AP-Fonden v Bank of New York Mellon SA/NV & Ors

A central premise of AP’s case is that the Sigma medium term notes acquired for AP’s account in 2007 were inherently unsuitable investments given its conservative risk appetite. As BNYM says, that in turn is largely premised upon the nature of Sigma itself. There was considerable evidence in that regard, from the parties’ experts, and from BNYM’s factual witnesses. Very different conclusions are drawn from it by the parties.

What is common ground is that Sigma was very similar to other entities commonly known as SIVs (or Structured Investment Vehicles), and that although Sigma considered itself to be a “limited purpose finance company”, it was considered by investors and analysts as part of the SIV sector. Some of these were “sponsored” by banks, some (like Sigma) were not. According to Mr Richard, as a result of increased demand for SIV securities by investment managers, SIV debt issuance experienced tremendous growth during 2005 and 2006, increasing 45% in 2006 alone. The evidence is that SIV assets were nearly US$300bn as of year-end 2006, and were projected to continue to rise at similar rates during 2007. The figure given by AP is US$400bn.

Sigma Finance Corporation was incorporated on 5 April 1994 in the Cayman Islands. The founders were two individuals, Mr Nicholas Sossidis and Mr Stephen Partridge-Hicks, who had previously (as Ms Demmler put it) “invented” the SIV when at Citigroup. They managed Sigma through a London-based English company called Gordian Knot Ltd, which as investment manager was regulated by the FSA.

Of all the SIVs, Sigma was the biggest issuer, with issued debt peaking at US$55bn in 2007. In July 2007, its debt had a weighted average life of 1.16 years. This implies a huge rate of issue, showing, BNYM says, that by July 2007 Sigma was issuing (and investors were purchasing) debt at a rate of almost US$1bn per week on average.

Stated simply, its business model was to issue securities in the form of short-term debt, and use the proceeds to buy higher-yielding long-term debt, profiting from the difference. This is aptly called “maturity arbitrage” in some of the material. The securities which Sigma issued were in the form of commercial paper (“CP”) and medium term notes (“MTNs”). There were a considerable number of such issues, with varying maturities. CP maturities were less than a year, whereas MTNs had maturities of up to five years, with several years being typical, according to Mr Servis. However, although BNYM has compared Sigma’s business model to that of a bank, it was not in my view seriously comparable.

The MTNs in question were issued by Sigma’s wholly owned Delaware incorporated company, Sigma Finance, Inc, pursuant to a Private Placement Memorandum dated November 1, 2005. This allowed for issues up to US$50bn of medium term notes (MTNs) with maturities in excess of 300 days from date of issue. The notes were guaranteed by Sigma Finance Corporation. There were relatively complex governing law provisions which are not relevant for present purposes. The notes were issued in book-entry form through The Depository Trust Company (DTC). The notes were floating rate notes (a floating rate note is classified as a “fixed income security” because the interest rate is reset at specified intervals, so that at any given date it is a fixed rate).

As explained by Mr Servis, and not in dispute, under the market conditions that prevailed for the first twelve years of its existence, Sigma was able to fund itself because investors typically rolled over its debt into new issues. What this meant was that when Sigma’s MTNs reached maturity, repayment was funded by issuing further MTNs, rather than by selling assets in its portfolio.

Under the terms of the Private Placement Memorandum, the holders of MTNs had security rights by way of a “first priority floating lien” over Sigma’s asset portfolio, which provided a reassurance of its ability to repay. The MTN holders had senior status by virtue of the subordination of the capital note holders in the security structure. This provided a capital cushion of some US$4bn, as against debt outstanding (according to a table produced by BNYM in closing) of about US$53.4bn in August 2007, and about US$24.3bn in August 2008. However, importantly, the memorandum stipulates that assets used as collateral in repurchase agreements (repos) took priority. There is a dispute between the parties as to the practical effect when repo transactions became a prime source of funding for Sigma as the financial crisis deepened.

Probably the most important reassurance for investors, however, was not the lien over the assets. It came from the credit rating agencies, whose endorsement was crucial to Sigma’s business model. Sigma’s senior debt was AAA rated by all major credit ratings agencies from its inception until April 2008. This includes the time in March/April 2007 when the Sigma MTNs were purchased for AP. The AAA rating was also essential in keeping the funding costs of the securities low, and thereby enabled the arbitrage that made the whole structure profitable.

In common with other securities lending agents, BNYM maintained an “Approved List” of issuers considered suitable for cash collateral investments. Only issuers on the Approved List could be purchased as collateral for securities lending clients. Sigma was added to the Approved List of Mellon Bank on 19 August 2005 (this being pre-merger), on the basis of a review conducted by Ms Demmler.

AP’s case is that the review was deficient, and Ms Demmler did not understand the complexities inherent in Sigma MTNs as securities. Among the points taken, AP says that her review was brief and outline in nature, that there was no legal review of Sigma’s documentation, and that Ms Demmler was not aware that the securities were “Non-standard Investments” under an internal policy document. Criticism is also made of the fact that she did not mention in the review that the Sigma MTNs were “Rule 144A Restricted Securities” or that they were “Private Placements”.

