hurdle for the D20 case on the making of an actionable representation. I do not consider that they can be overcome; that conclusion aligns with the fact that the above points were not dealt with by L30 in oral closings, with submissions focussing instead on whether a document of a third party could give rise to a representation, whether Mr O’Driscoll accepted that the representation was made and whether it conflicted with disclaimers.

It follows that for the reasons I have given the CDO representations all fail at the first stage: I conclude that the alleged CDO Representations were not made.

Although many of the representations pleaded were not pursued I should add that the decision not to pursue them was plainly right. It is not the case that L30 has (under some pressure from the bench) pursued the wrong representations. The narrowing of the case did focus on the stronger end of the pleaded representations. The reality was that the vast majority of the representations were doomed to failure for one or more of the following reasons:

In other words the Series Memorandum made it clear that CS may have material information arising from such roles but that no such disclosures would be made. The Category C Representations plainly seek to go behind this clear agreement.

There were also issues (to which I have alluded in passing in relation to the honesty representations) as to the way in which the representations depend upon numerous highly complex or elastic concepts. As I have noted it is hard to see how such difficult and complex concepts could be suitable for implication and could only be addressed expressly (which in many cases they were: see above). For example CS asked: “How was the reference to RMBS which CS had ‘structured, arranged, underwritten, issued, marketed or sold’ (which features in each of the Category C ... Representations) somehow said to be conveyed implicitly? How would the reasonable observer have picked up, from the mere offer of sale of the Notes, that CS was making representations covering all of these highly specific roles/activities and not merely some or other of them?” These were but some of the questions which arose and which L30 did not seriously address. They do however serve to illustrate why the implied representation case was in large measure doomed to failure.

There was also a fatal lack of clarity and rigour in L30’s case in that it sought to imply Representations which were couched in the conditional alternative – either that RMBS would be affected or that there was a “material risk” that those RMBS might be affected. The litigation charms of this approach were easily understood – if the case could be posited on material risk/might there would be no need to show that any of the CS RMBS (ie the seven particular CS RMBS) were actually the subject of any alleged misconduct. But such a formulation simply cannot provide a basis for a clear implication as the authorities cited at [123(3)-(4)] of Marme make clear. Only in such a large basket of proposed representations could such a defective argument hide away.

At this point I return to the point dealt with at paragraph ‎284 above. As I noted there, it appears that the reason why this case was not pursued until after it was time-barred was because (even with careful attention being directed to possible claims which could recoup losses on CDOs) no-one at L30/KfW saw a connection between the CDO transaction and the RMBS Misconduct. They did not see it because it did not present itself to their minds. It did not present itself because it was not clear, or obvious; that is precisely why there are no misrepresentations. In effect, the reason the claim became timebarred is because there was no claim to see.

My conclusions to date mean that this issue does not arise. However in the light of the fact that there has been full debate on the point, and that I have been asked by L30 to revisit my conclusions in the case of Leeds City Council v Barclays Bank [2021] QB 1027 (a case in which I had granted permission to appeal but which settled just prior to the hearing of the appeal) I will deal with this point in some detail.

The core requirement is to establish that the claimant was materially influenced by the representation. The question of inducement is centrally concerned with establishing a causal link between the conduct of the defendant and the conduct of the claimant. It is a control mechanism: if the defendant told a lie, but that lie had no effect on the claimant’s thought-process or conduct, no remedy will be available. The overarching question is whether the representation operated on the mind of the claimant or materially influenced the claimant.

The issue which has caused much debate in this case and elsewhere is the balance between the authorities which say that the representee must be aware of the representation in the sense that it must be the subject of “contemporaneous conscious thought” or “actively present to his mind” and those which indicate that the claimant is assisted by an evidential presumption (“the presumption of inducement”) which infers inducement from the facts that a material misrepresentation was made and the claimant did enter the contract.

It may be not insignificant that this debate has been conducted almost exclusively in the context of cases where – as here – representations were found not to have been made at all.

L30 of course submits that awareness is not a prerequisite. It submits that “commonly, … implied representation will be a matter which the claimant tacitly, rather than consciously, has in mind”. 

It submits that the correct analysis can be summarised as follows: 

This approach is said to be supported by the following authorities:

L30 contends that the cases suggesting a requirement of awareness (Property Alliance Group v Royal Bank of Scotland [2016] EWHC 3342 (Ch) (“PAG first instance”) per Asplin J, Marme Inversiones 2007 v Natwest Markets [2019] EWHC 366 (Comm) per Picken J and Leeds are obiter – and have taken a wrong turn.

In relation to Leeds specifically it was submitted that: 

If (contrary to its primary submission) awareness is the test L30 submitted that:

As my views on the subject align significantly (though not entirely) with the submissions of CS I will not rehearse those submissions here, save to say that they were presented with great clarity and élan by Mr Sher. 

I will start with what Leeds says, or attempts to say. At the heart of it is what I saw as a necessary and logical bridge between the representation and inducement (Leeds [67], [70]). The starting point – which is not controversial – is that representation has to cause inducement (see Lord Selborne in Smith v Chadwick 9 App (1884) Cas 187 “it must have produced in his mind an erroneous belief, influencing his conduct.”) and that causative link has to be at least capable of being discerned. 

That bridge consequently has to exist in all cases of representation: express, implied, by words and by conduct or any combination thereof (Leeds [65]). It is sometimes referred to as “understanding” or “conscious thought” or “active presence”; or it may be referred to as “awareness” (Leeds [57]). The formulation of “active presence” is classical – it dates back to the judgment of Bowen LJ in Edgington v Fitzmaurice (1885) 29 Ch D 459. That was a C19 prospectus reliance case which dealt with the competing claims of assumption and representation. The other formulations have arisen in other cases subsequently.

Reliance was placed by CS on Smith v Chadwick with Mr Sher submitting that in order to succeed in a claim for deceit, the claimant must show not only objectively that the representation had a meaning which was false, but also that the representee understood the representation in that false sense. This is true; however as I noted in Leeds at [68], Smith v Chadwick was dealing with the question of a representation which was capable of being understood in two senses. In this context, while it is true that Smith v Chadwick and the extensive line of cases following it in essence endorses the requirement of awareness/active presence of the representation, those cases have a specific focus (the particular sense of the representation) which is different. I therefore consider it preferable to negotiate by reference principally to the authorities where the question of awareness/active presence itself was in issue. However Mr Sher does rightly note that the issue of ambiguity is also likely to arise where representations are implied, simply because of the difficulty of spelling out a particular representation without words.

The way in which the requirement of awareness/active presence is characterised and what one might expect to see to satisfy that requirement will differ according to the circumstances of the case. An oral express representation might well best correlate to contemporaneous conscious thought – but it is in such cases that the understanding/awareness requirement is least often problematic on the facts. Simple representations by conduct may well correlate most closely to awareness with a quasi-reflexive understanding (Leeds [146-8]). That does not mean (as may have been suggested in some subsequent considerations of the case) that there is no such requirement of awareness or understanding in simple representation by conduct cases. What it means is that tests such as “contemporaneous conscious thought” are often inapt in such cases and other ways of expressing how the bridge between representation and reliance takes effect will be more apt. I remain of the view expressed in Leeds that the awareness/understanding bridge is however always present and it is always distinct from assumption – though in simple cases the process of awareness and understanding may look like assumption. 

That leaves a question as to how the presumption of inducement and the counterfactual of truth fit in. 

As to the presumption, my conclusion remains as stated in Leeds at [140] that logically the requirement of awareness precedes the presumption of inducement. This is consistent with Marme [2019] EWHC 366 (Comm), Vald Nielsen A/S v Baldorino [2019] EWHC 1926 (Comm) at [154]; BV Nederlandse Industrie Eiprodukten v Rembrandt Enterprises [2020] QB 551 and ACL Netherlands BV v Lynch [2022] EWHC 1178 (Ch) at [515(7)]. No argument was really addressed to this point, though I shall return to it below. More in issue was the counterfactual of truth – here re-dressed as the “helpful test” in another guise or eliding into it.

Dealing first with the counterfactual of truth pure and simple, I consider the arguments that such a test can stand in for the relevant test of understanding/awareness are doomed to failure. The counterfactual of truth is not without purpose: it can be used as (another) helpful test applicable in some cases. For example, particularly where there is some doubt about whether a representation has been received or understood, the counterfactual of truth can provide some evidence which is relevant to deciding the question of whether there has been receipt or understanding.. So it can be helpful in some cases. That is effectively the point being made by Flaux J in Parabola Investments Ltd v Browallia Cal Ltd [2009] EWHC 901 (Comm) [106]. ]: “where the evidence is that, had the claimant known the true position, he would have acted differently, that in itself demonstrates that the fraudulent misrepresentation, which by definition does not reveal the true position, “was actively present to the mind” of the victim of the fraud to a sufficient extent to establish inducement.” (Footnote: ³)

But it only works when there are not competing causes for the action; and particularly the possibility – of huge import in cases such as this one where representations of honesty are posited – that the representee's own pre-existing (mistaken) assumption may be operating. Christopher Clarke J in Raiffeisen at [187] makes clear that the counterfactual of truth is not necessary and:

“may not be sufficient, either. If it were, a claimant who gave no thought to any representation, or did not understand it to have been made, might be entitled to recover.”

As Mr Sher pointed out “the counterfactual of truth does not work as a gatekeeper because it does not distinguish between, on the one hand, mistaken assumptions not caused by the conduct and on the other, an understanding caused by the conduct or the representation.”

Testing this against the situation in DPP v Ray (aka Ray v Sempers [1974] AC 370, discussed in Leeds from [105]), CS suggested that the waiter would almost certainly assume that the customer is not a terrorist who intends to attack the restaurant. The counterfactual of truth is that the waiter would not serve the customer if they knew the truth. But it cannot sensibly be suggested that the customer is impliedly representing that he is not a terrorist. CS suggested that the best way through this was what they termed the “open question test” i.e that a hypothetical questioner asks what the representee understood was being said by the conduct and that on this hypothesis the waiter would never give an answer in favour of implied representation that the customer was not a terrorist. I part company with CS here, in that I fear that such a test remains too blunt an instrument. One might think that it is rather dubious whether the waiter in Ray would have answered the open question in favour of the implied representation found. Open questions are, as any examiner in chief knows, fraught with danger.

I also rather doubt whether the “open question test” would have worked for the implied representation found in PAG – a degree of leading would have been necessary because of the co-existence of the assumption that the other party was honest. Mr Sher might suggest that the answer suggests that PAG may require reconsideration; I obviously cannot. But further if there is to be a test it must leave room for the co-existence of assumption and representation; the open question test will struggle here.

Overall I deprecate the tendency to seek a single universally applicable test in a situation of this sort. In the case of implied representations one is discerning something on the basis of anything from simple conduct to a combination of actual words and conduct against a background of considerable complexity and sophistication. Ultimately, as here, in complex situations it may well be that as a result a clear representation cannot be spelled out; and it is thus that this debate can become a little artificial, because understanding/awareness is so often falling to be considered when there is no representation. But even so one cannot shut the door to the possibility (as in PAG); and in those cases an open question test will be as defective in a different way as the counterfactual of truth. In addition this search for a single test creates an extra formal stage within the process of analysing misrepresentation and hence adds to the already sufficient complexity.