AP accepts that Ms Demmler did her best to assist the court, which is correct. It goes on to submit however that her evidence as to what she did and, more significantly, what she did not do, during the period from approval onwards revealed a “shockingly imprudent approach to credit analysis and a complete failure of reasonable diligence and competence in the performance of her duties as credit research manager and analyst in relation to the Sigma MTNs”.

Having seen her give evidence over a day, I reject those criticisms. Ms Demmler began her career at Manufacturers Hanover, which became Chemical Bank and was eventually absorbed by JP Morgan. She joined Mellon Bank in 1995, and by 2005 was very experienced in her role. Though the Sigma securities benefited from a AAA rating, she said (and I accept) that she never relied solely on ratings.

Also, and importantly, Mr Tant reviewed and approved Ms Demmler’s analysis of Sigma. He said that the aim of the Approved List was to ensure that clients were invested only in high quality assets. While client investment guidelines would include rating criteria, the bank’s approach was to individually research securities. He said that, “This reflected the fact that Securities Lending and other cash investment strategies were intended to represent relatively low-risk investments in which preservation of principal was a priority. That said, any investment programme carries with it some degree of risk”. Mr Tant trained as a credit officer in 1983, when he joined Mellon Bank, and in 2004 he was appointed the Chief Reinvestment Officer for Mellon’s securities lending group. He was a thoughtful witness.

In view of the criticisms made by AP of Ms Demmler’s alleged lack of understanding of repos, it is to be noted that the review states that MTNs were secured over Sigma’s assets “subject to carve out for repos … The only priority claim ahead of Senior Notes are security interests granted under repurchase agreements”. This was a correct summary of the position. Further, AP’s criticisms of her performance need to be seen against the evidence of how widely Sigma was held as collateral by securities lenders.

The adequacy, or otherwise, of Ms Demmler’s review is ultimately a matter of expert evidence. Though Mr Servis described her analysis as “short and simplistic”, he has no relevant experience of credit analysis in this context. BNYM’s expert, Mr Richard, on the other hand, does have more relevant experience. He regarded the review as detailed, pointing out that it was supplemented by information from the credit rating agencies and direct communication with the managers, Gordian Knot.

His conclusion was that, “The detailed review shows that by identifying key risks as well as mitigating factors, Ms Demmler had considered the same types of issues that I would have focused on in performing a similar analysis. Based on the nature of this review, it is my opinion that BNY Mellon acted reasonably in evaluating Sigma for inclusion on its Approved List”.

I prefer the evidence of Mr Richard to that of Mr Servis in this respect. It may be correct, as AP says, that Ms Demmler did not fully appreciate the complexities inherent in the Sigma MTNs, which the crisis was in due course to lay bare. However in my view, this is a hindsight criticism. As I explain later, the evidence is that most of the investors in SIV senior debt were money market funds, which BNYM describes as “conservative investors” (and I consider that they were perceived as such), with most of the rest being held in securities lending cash collateral pools.

It was Ms Demmler who came up with the term “money good”. She said in cross-examination, “My role was, you know, look at the assets and determine if we thought they were essentially what we call "money good". Were they going to pay out at maturity”. The phrase was much cited by AP in its closing submissions. AP seeks to avoid the obvious conclusion that Ms Demmler was justified at the time in regarding Sigma securities as “money good” (as did the market generally) by attacking her analysis. No valid criticism can, in my view, be made of her review. I reject this ground of AP’s challenge to the acquisition of the securities.

The evidence is that there was a marked increase in issuance in the SIV sector over the period 2006 and early 2007. In retrospect, this was the peak of the bubble that had developed for asset-backed securities. The so-called credit crunch is usually taken to have begun in mid 2007, in June, July or at the latest August 2007, developing into a fully blown global financial crisis in September/October 2008. It is not seriously in dispute that from the onset of the credit crunch there was an attendant drying up of demand for the securities issued by the SIVs, which meant that their business model became unviable, though this was not immediately recognised at the time.

BNYM relies on a report issued by Moody’s on 20 July 2007 entitled, “SIVs: An Oasis of Calm in the Sub-Prime Maelstrom”. This report was one of a mass of reports and commentaries adduced by the parties in which rating agencies, banks, investment advisers, the financial press and others expressed opinions, to which the parties referred to substantiate their respective cases. Both parties found ample ammunition in the material. While the articles can be good indicators of contemporary market sentiment, and helpful in charting the course of the crisis, it would not be appropriate, in my view, to give undue weight to any particular opinion. This material was mainly relevant so far as it shed light on the primary evidence.

An example can be given as follows. In its closing, AP cited a Wachovia Securities article dated December 18, 2007. This states that “… with a ~67% spike in June MTN issuance at Sigma Finance, the largest SIV on the street, we may have seen an attempt at shoring up its position in anticipation of what occurred in the latter half of the year”. The AP Sigma MTNs were acquired in March/April 2007, so this is of doubtful relevance, even if correct. Further, the article also provides support to BNYM’s case, saying that, “Many in the sector likely anticipated 2007 to be a record year for SIVs, and as the troubles in the mortgage and hedge fund markets intensified during July and August, both Moody’s and Standard and Poor’s published articles praising the performance of SIVs and their relative insulation from the troubles of the broader credit markets”.