I therefore maintain the view that, like the question of deciding whether a representation was made, the question is a nuanced one; and that the court and the parties are better focussing their minds on the need for the bridge (however expressed) and the reason why it is necessary than exhausting themselves in the search of the elusive single perfect test. 

I should here deal with the deployment of Geest by L30. At some points this appeared to be the central plank in the argument of L30. Mr Lord submitted that if the representation was being decided via Geest it was illogical or incoherent for the same test not to be deployed at the understanding level. Obviously that does not work on the approach I have taken because I consider that Geest is or may be a factor in one element of the decision as to the existence of the representation; what it is emphatically not is “the test”. But secondly even if Geest were “the test” at the stage of finding a representation, this approach of engaging it again at the inducement stage would collapse the stages of finding an operative misrepresentation. Perhaps inspired by reliance on the Spice Girls case, two [stages] become one. But having distinct stages is not incoherent – it is disciplined, and more significantly it is part of the gatekeeping function of the law because absent these stages non-disclosure becomes misrepresentation. This is of course legal heresy. 

Mr Sher pointed out that a similar argument was run in Marme [285] where Picken J pointed out that it was unsupported by authority. In Leeds [94] the argument appeared in the written submissions, but was not pursued orally.

Mr Lord also suggested that this approach of decoupling the test from Geest created a rogue’s charter, because a representation could be found, and then the second stage could not be surmounted. However there is no unfairness in not providing a remedy in respect of a representation which has not been received or comprehended. The greater unfairness is to provide a remedy when the standard requirements of the cause of action are not satisfied.

What is more the cases where this is a problem will be very few and far between; as I have noted, more often where there is a problem at stage 2 there has already been a fatal problem at stage 1.  It should be borne in mind that:

Next I note that it is not fair to say, as L30 does, that the cases suggesting a requirement of awareness are limited to PAG First Instance, Marme and Leeds. Those cases come from a fairly lengthy line of authority starting with (as I noted in Leeds at [60-8]) Arkwright v Newbold and Smith v Chadwick via Raiffeisen [80], BV Nederlandse [32], CRSM [224], Brown v Innovatorone [882][906], Foster v Action Aviation [2013] EWHC 2439 (Comm) [100-1], Leni Gas and Oil Investments v Malta Oil [15][164] and Vald Neilsen v Baldorino [2019] EWHC 1926 (Comm) [152]. Of these I pick just two lines, both characteristically clear statements by Hamblen J:

What then of the cases on which L30 relies? I conclude that there is nothing within any of them which either binds me or which persuades me that I need to revisit the analysis above.

I will take Volkswagen first and at length as the most recent case which considers many of the others.

Volkswagen was, as noted above, a case where buyers of cars claimed that Volkswagen had made implied representations, that the vehicles sold to them were in compliance with all legal requirements. The claimants alleged that had they known the true position, they would not have bought the car. There was no evidence that the claimants consciously had in mind the representations. Volkswagen applied to strike out the claims on the grounds that reliance was not established. That application failed, with the judge holding that the question whether reliance could be established in this case, notwithstanding the lack of conscious awareness, gave rise to a triable issue. In his conclusions at [94] he noted the very different context to that which prevailed in Leeds:

“The case before me is very different from Leeds (and indeed PAG and Marme). The conduct, and the representations to be implied therefrom, as pleaded at para 63 of the GPOC (albeit at some length) are both in fact relatively simple. They are not to be spelled out from a complex web of communications. And while in Leeds it was to be assumed that the implied representations have been made out, it cannot be denied that the whole context was one where the implied representations might have been difficult to establish and indeed were positively rejected in PAG and Marme.”

En route to that conclusion Waksman J conducted a review of the authorities. He accepted that there were cases in which reliance had been established by reference to the “counterfactual of truth”. Amongst those he numbered the following.

DPP v Ray (considered in detail in Leeds at [105] and following). At [94] Waksman J saw in the passage at [1974] AC 370 at 386:

“on the facts as found it is clear that all concerned (the waiter, the respondent and his companions) proceeded on the basis that an ordinary customer would pay his bill before leaving. The waiter would not have accepted the order or served the meal had there not been the implied representation.

a clear invocation of the [counterfactual of truth]. The analysis appears to proceed at least at some points on the basis that there is a "conscious thought"/counterfactual of truth dichotomy, rather than the more nuanced approach I was attempting to convey in Leeds. Later in the judgment he considered that if (as I saw it) one viewed Ray as a case of awareness, in the sense of quasi-automatic awareness, it offered a parallel for the VW purchasers. That analogy might well be seen as apt based on the much simpler factual background,”

Gordon v Selico. In differing from my analysis in Leeds Waksman J placed some weight on the fact that Gordon v Selico was not cited. While, particularly given the fact that Waksman J was considering a preliminary issue/strike out application, I see that the very fact of the case might be seen as a factor, I do not consider that it offers any real support for the analysis of L30. Before Waksman J Volkswagen argued that the case was distinguishable as a concealment case. He rejected that and concluded that it appeared to provide support for the counterfactual of truth analysis: “while the claimant could readily say that had he known of the dry rot etc he would not have purchased, it is hard to see how he could have been ‘consciously aware’ of the representation as distinct from making an assumption.” However I cannot agree with this. The judgment cited in Volkswagen was the appeal, in which the main issue was whether attribution of dishonesty by the contractor to the builder was correct. It is clear from page 233 of the report that that was underpinned by the following issues which had been decided at first instance: (i) that the concealment of dry rot amounted to an implied representation as to the absence of dry rot and (ii) that this had been relied upon by the purchasers to their detriment. These aspects were not appealed and therefore form no part of the reasoning in the appeal. The appeal in Gordon v Selico therefore has no significance.

I will however deal for completeness with the judgment at first instance, which did deal with these points. That judgment [1985] 2 EGLR 79 by Goulding J may have been an extempore judgment. There was a considerable range of issues: deceit, breach of terms of the lease, breach of terms implied by the lease, negligence, breach of trust. There was no detailed consideration of either the making of a representation or reliance upon it. Instead at page 8 of the judgment the judge considers the credibility of the witnesses who joined issue on the question of whether the agent of the defendants, Mr Azzam, knew of the dry rot. He alludes in passing to the fact that “I have evidence which, if believed, leads irresistibly in my judgment to the inference that Mr Azzam deliberately covered up active patches of dry rot without any attempt at eradication”. He then passes seamlessly on to saying “In conclusion I find that the fact of deceit alleged by the plaintiffs is proved.” The consideration of representation and reliance comes next thus:

“The law must be careful not to run ahead of popular morality by stigmatising as fraudulent every trivial act designed to make buildings or goods more readily saleable, even if a highly scrupulous person might consider it dishonest. But it is to my mind quite a different matter for an intending vendor to hide so sinister and menacing a defect as active dry rot. The case is fairly comparable, in my view, with the concealment of cracks indicating the settlement of foundations, considered in Ridge v Crawley …, and, in the Court of Appeal …. There the plaintiff relied on a combination of words and conduct, but I believe it to be the law that conduct alone can constitute a fraudulent misrepresentation (see Horsfall v Thomas (1862) 1 H & C 90 and Smith v Hughes (1871) LR 6 QB 597). In my judgment the concealment of dry rot by Mr Azzam was a knowingly false representation that Flat C did not suffer from dry rot, which was intended to deceive purchasers, and did deceive the plaintiffs to their detriment. I am satisfied that the plaintiffs would not have entered into a contract or accepted the lease had they known there was dry rot inside Flat C.”

What is apparent from this is that (i) this is not a case where the issues were broken down with at all the degree of thoroughness that has occurred in this case (ii) to all appearances the question of reliance was only ever considered in broad terms and (iii) no-one put in issue whether the representation found was understood by the plaintiffs. I therefore consider that this earlier judgment adds nothing to the sum on knowledge in this context. On this basis, despite the doubts so politely expressed by the editors of Clerk and Lindsell at 17-35, I do not consider that any disharmony between this case and the present analysis should cause concern (Footnote: ⁴).

The most significant of the judgments relied upon by Waksman J and thus L30 is indubitably Spice Girls v Aprilia [2002] EMLR 27. Waksman J disagreed with Leeds on the question of whether awareness was live in this case (see [92(2)]), and at [78-84] cited lengthy passages from the judgments at first instance and on appeal leading to the following passage:

"It is hard, in my judgment, to resist the conclusion that while the points as to awareness and CFOT (Footnote: ⁵) [were] specifically taken, this case establishes that at least sometimes, an implied representation, intended by the representor to be relied upon by the representee, which is accompanied at least by CFOT, can be sufficient”

On my reading the reference to [were] should be a reference to [were not]; there certainly were issues as to (i) understanding in the sense alleged and (ii) inducement. However at the level of the appeal, those hinged on the absence of a witness to these matters, and on understanding the battle was really as to whether the representation would reasonably be understood in that sense or not. There were not issues as to awareness per se, or as to counterfactual of truth per se.

The passage in the Leeds judgment may not be optimally expressed (De Verneuil Smith and Day in “Reliance: a comparison between the common law and s. 90A FSMA” perhaps rightly chide it for being “a bit thin”); however the facts remain:

It follows that I conclude that there is nothing in Spice Girls (even post Leeds commentary) which justifies a view that inducement can be established by reference to the counterfactual of truth without satisfying at least an awareness test. I note that although Messrs De Verneuil Smith and Day critique some of the Leeds reasoning (and do take real issues with the “contemporaneous conscious thought” approach) they do at least implicitly, in their consideration of Ray at p 390 accept the need for a causal link between representation and action.

In Man v Freightliner [2005] EWHC 2347 (Comm), Moore-Bick J found there to be an implied representation that accounts had been prepared honestly (see [78]). The judge held that Man had relied on this representation. That involved no finding that Man consciously had in mind the representation. Instead, the analysis was tested by reference to the counterfactual of the truth at [123]: “I think it is clear that the essential honesty of the accounts was fundamental to the discussions between the parties. Dr. Schubert said that MN would have withdrawn from the negotiations immediately if it had discovered that Mr. Ellis had been deliberately manipulating ERF’s books in any material respect, and one can well see why.” 

Other cases relied upon by L30 include UBS v KWL [2014] EWHC 3615 (Comm). L30 submits that Males J held at [785] that reliance was established with no finding that the representee had any conscious awareness of the representation but rather by asking what the representee would have done had it known the true state of affairs: “I have no doubt that Depfa relied on both misrepresentations in entering into the Depfa Back Swaps. If it had been told that UBS knew that the intermediary advising KWL was dishonest or that the transaction which UBS had already concluded was tainted by a conflict of interest affecting that intermediary, it would have had nothing to do with this proposed deal. I accept the evidence of the Depfa witnesses to this effect.” 