This optimism proved to be hopelessly wide of the mark. The favourable reporting on SIVs did not last. AP quotes an article published on 20 September 2007 by the credit rating agency Fitch, saying that, “The SIVs ability to secure ongoing funding is core to its business model, which centres on maturity arbitrage by financing long-term assets to a large degree with short-term liabilities. … Continuous access to funding via the CP and MTN markets is critical for SIVs to maintain normal operations”.

That funding was becoming unavailable. In short, the market for the product dried up. Further, the assets held in the SIVs’ portfolios were suffering with the rest of the market. SIVs began to go into “enforcement”, which meant that they had failed their market value test, and were wound up, beginning with Cheyne Finance on 28 August 2007. BNYM points out that the initial view of the receivers was that its assets were not exceeded in value by its liabilities (Re Cheyne Finance plc [2007] EWHC 2116 (Ch), at [4], Briggs J), but this does not affect the overall picture.

An internal Standish email of 28 August 2007 says that, “Today's news is as close to a death knell as one can get for SIVs. The well-managed ones are going to suffer funding problems no matter what, after Cheyne”. This was an accurate appraisal.

The last securities which Sigma itself was able to issue were issued in October 2007. This is highly significant, since its business was the issue of securities. From that time, barring a revival in the market, Sigma was (to use the phrase used by one of the witnesses) in “run off”. However it should not be assumed that this was an orderly process. One way or another, Sigma had to use its asset portfolio to meet maturing liabilities. It did so largely by raising funds through repos, that is, loans collateralised on its portfolio, and what have been called ratio trades, by which it redeemed notes at par against the purchase of its assets at a ratio (usually) of three to one. This was not a process which could go on indefinitely. An important question in the case is as to the extent to which BNYM nevertheless still regarded Sigma as essentially sound, in the sense that its assets, capital and ability to continue to raise money were such that (as BNYM contends) it reasonably concluded that noteholders were likely to be paid on maturity.

In early April 2008, Sigma was downgraded below AAA, though it was still investment grade paper. (According to BNYM, the US$5.5m tranche maturing 12 March 2009, which was less than a year from maturity, continued to have the high ratings.) It was this downgrade, and more particularly the drop in the indicative price being provided to the bank by its pricing vendors, that led to clients (including AP) being contacted by BNYM in May 2008. The adequacy of such contact raises another important question in the case.

At the beginning of September 2008, Sigma was still rated A by Standard & Poor’s (“S&P”), and A3 by Moody’s. BNYM relies on S&P’s report of 12 September 2008, to the effect that as of 10 September 2008 Sigma had assets of over US$27.5bn to meet US$23.5bn of secured liabilities, including US$17.3bn in repos and US$6.2bn of MTNs. This is discussed further below, but notwithstanding passages from the reports of the agencies, I am satisfied that by now Sigma was hugely vulnerable.

The bankruptcy of Lehman Brothers on 15 September 2008 precipitated a drop in the value of many of the investments that made up Sigma’s portfolio. That, in turn, led to increased margin calls from repo lenders. On 30 September 2008, Sigma defaulted, and its securities were downgraded by the rating agencies to junk status. Receivers were appointed on 6 October 2008.

BNYM calculates that on 6 October 2008, some 97% of the portfolio (just under US$27bn) had been seized by repo counterparties, leaving assets of US$700m to meet remaining MTN liabilities of US$6.2bn comprising outstanding MTNs, including AP’s (Re Sigma Finance Corp [2008] EWHC 2997 (Ch) at [7]).

There subsequently ensued litigation over the correct interpretation of Sigma’s security trust deed. There was a difference of opinion in the courts as to priority of payment, and BNYM says that by the time the Supreme Court (by a majority) allowed the appeal from the courts below, the assets had “erroneously” already been sold at fire-sale prices for just US$450m. In any event, there was bound to be and was a large shortfall.

Having set the scene factually, it is now necessary to examine the legal relationship between the parties, which is based on the SLAA entered into between them.

Prior to its relationship with BNYM, AP began a securities lending programme in March 2002 with Citibank, which was at that time AP’s custodian. As Ms Hammarlund explains, a positive decision was made not to allow Citibank to invest any of the cash collateral. In 2004, consonant with its obligations as a public authority to ensure that it was receiving a competitive service, AP decided to tender for a new custodian relationship. The focus of the procurement was on custody services, but AP was also looking for the custodian bank to provide securities lending services. Having now had some experience of the way securities lending worked, Ms Hammarlund made it clear that AP had also decided that it now wanted a cash collateral programme.

On 18 February 2004, AP issued a notice inviting tenders for “global custody and related services”. There were four preferred bidders, each leading securities lenders, including ABN AMRO Mellon Global Securities Services B.V. This company was originally a joint venture between ABN AMRO and Mellon Bank, formed to enable Mellon to market its services to clients outside the US. BNYM is the legal successor to that company, and for convenience I shall continue to use the term “BNYM” to describe it.

In its responses, BNYM addressed various issues raised by AP, and provided information about its pooled funds for cash collateral investments, and investment parameters, as well as the possibility of a segregated account. It stated that “our strategy for collateral investment centers on preservation of principal and maintenance of proper liquidity by managing interest rate risk”. I agree with AP that this was a clear statement of how BNYM would approach its task if selected.