Here I tend to agree with CS that this is to take a very short passage of a very long judgment out of context. The section on fraudulent misrepresentation starts at [642]. If one looks at [666] of his judgment in relation to an earlier alleged representation (found not to have been made) the judge says that:… “[E]ven if such a representation had been made … there is no evidence (or at any rate none that I accept) that KWL was conscious of any such representation having been made to it, that it relied upon such a representation or that the representation played any part in KWL’s decision … There is nothing in Dr Schirmer’s evidence, and certainly no evidence which I accept, to suggest that he was ever conscious of the representation which KWL claims to have been made.” That shows that at least at this stage in the judgment the judge was applying the approach which follows from Edgington v Fitzmaurice.

Later in the judgment the judge concludes, based on what appears to have been a devastating cross examination of a witness on “general banking principles” by the “tag team” of Mr Lord KC and Mr Railton KC that:

I accept that this appears to show a conclusion of reliance being reached essentially by a link to assumption. However it must be borne in mind that this question of awareness does not seem to have been live at this stage of the argument. As such the case is somewhat akin to Gordon v Selico. Both are cases where there was essentially an active and conscious falsification of a fundamental representation which was being made implicitly; an active falsification layered on top of what would anyway be assumed. Plus – and significantly – in both cases the judge had on the evidence concluded that a sufficiently clear and clearly intended representation had been made. It may well be that in both cases it was considered imprudent to argue the point. But the fact remains the point was not live, and the case therefore cannot add anything significant to L30’s argument.

A similar point arises in relation to IFE v Goldman Sachs. The point did not arise because it was held that the implied representations as alleged were not made. It is accepted by L30 that the point was only addressed obiter by Waller LJ on appeal.

I conclude that the law does require that a representation (however made) is received by the representee and that to satisfy the requirements of reliance the representee must be aware of it/have it actively present to their mind when they act on it.

Mr Lord endeavoured to persuade me that in this context a test of “present to the mind” would suffice. This I regard as a dangerous step away from where the authorities stand, essentially because such a test too easily elides into assumption. “Actively present to the mind”, becomes “present to the mind”, from which it is but a short distance to “at the back of my mind” - which was exactly where L30 wanted to go because that was what Dr Bauknecht said. But where something is at the back of the mind is it because of assumption or representation? Very often it will be only because of an assumption unconnected to what has been said or done. Accordingly I consider that this approach merely provides another route to dissolve the distinction between representation and assumption – which is also the division between representation and non-disclosure. This is a problem which is particularly acute when (as here) one is dealing with implied representations as to honesty, because it turns every contract into a contract of utmost good faith.

I return finally to the fault line which plainly exists in the cases. The authorities above show that there are cases where on the particular facts in play reliance/inducement is found without any distinct evidence of understanding or awareness being identified. It is fair to say that when placed under the analytical microscope those cases do offer a different flavour to those which I have decided establish the orthodoxy which applies where a question arises as to reliance based on an issue as to the operativeness of a representation. But it can be seen that these cases where no evidence of understanding/awareness is separately discernible are, in type, very different cases to the ones where this issue is really live. There are two hallmarks which appear to drive the distinction. The first is that in the Gordon v Selico/Spice Girls type case the representation is simple and cannot well be missed by the representee. There is no issue as to whether the representation is discerned so as to be operative. In the Leeds and current type case, the representation is being said to be capable of being implied despite complex contractual provisions, usually in complex multifaceted transactions. That pushes to the fore a need to see what has been registered or understood.

The second (related) distinction is that in the Gordon v Selico/Spice Girls type case the representation is one which is at the heart of the transaction. Are you buying a good flat or a money pit? Are you buying the star power of a flourishing or an imploding group? Both of these factors mean that the question of awareness is one to which the answer is obvious (the representee cannot miss it) and where the whole question of reliance is, in the light of that, likely to be susceptible of being decided by the presumption of inducement. As noted above, it would have taken a brave counsel to raise this point in Gordon v Selico (when the court had been prepared to find a representation). In Spice Girls a similar flavour emerges from the use of the word “inconceivable” in [72] of the Court of Appeal judgment. In the banking cases awareness is far from obvious; so much else is being said more explicitly that there are real questions as to whether a particular implicit message is received and understood. A proffering of a complex transaction such as this is not analogous to the ordering of a meal or the raising of an auction paddle.

All of this of course feeds back in to the question of making of the representation. In these more complex cases the question will often arise either contingently (as here – arising only if I am wrong about the making of the representations) or (as in Leeds) based on an assumption which might never be made out; but the facts that drive the first conclusion impact on the situation where causation arises. In the simpler cases the answer is often either so obvious that there is no dispute, or the court may be prepared (instinctively, but I would suggest technically incorrectly) to elide the question into the presumption of inducement. That is in essence the point being made by Longmore LJ in BV Nederlandse v Rembrandt Enterprises [2019] EWCA Civ 596 [2020] QB 551 at [45] to which Mr Lord alluded in reply. None of this however affects the logic of the argument. Because points are not in some cases taken, does not mean that they are not there.

L30 effectively conceded that if the test was “active presence” or “contemporaneous conscious thought” its case could not succeed. That was an inevitable concession in a situation where there was no evidence from L30 as to anyone reading the Termsheets, still less as to the relevant understanding on their part. Dr Bauknecht never saw the 16 May email or the draft Series Memoranda. Indeed, on this basis the case in relation to the Category A and B representations fails even apart from the interesting question of awareness.

For A2, had this been surmounted, there would also be the difficulty that IKB relied on its own checking process to confirm ratings.

A case was advanced by reference to Dr Bauknecht and Mr Schirmer’s evidence on the Category C representations and representation D20. As noted above their evidence was problematic in that they had been “refreshed” before giving their statements, in a way which would not comply with PD57AC, in particular being supplied with documents which they would not have seen at the time and via an iterative process. Particularly on such a delicate topic, where recollection is being sought so many years ago, the effects of this are to leave the court in real difficulties in being able to establish that anything is a genuine memory.

Even had the evidence from the witnesses been such as to on its face support a case of receipt and understanding of representations I would as a consequence have had some difficulty in accepting that evidence, because I could not accept it as recollection of the facts. I conclude this without intending any disrespect to Dr Bauknecht and Mr Schirmer, who were plainly doing their best in difficult circumstances. The conclusion would if necessary be driven by the extent to which the statements made in the witness statements were revealed, on unpicking in cross-examination, to be reconstruction. The witness statements by the end of oral evidence bore the hallmarks of having been driven by the legal team’s analysis and approached by reference to what one might term case theory, rather than by eliciting genuine thoughts and recollections.

But fortunately I am not driven to this point; as matters transpired the evidence did not begin to come to that level, and even assuming that the evidence could be accepted as genuine recollection I would conclude that it did not satisfy the requisite test on awareness.

It was suggested by L30 that if the test was finessed to “quasi-automatic awareness” the answer was different, in that Dr Bauknecht would have taken the representations as “obviously the case” and these matters “encompass matters that would have been in my mind whilst I assessed the transaction and considered whether to proceed”; and that his oral evidence was that these things would have been “in the back of my mind”.

However in my judgment both these passages of evidence are not evidence of fact at all, but are plainly reconstruction. In essence therefore L30 offered no positive evidence on this. Further even if one were to accept it as a statement that Dr Bauknecht (to the best of his compromised recollection) did have this in the back of his mind, the evidence goes no further than evidencing an assumption. That is not changed by the background of trust; the background simply deepens the assumption. I do not accept the submission that they show that Dr Bauknecht “plainly had the requisite degree of awareness in relation to the substance of representations.” They do not evidence any sense of receipt of or understanding of the making of a representation. They illustrate precisely the danger outlined above of the tweak to the test which L30 proposed.

Further Dr Bauknecht’s evidence that he struggled with the concept of implied representations supports the conclusion that there was no such representation, or that if there was it was not understood by him or present to his mind.

There was also a pleaded case as to Mr Schirmer’s reliance. That was not strenuously pursued by L30 in closing. To the extent that it was it does not differ in any essential respect for the case based on Dr Bauknecht’s evidence. Mr Schirmer’s evidence was also based on reconstruction of what his understanding would have been. In his oral evidence Mr Schirmer readily accepted that his position was based on an assumption that CS would be honest based on his general expectations of his counterparties, deepened by the trust built up over previous deals. It was he said, “the basis on which I deal with them”.

Representation D20 is one said to have been addressed to and relied on by Mr Schirmer. Again even leaving aside the issues with Mr Schirmer’s obvious lack of recollection, his oral evidence went no higher than agreeing that he assumed honesty. There is on a fair reading of his evidence no basis for concluding that he understood representation D20 to be made or relied on it.

It therefore follows that the case on the CDO representations would fail at the level of reliance also.

In the light of the conclusions which I have reached already, the issues as to the RMBS Misconduct not only do not arise, but do not arise unless multiple conclusions were different.

It is therefore strictly speaking unnecessary to deal with any of the multiplicity of issues on RMBS Misconduct which occupied such a very large part of the trial, and doubtless the preparation time of the parties also. However for completeness I here give a summary of the reasons why I would (if necessary) have found that the case on RMBS Misconduct also failed, and failed by a very clear margin.

Further, largely in deference to the immense amount of time, work and cost which has been expended on this issue, I include at Appendix 1 to this judgment slightly more detailed reasoning on each of the strands of RMBS Misconduct.

The case on RMBS Misconduct relies on two things. The first is taking everything within the DoJ SoF and The 2017 Press Release (fact and commentary) as true fact. The second is the sheer number and complexity of the allegations (see above and Appendix 3 for the many elements of that case) – a “some mud must stick” approach.

This situation reinforces the point that both of L30’s overarching approaches must be rigorously interrogated. Both, when interrogated, can be found to be faulty. As to the DoJ SoF, this has been dealt with earlier. I have concluded that save as regards the Approval Representation the SoF has no legal consequences. It is evidence of the facts stated in it; but no more. The facts stated in it do not (save in one respect, as regards the Approval Representations) establish a basis for the RMBS Misconduct allegations. The rejection of the simple read-across from the SoF naturally impacts the RMBS case.

On the latter point (complexity) it is telling that at no stage has L30 been able to break out from the morass of allegations a single "hero" point. Even in closing L30 declined to prefer one strand over another, calling all elements “fundamental”. This meant the RMBS Misconduct was presented largely as a block, necessitating Mr Lord KC making submissions at a speed and in a telegraphic form which was accepted to be impossible to follow live in order to cover the myriad of highly detailed points. A prime example may be given: “Subparagraph or row (35), the reference there is to subparagraph (68) below, but that chiefly concerns Mr Sack’s evidence in cross−examination on what he thought the email meant. That is of course irrelevant to Mr Vibert’s knowledge.”

This absence of a “hero” issue is a striking point when one bears in mind that for one of the aspects – the Approval Representations – it was accepted by CS both that the representations had been made and that they were (objectively) false.

Those two hallmarks indicate the problems in L30’s case. The remainder of the case disappears when the structure is broken down.

It is instructive first to look at the case on the presumptive “bull point” of the Approval Representations.