A number of meetings and exchanges took place between AP and the tenderers, including BNYM. On 30 July 2004, BNYM sent a response to various questions which had been put to it by AP. It explained that its Global Securities Lending Asset Review Committee met bi-weekly to review issuers as well as reinvestment issues. It explained that credit research was the function of the bank’s Compliance Risk Management Group. It explained the credit review process by which the financial stability of issuers was monitored. Through the credit review, it said, the bank “…is able to take immediate action to remove a name from our Approved Issuer List and/or recommend selling the issue, if appropriate”.

In the event, AP retained BNYM to provide global custody services pursuant to a Global Custody Agreement dated 25 October 2004 (“the GCA”). The GCA envisaged that BNYM would also be engaged to provide securities lending services, and there were continuing discussions in that regard in which AP was assisted by an external consultant (Mr Roland Nilsson).

At a meeting of AP’s Risk and Returns Committee on 29 November 2004, Mr Grunditz recommended the segregated account option, and the decision was taken to proceed on that basis. A Securities Lending Authorisation Agreement was duly signed on behalf of AP and BNYM on 22 December 2004. It was agreed that the “net securities lending revenues” would be shared. The agreed split was 85% to AP and 15% to BNYM (the most favourable on offer from the four tenderers).

AP said in opening that the SLAA was a freestanding agreement, separate from the GCA. They were clearly different agreements, but equally in my view they were linked agreements, and BNYM performed both roles, in accordance with what the evidence suggests is the usual practice in the securities lending industry. (This is of some relevance in relation to terms in the GCA which dealt with conflicts of interest—see below.)

By the SLAA, which was subject to English law and jurisdiction, AP authorised BNYM to “establish, manage and administer a Securities Lending Program in accordance with the provisions hereof (the "Program") with respect to the lendable securities held by [AP's] Portfolio”. Not all securities were lendable, for example those of Swedish issuers, as AP later made clear.

As AP’s lending agent, BNYM was authorised to lend securities held for AP to counterparty borrowers, and was authorised to invest and reinvest cash collateral in accordance with Investment Guidelines attached to the SLAA as Exhibit B. This authority is given by clause 6 of the SLAA, which also stipulates for the cash collateral to be invested in a segregated rather than a commingled account:

Clause 10(a)(ii), is part of a “Risk of Loss” clause. By it, AP agreed that losses resulting from investment of the cash collateral were for its account, except where the losses resulted from BNYM’s negligence. AP:

“… acknowledges and agrees that any losses of principal from investing and reinvesting cash Collateral (“Principal Losses”) shall be at the Portfolio’s risk and for the Portfolio’s account except to the extent, if any, that such Principal Losses result from the negligence or wilful default or fraud of the Lending Agent, or the failure of the Lending Agent to comply with the provisions of this Agreement.”

Clause 10(b)(i) SLAA is an important provision which defines the “Standard of Care” required of BNYM:

“The Lending Agent shall perform its obligations under this Agreement with the care, skill, prudence and diligence which, under the circumstances then prevailing, a prudent person acting in a like capacity and familiar with such matters would use in the conduct of an enterprise of like character and with like aim”.

Clause 10(b)(ii) SLAA is a “Limitation of Liability”, providing that

“Except to the extent, if any, otherwise specifically provided herein, the Lending Agent shall not be liable with respect to any losses incurred by the Client or the Portfolio in connection with this Agreement or the Program, except to the extent that such losses result from the Lending Agent’s negligence or wilful default or fraud in its administration of the Program, or the failure of the Lending Agent to comply with the provisions of this Agreement. Notwithstanding any other provision of this Agreement, under no circumstances shall the Lending Agent be liable for any indirect, consequential, or special damages with respect to its role as Lending Agent.”

Thus losses resulting from BNYM’s negligence and failure to comply with the SLAA do not fall within the exclusion (this is not a case raising issues as to indirect damages, etc., referred to in the exclusion). BNYM submits that it is not open to AP to advance claims for breach of fiduciary and regulatory duties, in respect of which it alleges neither fraud nor wilful default, but because of the view I take as to AP’s case in respect of such duties, I need not decide the point.

The Investment Guidelines are Exhibit B to the SLAA. This is headed “Första AP-Fonden U.S. Dollar Cash Collateral Reinvestment Account (hereinafter the “Account”)”. The guidelines have four sections (or clauses). The first is headed “Objectives”, and provides as follows:

“The Account shall be for the management of cash collateral supporting securities loans, the key objectives of which are to:

- keep all cash collateral and related investments in a segregated account(s) in the name of [AP]

- safeguard principal

- assure that all cash collateral is invested in a timely manner

- maintain a diversified portfolio of investments

- maintain adequate liquidity in the Account to meet the anticipated needs of [AP] and/or its investment advisors; and

- consistent with these objectives, to optimize the spread between the collateral earnings and the rebate rate paid to the borrower of securities.

The following standards have been designed to complement the preceding objectives.