As noted, it is common ground that there were representations and that they were literally false. The relevant phrase was that “all final loan decisions were made by CS senior underwriters, not third party contractors”. I have concluded in the more detailed section below that despite the fact that the Fulfillment Centres operated within tightly controlled parameters, referring any difficult cases back to CS underwriters, the fact that the vast majority of decisions on loans were made by the FCs under those parameters with only a very small proportion of risks being referred back to the CS underwriters, means that falsity was material (though the point is highly arguable).

So, in relation to this representation L30 establishes: (i) representation (ii) materiality (iii) falsity.

But falsity does not equate to deceit. While the objective meaning is agreed, as regards the Approval Representation the subjective understandings to which witnesses spoke cannot be said to be unreasonable or not believable. For example it is perfectly credible that “final loan decisions” could be seen as being either or both of the underwriting policy decisions as to criteria for easy cases, and the referred decisions on the grey areas/exceptions, which did come back to CS senior underwriters.

On that basis when one breaks down the case into what the individuals knew and the basis for saying that they knew the representations were false it becomes apparent that even though some (but not all) of the individuals (i) knew of the representation and (ii) knew of the facts which (objectively) made it false, there is simply no evidence that they subjectively knew of the falsity. That of course is a requirement: Ivey v Genting Casinos (UK) (trading as Crockfords Club) [2017] UKSC 67 which (as L30 summarised it) establishes that the question whether someone’s conduct was dishonest involves (i) ascertaining that person’s subjective state of mind and (ii) applying the objective standard of ordinary decent people.

Here there is no smoking gun on knowledge of falsity and no logical inference. Indeed the balance of the evidence suggests that the CS individuals identified by L30 as having the relevant knowledge subjectively did not know of the falsity. The same ambiguity in the language and evidence as to their understanding means that, whatever the agreed objective meaning of the representation, I cannot find that their conduct was dishonest by the standards of ordinary decent people.

Similarly the evidence is adverse to any conclusion that they intended the representation to be relied on in the false sense.

It follows that the case on the Approval Representations, though it gets some distance down the line, fails on knowledge of falsity. This is the closest that L30 gets to success on the RMBS Misconduct and this is not a borderline factual decision. It is a clear decision based on a complete absence of evidence. Thus even the simplest of the RMBS Misconduct cases fails.

When one deconstructs the other “strands” where the route to L30's case is complicated by extra hurdles, the margin of failure is even larger.

In essence, the only Representation which can be established as having been made at all is the Approval Representation. Of the other strands each, other than due diligence, fails on at least one basis at this preliminary level:

For many of the representations (other than the Approval Representation), even if they were established as having been made the case on falsity would also fail.

As regards knowledge and therefore fraud, the case breaks down entirely on many of the strands. The case of knowledge as finally pleaded (as it needs to be) against individuals encounters the telling problem that the material (such as it is) is exiguous in the extreme and is afflicted by the division between the two sides of the business: in order to try to bridge the knowledge gap allegations are made against individuals in both CDO and RMBS teams, yet in practice those who knew about the CDO and any representations in that context, were divorced from the RMBS context, and vice versa.

This problem led to a certain amount of cross-examination on the CS office floor plans, designed to demonstrate an interconnectedness between the desks which would enable this disconnect to be overcome. This evidence did not achieve its aim. Although L30 did elicit evidence that (for example) only one bank of desks separated Messrs Sack/Othman from Messrs O’Driscoll/Serageldin, that there was a lot of discussion which was not on email and that Mr Kaiserman often walked over to see people on the trading floor, that is a country mile from where L30 needed to be. The impression created by the floorplans and the evidence was not one of interconnectedness of these two desks but rather of the vastness of the business. Taking the floorplan, so small was the print that on an A3 version the names of the people at each desk were illegibly small, even to the junior members of the teams, and had to be captioned off at the side. The graphic of the layout resembled the Matrix.

Even within the RMBS-side witnesses there were issues as to the knowledge of the kinds of details which were involved in the RMBS Misconduct which relied on quite granular details of how loans were acquired and packaged and sold.

There are also repeated issues (which were relevant both to falsity and knowledge) regarding the fact that in the welter of extremely detailed evidence L30 has routinely elided the position as between the raising of issues as to acquisition of loans with loans that were securitised, and also as to issues arising on quality control, with issues which L30 says should have been picked up by due diligence. Thus heavy reliance is placed on documents where someone in CS is expressing themselves freely about QC results, on the basis that they are highlighting issues with DD, ignoring the fact that a QC issue does not mean that there was a breach of warranties or a DD issue owing to the facts that (i) both are judgment processes with a range of answers and (ii) they operated to different guidelines and on different material.

As regards many of the key individuals whose names have been advanced since the start of this litigation L30 lacks any link to the making of one set of representations or the other (or both). Thus ultimately L30 pointed at Messrs Marriott and Kimura as being their primary targets so far as knowledge was concerned with Mr Serageldin being just behind them. However:

For those individuals the case as to knowledge has a very slight base and never rises to the level of a credible case. For Mr Marriott for some representations the case on falsity and knowledge of falsity depends on a single point within an audit report which itself (i) documents that systems were being tightened and (ii) indicates that to the extent there was a problem someone else (Mr Fallacara) had signed off on the issues as resolved.

In relation to the other representations it rests on a single email chain, (mainly between Mr Vibert and Mr Sack) for example containing text such as the following: “FYI... this is indicative of what we spend our time on every day here in the field. We tell our investors that we do 100% due diligence on LBL, but then when we have a competent QC firm do an underwriting review, they flag all kinds of errors that our fullfillment centers did not catch….Our fullfillment process is broken.” Yet at the same time the wider context shows that:

While it was repeatedly and forcefully said that handwringing is no defence (i.e. that good intentions to improve process do not stand in the way of a conclusion of knowledge of falsity) the context makes the conclusion of knowledge of falsity impossible to draw. At best for L30:

For Mr Kimura similarly half the case depends on a handful of emails taken out of context, including another “emotional” email from Mr Vibert. For the other half, concerning three representations, it rests on a single email showing a prompt and concerned reaction to “concerns/problems the market has had with acceptable value variances” and overtly designed to address (i.e. solve) this issue.

Another person whose position is worth considering discretely is Mr Sacco. He is said by L30 to have been “integral” to the alleged RMBS fraud. There are two aspects here.

The first is that like so many of those named (other examples are Mr Daniel, Mr Vibert, Mr Kaiserman, Mr Nordyk, Mr Sack, Mr Fallacara, Mr Schoen, Mr Dodman, Mr Othman, Mr Salomon, Ms Baird) he had no ties to the CDO business and therefore no involvement in the CDO Representations. While some RMBS personnel did have involvement in the portfolio selection for the CDO, they were not any of those on the RMBS side against whom knowledge is alleged. In respect of those who are named the evidence seemed to go no further than that the CDO Group was selling CDOs linked to the credit risk of RMBS. There is no evidence that they knew any more than that.

There were no documents linking Mr Sacco to the Notes or the CDO business. His evidence was that he was “probably aware there was a CDO group” but that “I don’t know what CDO means”. L30 nonetheless invited me to find that “Mr Sacco is likely to have known that the CDO group sold CDOs linked to the credit risk of CS-securitised RMBS” and that his evidence that he hadn’t heard of a CDO until after the credit crash was untruthful evidence. The basis for this was (i) “Mr Sacco was no financial ingénu” and (ii) “He participated in a presentation alongside Mr Sohn (an ABS/CDO trader) where CDO trading was likely discussed.” That is a manifestly inadequate basis for such a conclusion. Parenthetically, it is perhaps illustrative of the strained nature of the case for a “bridging of knowledge” between the CDO and RMBS Groups, that the case against Mr Nordyk (another of the main targets vis a vis RMBS) in one respect rested on a suggestion that a reference in a single email to a visit to Las Vegas meant that he must have been to the conference of the American Securitisation Forum (and heard something about CDOs at a panel event) when there was also evidence that he had been to Las Vegas for family holidays and he on this visit did not intend to take his laptop, as he would for a work event.

Secondly even in the context of the RMBS Misconduct case against Mr Sacco, it was striking how far short of the target the case in relation to him came. The L30 case theory posited Mr Sacco, who was Head of the Credit Policy and Underwriting Group within the RMBS Group between 2003 and 2008, as a pervasive and somewhat sinister figure with extensive and guilty knowledge covering every single one of the strands. However Mr Sacco’s evidence was clear as to the limits of his role. The contemporaneous emails depict a man doing his best to hold a realistic balance between the demands of sales and the integrity of the underwriting process. That impression from the documents was only reinforced by Mr Sacco’s oral evidence which, although somewhat affected in terms of the freshness of his recollections by his repeated encounters with the deposition process, was candid, clear and credible.

Another key facet of the case, is that the evidence relied as linking someone to the making of the Underlying Representations either points to no-one or it points to a very large number of people. It has proved impossible for L30 to advance a case which posits just one or two people as involved in the alleged fraud. The case advanced is not one of a few “bad apples”; as noted in the more detailed sections below, the case against a number of people effectively forms a subset of a case against someone else (thus the case against Mr Schoen almost entirely sits within the case against either Mr Sacco or Mr Nordyk). L30 effectively has to pursue Mr Schoen (and Mr Dodman and Ms Baird – none of whom had any exposure to the CDO business) because their names are on the circulation lists of “key” emails relied upon against Mr Sacco and Mr Nordyk. Because of the exiguous nature of the evidence relied upon and its fairly broad circulation the effect of the case is to become one of systematic or wholesale fraud, taking in many people across different desks and departments of CS and even manuals and policy documents issued by CS. And such is the extent of what is alleged that it appears hard to divide those named from those who are not named – the fraud would effectively be entirely pervasive.

To be fair L30 did not shy away from this case. In closing Mr Lord put it thus:

“Messrs Sacco, Nordyk and Sack graphically corroborated the serious and systemic flaws within Credit Suisse’s RMBS business, their awareness of it and the failure by Credit Suisse so take any or certainly any meaningful steps to remedy those failures. The upshot: such serious and systemic misrepresentation of RMBS credit risk to investors and rating agencies to the knowledge of Credit Suisse is beyond serious argument and certainly proved on the balance of probabilities.”

Such a fraud is – of course – not impossible. But again it provides food for thought and question. Ultimately there was no evidence to support a “systemic fraud” case beyond the isolated documents forming the basis of each strand of misconduct and that was manifestly inadequate.

Perhaps a bellweather of this aspect of the case on fraud is given by Mr O’Driscoll. As the executive with direct responsibility for the Notes transaction he was the obvious target. The case commenced with a case in fraud very much maintained against him: it was clearly pleaded that it was to be inferred that he knew of the RMBS Misconduct not because of any particular knowledge, but because it was systemic and because he held a senior position in CS. No document even indicating knowledge on his part was ever pleaded. In cross examination Mr Lord was reduced to asking Mr O’Driscoll to confirm that bits of the DoJ SoF showed that CS had made representations and to posing in the most general terms: “I’m going to suggest to you, Mr O’Driscoll, that you have exaggerated the degree of separation of the CDO and RMBS groups within Credit Suisse at the time … And in those circumstances, …, I suggest it’s likely that you did know about the statement of facts issues or at least some of them by the time of the .. transaction.”