To be eligible for cash collateral investment, a security must be rated by both Moody’s Investor Service and Standard & Poors (“Rating Agencies”) at the required quality level at the time of purchase. …”

There then follow the remaining three sections. Section 2 begins, “Acceptable cash collateral investment instruments include”. There follows a list of instruments, and the required rating (except in the case of US treasuries). Thus section 2d) permits investments in instruments issued by domestic corporations (that is US corporations) including corporate and floating rate notes rated A2/A or better. Section 2h) permits investments in “Asset-backed securities rated AAA”. There was some debate as to which of these categories the Sigma MTNs fell into, but it is not now in dispute that they were asset-backed securities which were rated AAA at the time of purchase, and so eligible investments under the guidelines. Investment in some types of instrument was expressly prohibited, including derivative securities.

Section 3 begins, “Maturity and duration constraints are as follow”. Section 3a) sets a portfolio-wide weighted average maturity (WAM) of 30 days. Section 3c) requires floating rate notes to have a maximum term of two years, or in the case of asset-backed floating rate notes, a projected maximum life of two years and a final stated maturity not exceeding five years. Again, it is not in dispute that this covered the Sigma MTNs.

Section 4 begins, “Diversification standards are as follows”. There are two relevant ones (both of which are in dispute as regards the Sigma MTNs). Section 4e) provides that “No more than forty percent of the Account may be invested in floating rate … securities”. Section 4h) provides that, “The Account may not be invested in illiquid securities. For the purposes hereof, an illiquid security is defined to be an issue that is not sellable without the sale proceeds being materially impacted due to a lack of other trading activity in the security. …”.

There are a number of points in issue on the effect of the SLAA and particularly the Investment Guidelines. As identified by the parties these are:

Some of these are only relevant if I find in AP’s favour on its contentions. Thus, if the acquisition of the Sigma MTNs in fact complied with the “safeguarding principal” objective at the time of acquisition, then it may be academic whether (as AP argues) this objective was part of the mandate. If BNYM in fact monitored Sigma with due care over the relevant period, the disputed scope of their monitoring obligations may be likewise academic.

Was preservation of principal part of the mandate?

In the response which BNYM sent to AP on 30 July 2004 as part of the tendering process, it said that, “Chief among Mellon Bank’s reinvestment objectives is the preservation of your principal”. AP’s case is that this was also its chief objective. Despite a number of points that BNYM has raised, as a matter of evidence, and of common sense, I accept that this was indeed the position.

Based on it, AP has a breach of mandate case to the effect that the “safeguard principal” objective, which is contained in section 1 of the Investment Guidelines, was part of the bank’s mandate, in the sense of defining the bank’s authority to make investments.

AP contends that on a proper construction of the scheme of the SLAA Investment Guidelines, the key objectives as set out in section 1 set out the risk profile and objectives to be achieved in the management of the cash collateral in its account, and the following sections 2, 3 and 4 set out standards or “tactics” by which those objectives could be achieved. It says that the latter “tactics” were secondary, and designed to “complement” the key objectives which were, at all times, of primary and overriding importance. Based on this construction, AP puts its case in breach of mandate, contending (to quote its closing submissions) that “only investments which when purchased had a “near certainty of payment” were within the mandate and objectives agreed”.

A number of textual points are made by AP to support its construction. It is said to be necessary to give effect to the word “objectives”. It is said that passing the threshold of eligibility does not mean that the overall objectives under section 1 can be ignored. Reliance is placed on the evidence as to the approach of BNYM’s Credit Analyst Group, which (it is not in dispute) looked beyond the rating criteria.

BNYM submits that the “objectives” section in the Investment Guidelines should be read as setting out the basis on which the detailed provisions following were selected, rather than as themselves imposing separate obligations. It submits that under the SLAA, it was authorised to invest cash collateral in accordance with the Investment Guidelines. Nothing in the SLAA required it to apply a higher standard when making investments than that set out in the Investment Guidelines. The operative provisions of the Investment Guidelines, governing which securities could be purchased for AP, are those under sections 2-4. The “objectives” section does not have any independent effect and does not modify the formal requirements of sections 2-4 with regard to credit quality, liquidity or diversification.

My conclusion on the point of construction is as follows. Clause 6 of the SLAA, which I have set out above, authorises BNYM to invest in accordance with the Investment Guidelines in Exhibit B. This constitutes the bank’s authority, or to use the phrase used in the case-law in the banking context, the bank’s mandate. So, for example, a bank paying a cheque with a forged signature is said to have paid in breach of mandate, because the customer never authorised it, and the payment is treated as for the bank’s own account (Barclays Bank Ltd v WJ Simms Son & Cooke (Southern) Ltd [1980] QB 677 at 699, Robert Goff J).

In this case, AP seeks to construe the “safeguard principal” objective as part of the mandate. I do not accept that submission. As a matter of commercial sense, it is important that the authority of a lending agent to invest cash collateral on behalf of a lender is clearly defined, for the protection of both parties. As BNYM says, a requirement to “safeguard principal” is too vague to be of contractual effect as part of the mandate. As a matter of construction, I consider that the objectives section of the Investment Guidelines (as it says) states the key objectives for the management of the cash collateral, including the “safeguard principal” objective. But the scope of BNYM’s authority to invest, in other words its mandate, is defined in the subsequent sections. These are the sections which specify the type of instrument that is eligible for acquisition as collateral, the rating and maturity requirements of such instruments, and the overall diversification standards for the account.

Were the Investment Guidelines applicable at time of purchase only?

There is a dispute between the parties as to whether the Investment Guidelines applied at the time of purchase only, as BNYM submits, or whether they continued to apply so long as the investments were held on AP’s account, as AP submits. This is linked to an issue between the parties as to BNYM’s ongoing duty to monitor the investment portfolio for compliance with the guidelines. BNYM says that it was not under such a duty. The issues are relevant to AP’s retention claim.

AP submits that BNYM was contractually obliged to monitor the investments in its account post-acquisition on an on-going basis to ensure that the account remained in accordance with the Investment Guidelines. AP accepts that the guidelines expressly provide that ratings criteria were to apply on acquisition only. But, it says, BNYM held itself out to AP as having, and both parties understood BNYM to have, obligations to manage and to monitor the investments in AP's account post-acquisition for on-going compliance with the Investment Guidelines. The provision as to ratings only meant that there was no automatic sale of the security out of the account if a rating was lowered.

In particular, to quote from its closings, AP contends that “BNYM was obliged under the terms of the SLAA to manage and to monitor the investments in AP's cash collateral account post-acquisition - in order to achieve the key express objective to ‘safeguard principal’. This meant that BNYM was obliged to monitor and manage the account for compliance with the objective of ensuring that the securities remained ‘money good’ (i.e. near certainty of full principal repayment), and to ensure ongoing compliance with AP's investment guidelines”.

To give support to its construction, AP relies on the fact that section 1 provides that the account is to be for the management of cash collateral with the objective of maintaining a diversified portfolio of liquidity. AP also makes reference to material both before and after the conclusion of the SLAA. On 30 July 2004, BNYM sent AP a response to follow up questions raised by AP in the tender process. Under the heading “Continual Monitoring” it said that, “In addition to a formal yearly review, Mellon Bank’s team of credit analysts monitor our issuers on a daily basis. Daily monitoring is vital to stay ahead of any credit problems that may arise and to anticipate potential downgrades in advance of rating agency announcements…”. The response also said that, “…on the reinvestment side, Mellon Bank as manager of cash re-investment activities utilises a combination of automated and manual process and procedures ensures that cash collateral investments conform to investment guidelines”. I agree with AP that the latter appears to be an assurance that there will be continuing compliance with the guidelines.

Internal Mellon Bank securities lending policy documents (though made after the agreement was entered into) are also relied on by AP, and in parts suggest a continuing duty to monitor, including a statement that separately managed accounts "will adhere to client prescribed liquidity targets". As against that, what seems to be the most explicit statement of all in the policy documents is to the effect that “All investment guidelines are applicable at the time of purchase”.

AP put to a number of witnesses that management of risk implies a dynamic review, and a constant review of issuers, and Mr Blanchard in particular agreed. AP also relies on his evidence to the effect that the bank would have to monitor the client’s investments.

AP places particular reliance on the evidence of Mr Fort that, “Once investments were allocated to appropriate portfolios by the cash reinvestment group, one of the tasks of the portfolio managers was to monitor the adherence of their portfolios to their respective investment guidelines, as well as to analyse and review each client's asset liability mix and cash flow". In oral evidence BNYM’s experts Mr Zimmerhansl agreed with this statement and Mr Richard said that it was his experience as well.

AP also places reliance on the deposition of Mr Fort in the Oklahoma proceedings that I mentioned on 20 January 2011, where in a discussion on investment guidelines (and leaving out objections) he says as follows:

Finally, the four annual compliance reports dated 28 February 2005, 31 January 2006, 31 January 2007 and 31 January 2008 are headed “Compliance Checklist Investment Restrictions”. I agree with AP that they demonstrate that BNYM did in fact seek to monitor AP's account against the Investment Guidelines.

BNYM submits that the Investment Guidelines are to be construed as applying only upon the purchase of investments. They do not govern BNYM’s obligations with regard to existing investments.

BNYM submits that the purpose of the Investment Guidelines is set out at clause 6 SLAA, pursuant to which BNYM was “authorized to cause the investment and reinvestment of all cash Collateral held in respect of loans of the Portfolio’s securities in accordance with the Investment Guidelines”. The authority granted is “to cause the investment and reinvestment of all cash Collateral”. To cause an investment means, BNYM submits, to apply funds to the purchase of a security. This makes explicit that the function of the Investment Guidelines is to determine what can be purchased for AP’s account at any given time, rather than to provide a set of ongoing requirements governing whether or when to hold or sell investments once purchased. The effect of clause 6 SLAA is therefore that BNYM is empowered to invest AP’s cash collateral in any investment which, at the time of investment, satisfied the criteria of the Investment Guidelines.

Based on Mr Blanchard’s evidence, BNYM submits that while evidence of pre-contractual negotiations is of limited relevance, it is clear that AP well understood, on entering into the SLAA, that its Investment Guidelines would apply only at time of purchase. I do not accept this submission, which is contrary to the evidence of Ms Hammarlund.

In addition, BNYM relies on its expert evidence. Mr Richard’s evidence was that in the absence of specific language to that effect, “it was not standard [practice] to alert a client of a security falling out of compliance if the thought was to stay the course and to continue to hold the investment”. Mr Zimmerhansl said that continual monitoring does occur, but drew a distinction between continual monitoring and compliance with the original acquisition criteria.