In written closing L30 felt it could not go further than to suggest that the Court “could quite properly form the view that Mr O’Driscoll would have known at least certain of the systemic problems” outlined in the SoF and “may…conclude” that Mr O’Driscoll knew that the CDO Representations were false. L30 also stated in terms that while there were in their view unsatisfactory elements to Mr O’Driscoll’s evidence he was not a key target on the knowledge front.

One other feature of the L30 case was the extent to which it was said that adverse inferences should be drawn from the absence of witnesses. For example: “Due to Mr Daniel’s failure to give evidence, the prima facie case against them [sic] is unanswered and is made out; and/or an adverse inference should be drawn against them [sic] that they [sic] knew of the falsity and re liable for fraud as L30 alleges.”

But adverse inferences are not to be drawn at large. Although in Royal Mail Group Ltd v Efobi [2021] UKSC 33 Lord Leggatt at [41] has emphasised that the drawing of adverse inferences is a matter of common sense, there is some force in terms of analysis of when common sense should kick in, in the discipline urged in Wisniewski v Central Manchester Health Authority [1998] PIQR 324. Applying that test the level of evidence deployed is not such as to raise a prima facie case against the individuals in question, even were it not for the fact that many of them have good reasons for not being present.

In the circumstances the final element of the CDO representations – knowledge of falsity – and the need (or otherwise) to establish an intention to deceive in the context of implied representation - does not arise. At this stage of distance from my findings the process of analysing this point would be entirely artificial and I therefore do not deal with it, although I have alluded to some of the factual issues in the previous section.

In the circumstances I need not deal with some of the remaining issues, in particular the “Alternative Note” issue and the question of the availability of rescission.

Two issues do however remain: Negligence and Conspiracy.

As Mr Field wittily put it in closing – this is not a part of the case where the dispute has raged hottest. L30 relies upon section 14A of the Limitation Act 1980. Under section14A, the limitation period is postponed to the date three years from the date when L30 acquired the requisite knowledge to bring an action for damages in respect of the relevant damage. This includes, under section 14A(8)(a), knowledge that the “damage was attributable in whole or in part to the act or omission which is alleged to constitute negligence”. The relevant “cut-off” date for the purposes of section 14A is 15 November 2015 (being 3 years before the claim was issued). 

L30 submits that even if the CDO Representations were not dishonestly made, they were on any view carelessly made. The questions of what representations were made, reliance and falsity apply equally to the negligence claim.  

It follows that this part of the case fails both at the representation and reliance level for the reasons already given. So too does it fail on falsity. It also, for obvious reasons, fails on limitation, the issue being essentially similar save as to the critical date. 

Since the remaining points only arise on a multiple contingency they can be dealt with very briefly:

There is therefore no need to deal with the case on contractual estoppel.

Originally, L30 relied on several different strands which created the unlawful means: (i) the fraudulent misrepresentations or deceit; (ii) the negligent misrepresentations; (iii) breach of Irish Prospectus (Directive 2003/71/EC) Regulations 2005 (“Irish Regulations”) and Irish Investment Funds, Companies and Miscellaneous Provisions Act 2005 (“IFCMPA”); and (iv) dishonestly procuring the ratings for the Notes. Following a certain amount of judicial encouragement L30 sensibly streamlined its claims in relation to unlawful means conspiracy and Irish law. L30 recognised that (save in relation to Irish law) it does not add materially to L30’s other claims. It is further worth noting that L30 no longer pursues a free-standing cause of action under section41 of IFCMPA, which is the relevant Irish Regulation L30 seeks to rely on to establish its unlawful means conspiracy claim. 

The conspiracy claim was however pursued insofar as it relies upon breach of Irish prospectus disclosure law as the relevant unlawful means. Here it offers L30 a genuinely alternative route to the relief it seeks. It says that even if it fails in its case on misrepresentation (as I have found it does) it still has a good claim in unlawful means conspiracy.  

The essence of the claim is that the Defendants were party to a combination or understanding with each other with an intention to injure or cause financial loss to L30 (as the ultimate purchaser of the Notes) by use of unlawful means, with the object of achieving a sale of the Notes and/or the obtaining on a false basis of credit protection in respect of the reference obligations. The unlawful means consisted of breach of the Irish Regulations and IFCMPA. 

The relevant Irish Law provisions are as follows: 

It is common ground that these provisions do not give rise to any directly actionable liability. 

Two main issues arise here. The first is whether unlawful means can consist of breach of a foreign law. L30 submits that the weight of authority (Bank of Tokyo-Mistubishi v Baskan [2009] EWHC 1276 (Ch) per Briggs J at [932-942]; Digicel v Cable and Wireless [2010] EWHC 774 (Ch)at [22]-[23]; Lebara Mobile Ltd v Lycamobile UK Ltd [2015] EWHC 3318 (Ch) at [41] as well as my own judgment in FM Capital Partners v Marino at [94(2)(a)] favours the conclusion that it can and that there is no principled reason why an unlawful means should be limited to unlawfulness under English law, in particular where the law in question is an EU regime which underpins the relevant liability (with substantially co-extensive liability in English law).  

The second issue is that CS contends that because breach of the Irish Regulations is not actionable, it cannot supply the relevant unlawful means for unlawful means conspiracy. CS relies on Digicel, where Morgan J held that “non-actionable breaches of a non-criminal statute” were not “unlawful acts” for the purposes of conspiracy. L30 relies on BTA Bank v Khrapunov [2018] UKSC 19 [2020] A.C. 727 where at [15] the Supreme Court expressly left open the question whether breach of statutory duty may supply the relevant unlawful means. In any event, L30 says that it does not rely merely upon the (non-actionable) breach of the Irish Regulations; that breach translated into an actionable claim under IFCMPA section41. 

On these two interesting points my conclusions are as follows.

Foreign breaches: on this I would continue to consider that a breach of a foreign law may give rise to a claim in conspiracy. However the potential for “jurisdictional arbitrage” does give pause for thought as to how ready the court should be to go down this route. I also note that the Court of Appeal in Emerald Supplies v British Airways [2016] Bus LR 145 at [132] was sceptical:  

“As we have said, there is an issue whether foreign wrongs can be unlawful means for the purposes of these torts. But if they can, it would mean that the common law could provide remedies for breach of competition laws which those laws themselves had chosen not to confer”.  

Both of these points perhaps suggest that ultimately it may be necessary to have some kind of public policy/comity filter on the breaches of foreign law which may be held to be capable of constituting unlawful means. This would align with the example given by Mr Woolgar in argument – a case where if the conduct had taken place in England, it would be a tort or a crime; for example if there was a French company and a bribery takes place in France and it is thus French criminal law that governs, the bribery being illegal under French criminal law. There is no reason of principle why, that being a crime which is also recognised here, it cannot supply relevant unlawful means. 

On non-actionable breaches, the relevant portion of Khrapunov states as follows: 

“[11] Conspiracy being a tort of primary liability, the question what constitute unlawful means cannot depend on whether their use would give rise to a different cause of action independent of conspiracy. The real test is whether there is a just cause or excuse for combining to use unlawful means. That depends on (i) the nature of the unlawfulness, and (ii) its relationship with the resultant damage to the claimant. This was the position reached by the House of Lords in Revenue and Customs Comrs v Total Network SL [2008] 2 All ER 413, [2008] 1 AC 1174. The Appellate Committee held that a criminal offence could be a sufficient unlawful means for the purpose of the law of conspiracy, provided that it was objectively directed against the claimant, even if the predominant purpose was not to injure him. … 

[15] The reasoning in Total Network leaves open the question how far the same considerations apply to non-criminal acts, such as breaches of civil statutory duties, or torts actionable at the suit of third parties, or breaches of contract or fiduciary duty. These are liable to raise more complex problems. Compliance with the criminal law is a universal obligation. By comparison, legal duties in tort or equity will commonly and contractual duties will always be specific to particular relationships. The character of these relationships may vary widely from case to case. They do not lend themselves so readily to the formulation of a general rule. Breaches of civil statutory duties give rise to yet other difficulties. Their relevance may depend on the purpose of the relevant statutory provision, which may or may not be consistent with its deployment as an element in the tort of conspiracy. For present purposes it is unnecessary to say anything more about unlawful means of these kinds.” 

This does not necessarily provide the clearest possible signposting. In the end, despite Mr Woolgar’s skilful and engaging submissions I conclude that the provision in this case does not fall the right side of the line and that Khrapunov makes no difference. Whilst L30 submitted that the Supreme Court did not expressly state that non-actionable non-criminal statutes cannot constitute unlawful means, they also did not state that they did. The key point here is that Khrapunov does not really deal with the specific issue; while it leaves the door open for a breach of a civil statutory duty to be an unlawful act, it most certainly does not say that Morgan J was wrong in Digicel when he held that a non-actionable breach of a civil statute could not give rise to a claim in conspiracy. This answer also makes perfect sense in the context of the broader themes highlighted by the Supreme Court: although it would appear that the Irish regime is intended in broad terms for the protection of investors, investors cannot themselves action it and they have a separate mechanism for seeking compensation. In those circumstances the remedy in conspiracy is unnecessary and it is not squarely within the sort of duties that the Supreme Court envisaged might be an unlawful means for the purposes of a conspiracy. Accordingly I find that the claim in conspiracy fails. 

Had I concluded otherwise I would then have had to decide whether CS would be liable under either of the Irish Regulations, i.e.:  

As to the former, it is hard to see how the purchase of the Notes could have been “on the faith of” i.e. “in reliance on” either the Series Memorandum (which Mr Schirmer did not regard as a prospectus and which was finalised after the agreement to purchase) or the draft Series Memorandum which was a draft and not a document which should be captured by the legislation. In addition this was not an offering to the public, so on this basis too the transaction would not be caught by the Irish legislation. 

On this basis I do not need to decide the materiality point and will not do so. 

Finally it was contended that the conspiracy claim must also fail given the fact that the elements of combination and intent were not put to CS’s witnesses. I am not persuaded of this argument. In many cases it may be necessary or at least prudent to put these points. However I see no reason why combination cannot be spelled out from documents alone in some cases. Intent to injure is more difficult, but as I noted in FMCP v Marino inferring it from primary facts is possible and (for example) in the cases where intent to injure is the obverse of intent to promote one's own interests, it may well not be necessary to be addressed discretely. There was however force in Mr Field's point that, regardless of the necessity of putting the case to the witnesses, the evidence on intent at least was lacking. 

I therefore conclude that L30’s claim in misrepresentation fails, for the various reasons given above, and that the alternative claim in conspiracy also fails.

The case on RMBS Misconduct has to be considered against the relevant factual background. This is complicated because each RMBS would contain thousands of loans.

The main parts of the background which are necessary to understand the disputes in this area of the case are as follows.

The business model of CS’ Structured Products group on the RMBS side (the “RMBS Group”) during the Relevant Period was to acquire loans through various acquisition channels and then either to package these loans into RMBS, which CS sold to investors, or to sell them as part of a portfolio of loans to other financial institutions through “whole loan sales”.