In summary, BNYM accepts that it did perform ongoing credit monitoring of investments, but it says that it did so on a programme-wide basis via the Credit Analyst Group. It accepts that it is good practice to monitor investments on their approved lists. It does not follow however that absent an express provision in the Investment Guidelines requiring ongoing monitoring, it had a legal obligation to do so. Rather given the number and diversity of its clients, it is difficult to see (it says) how in practice such a duty could be applied.

The question of the post-acquisition application of the Investment Guidelines is a matter of construction of the agreed terms. The only express provision dealing with the point is in section 1, which states that to be eligible for cash collateral investment, a security must be rated AAA by both the main ratings agencies at the time of purchase. However, this does not settle the timing question as to the rest of the guidelines.

As to the contractual language, on BNYM’s case, I do not think that the words “to cause the investment” says anything about whether or not compliance must be ongoing. On the other hand, I consider that the provision that the account is to be for the “management” of cash collateral with the objective of “maintaining” a diversified portfolio of liquidity supports AP’s case that the intention of the parties was that there would be ongoing monitoring of the account by BNYM.

BNYM draws a distinction between an obligation to monitor an account for compliance on an ongoing basis, as new investments are being made for it, and an obligation to monitor the individual investments in that account for ongoing compliance. There is perhaps a distinction, but the concepts obviously overlap.

BNYM points out that the pre-contract negotiations and the implementation of the contract by BNYM are not admissible in construing the contractual terms. But I think it is common ground that good practice on the part of a securities lending agent such as BNYM can be taken into account in ascertaining the meaning of the written terms. This is because, “What the parties agreed, explicitly or implicitly, can only be judged against the factual background they knew, which must include practices of any particular market in which they operate and in which the agreement was made” (Crema v Cenkos Securities plc [2010] EWCA Civ 1444 at [45], Aikens LJ).

Whether as a matter of obligation, or good practice, it is clear that BNYM monitored the account, and monitored individual issuers. It would be surprising if it did not do so. I have set out the details above and need not repeat them. I reject BNYM’s suggestion that all this shows is verification that investments complied with the Investment Guidelines at their time of purchase.

However, there are limits as to how far this point goes. As BNYM says, many aspects of investments, such as the issuer, maturity date and asset type, will not be expected to vary over time anyway. Ratings can vary, but the Investment Guidelines expressly state that rating criteria apply at time of purchase. I think that BNYM is right to say that in terms of AP’s Investment Guidelines, the only provisions that could theoretically impose any ongoing obligations are those dealing with diversification and liquidity.

As to diversification, section 4e) of the Investment Guidelines provides that no more than 40% of the account may be invested in floating rate securities. The Sigma securities fell within this category. There are other similar provisions. BNYM says that it is not practical for the lending agent to manage the cash collateral account on the basis of an ongoing restriction, because it is outside its control to ensure compliance. The makeup of the account is liable to fluctuate significantly on a day to day basis, with the lending agent having no advance notice. An account may go in and out of the limit.

The conclusion that BNYM seeks to derive from this is not, in my view, supported by the facts of the present case. On 28 February 2008, following a review of AP's account, BNYM realised that floating rate notes amounted to 55% of its portfolio. Its contemporary emails say that this was not "a technical violation and is in fact an operational error on our part", and that the account "has been in breach for months". Quite clearly, BNYM regarded this as a breach of its contractual obligations. The matter was brought to AP’s attention, and resolved with AP’s approval by closing positions without taking losses. The account was brought back within the limit by 20 March 2008.

Turning to liquidity, section 4h) of the Investment Guidelines provided that the account was not to be invested in illiquid securities. However, monitoring the continuing liquidity of particular investments would be a far more difficult exercise than monitoring the 40% limit on securities with a floating interest rate just discussed. In fact, two expert market professionals are in dispute as to whether the Sigma MTNs ever satisfied the liquidity requirement, and the parties are not in agreement as to what liquidity means in this context.

Further, BNYM submits that it would make no sense to apply the liquidity requirement as an ongoing prohibition—the prohibition makes sense in the context of new investments, but if applied on an ongoing basis, it would imply that BNYM was under an obligation to sell securities at the precise moment when such a sale was most likely to expose the lender to an unnecessary loss. This in my view is a reasonable submission.

My conclusion on this issue is as follows. I reject BNYM’s submission that it was under no obligation to monitor investments for compliance subsequent to purchase, and that it was not obliged to monitor the credit quality of investments post-purchase. In my view, on the construction of the agreed terms, BNYM had an on-going duty to monitor the account for compliance with the Investment Guidelines. That duty included paying continuing regard to the key objectives, but as objectives, not as part of its mandate. (It is not necessary to add further words to the “safeguard principal” objective, as AP sought to do, because it is self explanatory.)

Beyond that, I consider that question of compliance depends on the nature of the guideline in question. BNYM was, in my view, under an ongoing duty to see that no more than 40% of the account was invested in floating rate securities, and if the limit was exceeded, to take steps to bring the account back into compliance.