Originators (i.e. original lenders to residential borrowers) offered their loans for sale, either individually or in “pools”, to CS and transferred the loan files, containing the underlying documentation on the borrower, the loan and the property so that CS could review the loans and conduct due diligence. CS contracted with third-party companies to carry out due diligence on the loans or pool of loans, either individually or on a sampling basis. If CS purchased the loans, they were then held on CSSU’s inventory where they earned interest.

The creation of an RMBS involved (i) a “Sponsor”, i.e. the CS entity which acquired the loans (CS used DLJ Mortgage Capital, Inc (“DLJ”) for this purpose); (ii) a ‘Depositor’, another CS entity which acquired the loans from the Sponsor – CS used CS First Boston Mortgage Acceptance Corp. and CS First Boston Mortgage Securities Corp for this purpose (“CSFB”), and (iii) an RMBS Trust, into which the Depositor transferred the loans.

CS’ business of acquiring (or originating) loans onto its own book was known as the Conduit. The Conduit had four separate “channels”,

CS acquired loans through four acquisition channels, known as (1) the “Bulk Channel”; (2) the “Mini-Bulk Channel”; (3) the loan-by-loan channel (“LBL Channel”); and (4) the “Wholesale Channel”. The Bulk Channel involved the acquisition of large pools of loans originated by third parties which had been underwritten to the originator’s underwriting guidelines and were priced for sale as a package. The Mini-Bulk Channel involved the acquisition of smaller pools of loans generally under $5m originated by third parties. The LBL Channel involved the acquisition of individual loans originated by third parties to CS’ underwriting guidelines that were priced and processed on a loan-by-loan basis. The Wholesale Channel was CS’ own origination platform through which it originated its own loans as a lender.

The word “Conduit” was used in different ways at different times. In particular, sometimes it was used to mean only three of the above channels (excluding the Bulk Channel), whereas at other times it meant the RMBS business, including all four channels.

CS acquired loans in the Relevant Period through two affiliate entities: DLJ for the Bulk, Mini-Bulk and LBL Channels, and CS Financial Corporation (“CSFC”) for the Wholesale Channel.

In respect of loan pools in the Bulk and Mini-Bulk Channels, acquisition was processed by CS “Trading Desks”. Trading Desks sat within “Agency”, “Non-Agency”and “Secondary Credit”. Each Trading Desk kept their own book of business and was responsible for the acquisition of different profiles or characteristics of loans, such as “prime”, “subprime”, “Alt-A”, “jumbo” and “second lien”.

The process of acquiring pools of loans through the Bulk and Mini-Bulk Channels involved sellers sending “collateral tapes” to CS. These contained the loan information of the pool sellers were marketing for sale. The Collateral Analysis Group at CS “cracked” the tape by uploading the loan data into CS’s internal “RPM” system, performing checks on the loan data, populating a pricing model and producing reports for the appropriate Trading Desk, which depended on the loan profile of the package. The Trading Desk then produced a bid price for the loan pool which was offered to the seller.

If the bid was accepted by the seller, then CS carried out due diligence on the loan package. Sometimes sellers imposed stipulations (called “trade stips”) on which the sale of the pool was conditional, such as certain due diligence requirements.

As a result of the due diligence results, which were sent to the seller and Trading Desk, certain loans might have been added to or removed from the loan package. CS and the seller agreed a final ‘settlement tape’ of the pool of loans it would purchase and the Trader made any appropriate adjustment to the payment price for the package.

Through the LBL Channel, approved sellers could access an online CS portal (known as “CWS”) and upload the loan characteristics for loans they expected to close within the next 5 to 60 days. An automated process compared the seller’s input loan characteristics with CS’s acceptable loan parameters and automatically produced to the seller a denial or a purchase price for the loan. If the seller accepted the purchase price, they could “lock” the loan (i.e. confirm it for purchase).

The Wholesale Channel process of CS originating its own loans began when an approved mortgage broker inputted the prospective loan characteristics into CS’s online portal (known as “WHS”). An automated process compared the input loan characteristics with CS’s loan parameters and automatically issued either an initial approval (with or without conditions) or a denial. If the loan was approved, the broker would be instructed to send the full loan file to the Fulfilment Center designated to underwrite CS’s Wholesale loans. If the loan was approved, WHS was updated and the broker could lock the loan.

Other than the Wholesale Channel, where CS was the underwriter, CS performed preacquisition due diligence on loans. CS’s due diligence process varied by acquisition channel:

Due diligence reviews principally involved three main elements: (1) a credit review; (2) a compliance review; and (3) an appraisal review. In the Bulk Channel, loans were allocated an “event level grade” referred to as EV1, EV2 and EV3 in respect of their credit and compliance review.

In the Bulk Channel, as a loan underwent due diligence, data on the loan and the due diligence conducted was added to a spreadsheet and uploaded to CS’s LoanApprove system. The outcome of the due diligence review was to designate a loan as either “eligible” or “ineligible”, but it was open to CS to purchase loans designated “ineligible” in any event. The results of due diligence were therefore considered as being advisory.

In the Mini-Bulk, LBL and Wholesale Channels, a loan would be approved or denied for acquisition. The Fulfilment Centers could raise “exception requests” with the CS Credit and Underwriting Group if a loan failed due diligence but appeared to have relevant compensating factors – these were determined by Mr Salomon, Ms Lombardi and Ms Baird. Once acquired, loans from all channels were held on the CS inventory. CS’s RPM system had a field showing whether a loan held on its inventory was “sourceable” or “unsourceable”, which determined whether a loan was eligible for securitisation. “Sourceable” meant that a loan could be securitised, “unsourceable” meant that it could not.

CS executed securitisations through both (1) principal transactions – i.e., CS acting as the issuer and underwriter of the RMBS using loans from its own inventory, via a CS-affiliated depositor selling into the RMBS Trust, and (2) third party transactions, where CS served as underwriter. L30’s case is limited to the principal transactions.

There are six strands to L30’s case as regards the RMBS Misconduct. The first was in relation to the Sovereign Transaction and the alleged purchase of ineligible loans, L30 criticised and/or accused CS of misconduct on a number of bases (including the acquisition of “EV3” loans and the ‘use of “trade stips”) (“RMBS Misconduct 1: Loan Compliance”).

The second element of the RMBS fraud arises from six examples of bulk pools (or groups of pools) where L30 claims that CS agreed to a trade stipulation which limited the size of the pre-closing sampling which CS would undertake as part of its due diligence process. It is alleged that the existence of these stipulations falsifies a series of Underlying Representations (“RMBS Misconduct 2: Due Diligence”).  

Third, it is accepted that CS represented to investors that all final loan decisions were made by senior CS underwriters, not third-party contractors (the “Approval Representation”). The crucial questions are whether this representation: (i) was materially false; and (ii) were made fraudulently.

Fourth, it is said that CS’s conduct in relation to Fulfilment Centres falsified certain Loan Compliance Representations in that the Fulfilment Centres adopted inadequate processes to assess the quality of the loans and, allegedly, knew of the true value of the loans that they were acquiring and nonetheless approved them for acquisition by L30 (“RMBS Misconduct 4: Fulfilment Centres”). 

Fifth, it is said that CS made a number of representations concerning the LTVs of properties securing the loans it was securitising, the essence of which was that the LTVs published in the Offering Documents were accurate and therefore that none of the loans had an LTV exceeding 100%. L30’s case in this regard is that CS was aware of and concerned about the risk of appraisals being too high and through its DD processes it had learned that the appraisals may be too high and that it had nevertheless applied tolerances and purchased loans within those tolerances and that the court should infer that those loans were securitised and falsified the alleged LTV, <100 LTV and/or Accurate Appraisal Valuation Representation. L30 also relied on alleged manipulation of Broker Price Opinion (“BPO”) values as falsifying these alleged Reprsentations (“RMBS Misconduct 5: LTV”). 

Sixth, it is said CS represented that it had conducted Quality Control (“QC”) reviews, the overall purpose was to approve the quality of the loans that had been securitised. This form of RMBS Misconduct encompasses two separate allegations. First, that the QC findings showed that loans acquired by CS had critical defects, which falsified the Applicable Laws Representation and the Rigorous DD Representation and second that the QC Representation was false because the purpose of QC was not to improve the quality of the loans acquired or securitised (“RMBS Misconduct 6: QC”).

There are also a number of themes running through the case on RMBS Misconduct, which can sensibly be dealt with discretely. 

The first is the making of representations via the presentations. L30, in order to succeed on various of the Underlying Representations, must show that the presentations conveyed the precise meaning alleged by way of representation. In one respect – the Approval Representation – it is conceded that the presentations were clear enough to accomplish this. In other respects however it is contended that the presentations were not so clear, or so final.

Several arguments were raised in the proceedings as to the nature of the words used, both written and then in the oral presentations which accompanied the slides. Aside from the Approval Representation when one stands back from the actual presentations themselves and looks at the context, one can readily see that:

The net result is that – aside from the Approval Representation – where Underlying Representations are alleged to have been made in the presentations, I conclude that those representations were not made.

The status of the SoF was hotly contested. For the reasons already given I have concluded that it does not preclude CS from disputing the making of the bulk of the Underlying Representations.

L30’s case is that there were four Loan Compliance Representations made by CS, the reason that they were false was due to the Sovereign/Countrywide Transaction and by reason of CS acquiring ineligible loans.   In broad terms these representations went to whether the loans complied with applicable laws, or underwriting standards or to CS’s due diligence and/or compliance processes.

The first stage to assess is whether these representations were actually made. L30 relies on the following representations: (i) the Underwriting Standards Representation; (ii) the Applicable Laws Representation; (iii) the Compliance Process Representation; (iv) the Rigorous DD Representation. L30’s case is that the first two Representations derived from the Prospectuses or ProSupps and the second two Representations derived from presentations.  

L30 alleges that the Underwriting Standards and Applicable Laws Representations were made in the Offering Documents for CSMC 2007-3 (specifically the Prospectus and the ProSupp).

In closing the case as to making of these representations L30 rested heavily on supposed admissions by witnesses. As indicated above I do not accept that the evidence relied on is relevant to this point, so this argument can be put to one side.

I conclude that no such representations were made.

As to the Underwriting Standards Representation: 

The Compliance Process Representation and Rigorous DD Representation were said to be made in the Presentations. As already explained these documents were not generally apt to give rise to representations and there is nothing in these representations which takes them outside the norm. More specifically the Compliance Process Representation effectively said no more than that CS had a process to check that loans complied with applicable laws. It did not say (as it would need to) that CS had a process which guaranteed that all loans complied with applicable laws.

Specifically;

Similarly, as regards the Rigorous DD Representation. The representation is said to be that CS had rigorous due diligence processes in place in respect of the loans it securitised. The point made by L30 is that the wording was within Presentation Slides between Q4 2004 and Q1 2007 which highlighted that CS had an industry leading mortgage platform, which achieved a level of loan performance better than its competitors through a “rigorous due diligence process and superior servicing”.

However:

Furthermore, the representations cannot be said to have been false. There are two strands to this. The first is based on the Sovereign/Countrywide Transaction and the second is based on the purchase of ineligible loans.

L30 contends that the Sovereign/Countrywide Transaction compromised the due diligence process through CS agreeing that it would only “kick” loans for fraud and subject to a 5% cap. Yet this was a “one off” transaction. This was due to the fact that Sovereign needed to sell off a very significant part of its inventory to raise capital and had to do so quickly. The fact that this transaction was unique supports CS’ submission that one cannot extrapolate CS’s processes generally based on this one-off transaction. The suggestion that it was not an isolated incident but happened with some frequency is simply unsupported by evidence (eg by way of Loan Level Analysis – an assessment of individual loans, by reference to their individual underlying loan files containing the original loan and due diligence documentation relied on by CS at the time, to establish whether they were in fact in breach of any particular laws or guidelines).

As regards the ineligible loans, L30’s case is based on two examples from the 100,000s of loans purchased by CS, of which one was apparently not purchased by CS and was certainly not securitised. Purchase is only the first step; to have any relevance in this context the ineligible loans would need to be securitised. The allegations that this was routine as part of CS’ business model therefore lack merit. As already indicated the basis for the inference of systemic flaws is lacking. It is not adequate to say “that must be the case because the Department of Justice wouldn’t have procured or wouldn’t have got a $5 billion payment from this bank if these were just a few mistakes made on the way, just a handful of mistakes in two and a half years.”

As regards the knowledge of the falsity, there was nothing significant adduced in evidence that would suggest that any of the RMBS employees were aware of this alleged falsity. Much of the knowledge of falsity case simply hinges on knowledge of the content of the Presentation Slides. The high point appears to be that Mr Nordyk knew that ineligible loans were on occasion purchased. But that goes nowhere without a link to securitisation, and if (as appears from the documents relied on) what Mr Nordyk knew about was the odd occasion – a situation some distance from the allegation of routinely securitising loans CS knew had been rejected.

More specifically:

The relevant representations for this section all concern due diligence. They include the Rigorous DD Representation, which has already been found not to have been made from the above.

The other representations involved were: (i) 100% Subprime; (ii) 100% HEAT; (iii) Variable Sample Size; (iv) Model; and (v) Model Threshold. It is fair to say that L30 did not press these particularly hard noting in closing that “there may be some scope to debate some of the Schedule 2 Representations”, only picking out the Rigorous DD Representation as being one about which there should be no dispute.

Each of these representations are said to have been made in the Presentation Slides. For the reasons already given I conclude that no such representations were made. More particularly:

As to falsity, L30 claims that there were pools of entirely subprime loans, which were not subject to 100% due diligence, which falsified the 100% Subprime Representation. L30 further claims that because of the fact that the loans in these pools were securitised into HEAT RMBS, it inherently falsified the 100% HEAT Representation.

As regards the 100% Subprime Representation, Mr Sacco gave evidence that there was an oral qualification to investors regarding large pools. Based on the evidence that was shown to this court, subprime pools were generally subject to 100% DD, with a small number of exceptions for very large pools from large originators to which Mr Sacco gave an oral qualification. Thus, what was actually stated to investors was substantially true.

As regards the 100% HEAT Representation, it was true. The only pleaded instance on this representation being made was in June 2005, with the latest HEAT deal having been HEAT 2005-4 (HEAT 2005-5 closed on 1 July 2005. According to the Agreed Structured Data Guide (“ASDG”), 100% of the loans in each of HEAT 2005-1, 2005-2 and 2005-3 were diligenced, and 99% of the loans in HEAT 2005-4 were diligenced. There was a raft of communications by CS with investors and rating agencies which pointed out the pools exempt from its 100% DD policy.

Otherwise as regards the Due Diligence Representations, it is alleged that these have been falsified due to a series of “trade stipulations” which limited DD in some way. The case L30 had to prove was whether this falsified what was said to investors or rating agencies about its DD processes. But the thrust of the evidence was that such trade stipulations were rare/very unusual (the suggestion being maybe half a dozen trades), and where the trade stipulations were present, there was nothing wrong with them, nor were they against what CS had actually told investors. They more generally limited “kicks” than sample size and where they did they did not change the way the DD was conducted on the loans that were sampled.

As regards the Model/Model Threshold Representations, the communications made (which cannot sensibly be read to say more than that CS used the model to help select the sample) were true.

As regards the knowledge of the falsity of the representations, the only place where this argument might gain traction would be with the 100% representations, where falsity and knowledge might be said to go hand in hand. The case on knowledge here hinges on establishing the full breadth of the 100% representations (i.e. that there were absolutely no exceptions). Nothing else will do for L30. There is no case on knowledge which can survive that. But as those representations were not even strongly pressed, that is neither here nor there.

In relation to the main target, the Rigorous DD Representation there is no adequate case that those accused of knowledge had knowledge of falsity. The headline point is that there was no motive adduced by L30 which would suggest a reason why Messrs Sacco or Nordyk would lie.

At a more granular level:

In relation to the Approval Representation, it is agreed between the parties that a statement in terms of the Approval Representation was made in the Presentation Slides. The Approval Representations come out of a single bullet point in a series of Presentation Slides.

The unique element of this representation is that following the DoJ SoF, CS essentially admits falsity. CS admits these to be false not because of the clarity of meaning of these statements, but because in the SoF it agreed: “Credit Suisse represented to investors, ratings agencies, and others, that “Credit Suisse senior underwriters make final loan decisions, not contracted due diligence firms.” For Conduit loans, these representations were false.”

It is therefore accepted that objectively those representations meant: all final loan decisions were made by CS senior underwriters, not third party contractors.

The question of whether the representation was materially false remained live.

What would certainly be material would be if decisions were made by FCs if there was any judgment to be exercised. On this the evidence given in court highlights the limited ambit of the FCs’ remit and the ample scope for those reading the slide to have interpreted it differently. The reality was that the FCs had no authority to approve loans outside guidelines/matrices which required an “exception”. They could turn down and refer risks within their guidelines; but they could not approve risks outside their guidelines. Where judgment had to be exercised it is common ground that such a loan had to be escalated to CS underwriters. They therefore operated within an underwriting “cage” with essentially no scope for independent decision making. This was reflected in the evidence, in particular Mr O’Driscoll's evidence was that the fact a loan went through an FC would not necessarily mean it was a “good or bad thing”, nor did Mr Kaiserman take that to be the case. In terms of what one might term content materiality therefore, the misrepresentation was unlikely to be material.

Materiality overall turns not just on content but also on the sheer statistical fact that it appears that over 90% of the loan decisions were made by the FCs. I have considered very carefully whether this must lead to the conclusion that the representation was materially false. I am ultimately persuaded that it was materially false.

This is not a straightforward decision, given that where it was false the FCs were making the decision precisely because the decision was formulaic and the risk fell within prescribed parameters. It was, in essence, a decision which was no more than a box ticking exercise. The problem essentially lies in the mistakes which demonstrably were made by the FCs. It was such errors which caused the discussions which are the subject of the next allegation of misconduct. Although in the abstract one might say it was a box ticking exercise, it was not a straightforward box ticking exercise. As CS said in closing, dealing with the correspondence on FCs: “the FCs were performing a very complicated and difficult job with real scope for human error.” And the proof of the pudding was in the eating: there were mistakes, and not just of one sort. If there was scope for such a range of mistakes by FCs, it seems to follow that fact that the percentage of decisions being taken by the FCs was so high is a fact which would be material.

The question of deceit is therefore critical here. It is worth rehearsing the target at which L30 must aim. It is that set out by Lord Hughes in Ivey v Genting Casinos [2017] UKSC 67, [2017] 3 WLR 1212 [74]:

“When dishonesty is in question the fact-finding tribunal must first ascertain (subjectively) the actual state of the individual’s knowledge or belief as to the facts. The reasonableness or otherwise of his belief is a matter of evidence (often in practice determinative) going to whether he held the belief, but it is not an additional requirement that his belief must be reasonable; the question is whether it is genuinely held. When once his actual state of mind as to knowledge or belief as to facts is established, the question whether his conduct was honest or dishonest is to be determined by the fact-finder by applying the (objective) standards of ordinary decent people. There is no requirement that the defendant must appreciate that what he has done is, by those standards, dishonest.”

One must then consider what the words on the various slides could reasonably be thought to mean, in particular whether, although objectively bearing the agreed meaning, they were ambiguous such that the individuals at CS who were said to have acted dishonestly must be taken to have subjectively understood them to say that which was untrue.

On this I am satisfied that the words used in the presentations were capable of more than one interpretation:

As regards the knowledge of the falsity, this was pleaded against a range of people. The headline figures are probably Mr Sacco (primarily as L30 submitted) and Mr Othman. As to them:

The evidence of the other lesser players not only cannot make good this deficiency; also the elements where the evidence is the same points up the frailty of the evidence even against the so called key players:

Consequently the case on knowledge of falsity fails.

This strand was based around the proposition that the FCs and their role falsified certain of the representations which also figure in other strands.

The Underlying Representations relevant for this stage are: (i) the Underwriting Standards Representation; (ii) the Applicable Laws Representation; (iii) the Compliance Process Representation; and (iv) the Rigorous DD Representation. All of these have been considered and found not to have been made. This point therefore fails at the first hurdle.

Nevertheless, as regards the falsity element, L30’s case is built around a number of different types of document in which criticism was made of the FCs. In particular L30 relies upon:

The Affidavits of Diane Johnson (of Ocwen) and Ron Szukala (of Lydian) (deriving from the MBIA Proceedings) were originally relied upon heavily by L30 but formed no real part of the case as closed. Although not technically dropped (as Mr Lord made clear) they were not pressed. There was a good reason for that, which was that they did not really cohere with L30’s own case. L30’s case was predicated on lamentable performance by the FCs, known to CS. Perhaps unsurprisingly Messrs Johnson and Szukala, who worked for the FC told a tale which had the FCs being beyond reproach. Certainly to the extent that these statements were relied on I did not find them persuasive, given the fact that they were not explored at trial and the apparent tension between them and documents which I did see.

L30’s case in this regard is that these communications highlight the systemic problems at FCs. In short, their complaint lies in the alleged fact that CS took no action to fix these issues and this thus rendered various Underlying Representations false.

The problem with this is that taken in any form of context these are isolated documents. On the whole they are documents authored by individuals who had limited involvement in due diligence and where the passages relied upon form one part of a wider (and often different) context. Nor do the communications show that there were systemic problems, but rather, as alleged by CS, “human error” within a business where the volume of loans being considered meant that it was completely inevitable that things would go wrong. So, for example – the Audit Report did give the RMBS and Conduit and Wholesale Channel a rating of C4 (almost as bad as it could be) but there was no notation of Significant Reputational Risk Issues and the issue (lack of documentation) was noted as “Agreed and completed”. Ms Lombardi was conducting a system review of an FC centre. QC reviews were looking at QC. Mr Vibert was a trader.

Furthermore, the allegation by L30 that CS did nothing to “fix” the matters identified and were unbothered by them goes against the documentary evidence that was produced at trial. For example, the very correspondence relied on repeatedly indicates a concern to do things better. The colourful language which L30 highlights as indicating a problem also indicates concern – even when the problems are not systemic. See for example Mr Othman’s “we need to start doing a better job… it is imperative that you take this email VERY seriously” in the context of stated income loans. Or one might highlight Mr Vibert’s “fulfillment process is broken” email which is plainly designed to seek improvements, and proposes some (albeit based on limited understanding of the underwriting process). There was plainly a spreadsheet which scored underwriters performances and from which underperformers were identified. It was clear that CS had produced training slides which highlighted “Recent Findings – QC, Whole Loan Sales, Repurchase”.

There is also a problem with the attempted “read across” from a loan being rejected in a whole loan sale to breach of loan level representations in the PSA. It does not necessarily follow given that different standards would be applied; this is an “apples and oranges” comparison. To ascertain whether the read across was valid one would need to look at each loan individually to identify the reason for the “kick”: and thus whether that kick was a measured response and whether that should have been picked up by the FC at the time of their review.

As regards the knowledge on the part of the falsity, there was no knowledge of the RMBS employees of the alleged falsity. A wide range of people are said to have the relevant knowledge: each of Messrs Kimura, Marriott, Fallacara, Sacco, Sack, Othman, Salomon, Daniel and Vibert. However the evidence does not bear out the point for any of them. Some of them had very limited involvement in DD, though others had more. Again for some of those who were alleged to be fraudulent much hinges on the inadequate base of the Presentation Slides.

The most obvious targets here (other than Mr Sacco) are Mr Othman and Mr Vibert – because they were the authors of some of the liveliest emails. As to them:

As to the other individuals:

This element of this misconduct concerns the manner in which CS reported the valuations it received in respect of the properties underlying the loans it securitised. It hinges on four separate representations: (i) the LTV Representation; (ii) <100% LTV Representation; (iii) the Accurate Appraisal Value Representation; and (iv) the Rigorous DD Representation. The LTV Representations were said to be made in the Offering Documents by reason of the inclusion in Offering Documents of tables of LTVs for the relevant mortgage loans for that securitisation.

This element of the case was dealt with for CS by Mr Brock whose confident and clear presentation brought much needed focus to an argument whose complexities had swept by at some speed in the course of trial.

I can start by noting that the Rigorous DD Representation has been found not to have been made.

As regards the LTV Representation and the <100% LTV Representation, CS submits that these representations are irreconcilable with the actual words that have been used in the Prospectus/ProSupp, which states in terms that:

I accept the submission that on any fair reading of these documents the statements as to value and composition were not representations of fact but statements as to the expected make-up of the trust. The statement as to actual value was within the PSA which represented that the information in Schedule 1 (the “Mortgage Loan Schedule” or “MLS”), a spreadsheet containing details, including LTV, of each loan actually transferred, was “complete, true and correct in all material respects”.

It follows that the LTV Representations were not made.

As regards the Accurate Appraisal Value Representation, this again too was not made. L30 attempts to make an argument that the Accurate Appraisal Value Representation concerned the accuracy of the appraisal at the time it took place, but this is even more difficult when the steps taken by CS as part of valuation due diligence were not geared to value at the time of origination. In order to find an implied representation it needs to follow (as L30 accepts) as a matter of necessary implication from the reporting of LTV values to investors. But since the LTV values were (as set out above) not reported to investors qua values, but simply as an indication of expected (and subject to change) composition, it is impossible that the Accurate Appraisal Value Representation can be a necessary implication. The L30 case on this is made still worse by the fact that the implication is one which has to be into the Prospectus/ProSupp, which was as already noted a detailed document into which had doubtless gone days if not weeks of careful legal draftsmanship.

That conclusion is consistent with the fact that it took L30 some considerable time to discern and plead this apparently necessary and obvious implied representation. This argument is not saved for L30 by the submission that Mr Sacco and Mr Nordyk accepted that such a representation was made. Those “acceptances” were (as noted elsewhere) not matters on which the witnesses’ answers can be determinative. They were also extracted by a process of putting to them points in a generic way, out of full context – and even then the answers were somewhat hedged. So Mr Nordyk says “if we are putting forth an LTV…”

As regards falsity of the representations via appraisal inflation (the awareness of the risk of overhigh appraisals combined with the application of “tolerances”), and BPO manipulation L30 puts forward the argument that the internal communications in CS showed that the appraisals were significantly inflated, as well as that the Structured Data showed that BPO values were being manipulated to comply with CS’s permitted variances. There are a number of problems with this. The first on appraisal inflation is that L30 does not allege that the LTV ratios in the LTV tables were calculated wrongly according to the definition in the ProSupps. L30’s revised case is that LTVs meant something else via the Accurate Appraisal Value Representation. This is not a convincing argument, but even on its own terms it hits problems. The difficulties with this argument are (i) that CS had in place DD processes in order to mitigate the risk that the property’s appraised value could not be reasonably supported and (ii) the purpose of the DD was to test the reasonableness of the original appraisal, not to test the value of the property at the time of the DD. The DD had a different function to the one which L30 alleges. Further these claims proceed from the false predicate that BPOs were more accurate than appraisals (when – as Mr Brock illustrated - they were “drive by” exterior valuations not performed to the same standards required of an appraisal - and so likely to be less accurate).

As regards falsity via the alleged BPO manipulation these allegations are a triumph of form over substance and are in any event unsupported by contemporaneous documentation, being based entirely upon the Structured Data. Even the structure of the allegation is lacking in credibility: the alleged form of the manipulation is one carried out by hundreds of BPO agents (i.e. not CS employees), who had no (even suggested) motivation to carry out such a manipulation. The case on “primary” manipulation was clearly inconsistent with the documents which showed that the BPO agents were not aware of the original appraisal value. The case on secondary manipulation ignored the facts – manifest in the documentation, as Mr Brock demonstrated - that changes in BPOs were made following the broker having sight of the (more thorough but earlier) appraisal, which included information they would not have had. This was a point with which L30 failed to grapple. It also ignored the further fact – explained by Messrs Nordyk and Sacco - that CS did not need such manipulation to accept loans outside CS’ tolerances’ because a due diligence manager in CS could have accepted them as an exercise of discretion.

L30’s case in opening on this very serious allegation was that that the “numerical facts are so striking that the only sustainable inference is that the BPO values were being manipulated to fall within tolerance, and thereby enable CS to acquire and securitise those loans.” In closing the case was maintained contending that “the skewing of BPO results to the particular variances is evidence that BPOs were being manipulated to comply with those variances. The tables of data speak for themselves, and the Court will have seen from those tables the remarkable drop-off precisely around the applicable variances (0.85 and 0.9) at the relevant times.” This was posed as a matter of pure assertion – with no expert evidence sought or offered to assist on how the data had been dealt with and whether the inference to be drawn was statistically permissible, let alone inevitable. Mr Brock’s submissions indicated likely fallacies in the approach adopted by L30 – most notably (but not by any means solely) –selection bias. Those issues had been pleaded by CS but were never grappled with by L30, save to assert that there was nothing wrong with relying only on those pools which it sees as exemplifying the fraud. That does not, in my judgment, offer any real answer. A case in fraud based solely on statistical evidence would require a far more thorough grounding.

Finally the case on knowledge/fraud was again lacking. There was in effect no evidence which suggested that CS knew of the manipulation – not least because the case was entirely based on L30’s own workings without any opportunity to look at the figures or put them in context. For many of those involved they would have had no link with the Bulk Channel such that they could be thought to have any knowledge of any manipulation (eg Messrs Salomon and Othman and Ms Baird).

Similarly, there was no knowledge of the alleged falsity of these statements by any of the RMBS employees.

L30 relied heavily on a number of emails as evidencing knowledge of falsity. For example:

But these emails were yet more emails addressing the points before the eyes of those involved – i.e. making the system work. They provide no evidence of knowledge of falsity and fraud. On the contrary they appear to show a robust and honest approach to stress testing and where necessary improving systems.

This final strand of the alleged misconduct which relates to QC findings and quality control generally hinges on: (i) the Applicable Laws Representation; (ii) the Rigorous DD Representation; and (iii) the QC Representation.

The Applicable Laws and Rigorous DD Representations have already been considered at earlier facets of the RMBS Misconduct and found not to have been made.

The QC Representation alleged is: “the Purpose of CS’s QC Process was to increase the quality of the loans it securitised”.

However this representation cannot be made out. First, the representations were alleged to have been made from Q4 2006 and were said to derive from the “Post Close Quality Control” Slide, which states:

“Objective – Increase the Quality of loans originated/acquired by testing loans diligence/not diligence”.

In other words the base document for the representation refers to the quality of the loans that had been originated/acquired (not securitised). In addition, L30’s case is that the bullet point alone encapsulates all elements of QC. However when taking the whole slide, in its proper context, it can be seen to refer to the “random” regular monthly 3-5% element of QC and not other (distinct) types of performance based QCs or targeted/ad hoc QC or 100% Wholesale Channel sub-prime QC. The documents demonstrate not just that this statement referred to one part of QC only but also that other elements of QC had different roles to play. So, for example, performance based QC was used to identify loans that might exhibit a breach of the representations and warranties made by a seller to Credit Suisse in a MPLA.

It follows that on no sensible reading of the material relied upon can a representation to the effect alleged be discerned.

It follows that RMBS Misconduct 6 fails at the level of making of the alleged representations.

As regards falsity, L30 alleges two types of falsity. The first is that QC findings showed that the Applicable Laws and Rigorous DD Representations were false. The issue here is that the L30 case assumes that a “critical finding” in a QC report necessarily means that there was a breach of either or both of the Applicable Laws Representation or the Rigorous DD Representation. Quite aside from the obvious point (not really addressed by L30) that there has to be room for differences in judgment in an evaluative process, there are more hard edged flaws. The QC and DD or originator guidelines were different and the QC process occurred later with later occurring information as part of the inputs. As Mr Sack said: “what I had in mind, is if you do a third party QC and they say: we disagree with the credit decision, that's not actually an analysis of a rep and warrant repurchase obligation” – and this led him to seek changes to the QC system which would enable CS to put back loans using the QC finding.

The second element of falsity is based on the changes to the QC thereafter, which is said to have falsified the QC Representation when made. There are two points to make here. The first is that again L30 wrongly elides the different forms of QC process; the only type of QC that actually underwent any form of change was the performance-based element of QC, not the random element of QC. Secondly the material used by L30 to support its case is not material which goes to actual change, but only to discussions of potential changes. This could not therefore falsify anything. And thirdly in context it would appear that what was being debated was the possibility of using QC for extra purposes, not cutting it back or dispensing with it.

Thus this element of RMBS Misconduct fails at the second hurdle, as well as at the first.

Similarly to the other strands of misconduct, although fraud was alleged against a wide spectrum of individuals (Messrs Sacco, Nordyk, Sack, Kaiserman, Daniel, Salomon, Schoen, Fallacara, Othman and Vibert) there was nothing adduced in evidence which suggested any knowledge of the alleged falsity by the RMBS employees.

APPENDIX 2A: CDO REPRESENTATIONS (PLEADED)

APPENDIX 2B: CDO REPRESENTATIONS (REVISED FOR TRIAL)

APPENDIX 3: RMBS MISCONDUCT ALLEGATIONS