In my view, however, the guideline which provided that the account was not to be invested in illiquid securities raises different considerations. I prefer BNYM’s case that it applied at the time of acquisition of the securities in question, but not on a continuing basis. I consider that this gives a sensible commercial meaning to section 4h) taken in context. The alternative finding that AP asks me to make, namely that “BNYM was obliged to monitor and manage the account with the objective of ensuring that the securities remained ‘money good’ (ie near certainty of full repayment)”, would have been completely unworkable.

In any case, I cannot see that the contract required the bank to dispose of the securities upon them becoming illiquid. That would be inconsistent with the “safeguard capital” objective if it resulted in an unnecessary loss, and would make no commercial sense. AP came close to accepting this obvious point in oral closings.

What was the scope of BNYM’s duty to communicate with AP?

In its closing submissions, AP contends that “full and open communication between BNYM and AP on matters concerning AP's Account was in accordance with BNYM's fiduciary responsibilities to AP, was expected (on both sides) and was in accordance with market practice”. AP appears therefore to put its case in this respect on the basis that BNYM owed it fiduciary duties. The main finding it seeks is that “BNYM was obliged to provide full and open communication between BNYM and AP as soon as there was a problem with a security which threatened the near certainty of repayment in full at maturity”.

BNYM’s case is that the level of communication required by it was set out in the Global Custody Agreement and, specifically, the Service Level Agreement annexed thereto which provided for BNYM to report via its online Workbench system. In addition, AP requested reports to be sent to it daily, which included information on the rating of securities. As regards other information such as market concerns, that formed no part of the contract with BNYM. It relies on the fact that AP's own evidence is that its witnesses did not expect BNYM to be providing it with information about investments, expecting the bank to manage the portfolio itself.

The first question to address is what was specified by way of communication from BNYM. The GCA refers to a Service Level Agreement, though the copy in the documents is unsigned. It states in very general terms that BNYM will provide AP with “details of collateral holdings proving they are covered by the appropriate percentage, as agreed in Securities Lending contract, against their outstanding loan positions via Workbench”.

Ms Hammarlund describes in her evidence what was provided through Workbench, which is a web-based information reporting tool. There were various reports available to AP. This included pricing information on the securities, which (as described later) was provided by IDC, a pricing vendor. Mr Grunditz explains that his understanding was that the market value figures on these reports were only updated on a monthly basis, and I believe that this is correct.

When it entered into the agreement, however, AP also required a daily report, which showed the securities held in the cash reinvestment portfolio and certain other information including maturity dates, ratings, par values and amortized cost. These daily reports did not include pricing. The example I have seen runs to five pages, and shows both acquisitions of the Sigma MTNs under “SIGMA FINANCE”.

Mr Grunditz says that he reviewed the holdings reports on a relatively infrequent basis. In cross examination, he said that he reviewed them perhaps once a quarter, and once every six months there was a random sampling of the portfolio. He said that AP did not have the resources to carry out any in depth analysis. Although it had a risk-management system called “Dimension”, for technical reasons it was not possible to monitor the cash collateral portfolio through the system. He said that BNYM had emphasised its monitoring of investments as one of its selling points in the tendering process, and AP relied on the bank to do this.

So far as factual, I accept Mr Grunditz’s evidence in this regard. However, it does not alter the fact that AP knew from the daily reports that BNYM had acquired the Sigma MTNs as collateral for its account. I agree with BNYM that neither Mr Grunditz nor any other AP officer seems to have done very much to monitor the risks to which AP was exposed through its cash collateral portfolio. Their response was that they trusted BNYM in that regard, and could not practically interfere in its discretionary mandate. This is a reasonable response in itself, but as investment professionals I consider that they could have done more.

My conclusion is as follows. In agreement with BNYM, I consider that the alleged duty of full and open communication on matters concerning the securities lenders’ account is imprecise. As BNYM says, it was a discretionary investment manager, not an adviser of a party which was going to take its own investment decisions. Its role under the SLAA was to manage AP’s portfolio within the bounds of the Investment Guidelines.

The bank says that if the court were to find that discretionary investment managers were required to keep clients “fully informed” about their portfolios, including providing information on any market concerns and their own views relating to the investments in client portfolios, this would cause an “earthquake” in the discretionary management industry, both for managers (who would be obliged to completely overhaul their reporting) and their clients (who would find themselves on the receiving end of masses of information).

I accept BNYM’s submission in this respect. Although Mr Blanchard seemed at one point in cross examination to express the view that the bank was under a duty to disclose all material facts to the client, he was on the sales side of the business, and may not have understood the implications of what he was being asked. Mr Zimmerhansl made it clear in re-examination that this was not his view.

The facts of this case illustrate that this was not the intention of the parties. AP’s witnesses made it clear that they did not expect the bank to be providing it with information about investments beyond what was agreed. Quite the reverse, they expected the bank to get on and manage the portfolio without reference to AP. Up to May 2008, I am satisfied that BNYM provided AP with the information that the parties had specified, and that was adequate. That did not preclude AP asking for more information.

In normal circumstances, BNYM was under no obligation to do more than this. I consider AP’s case that BNYM was “obliged to provide full and open communication between BNYM and AP as soon as there was a problem with a security which threatened the near certainty of repayment in full at maturity” is unworkable, and I reject it.

However, the question remains as to the bank’s duty to communicate with its clients if things went wrong. This was the subject of considerable cross examination by AP of the bank’s witnesses. They put it in slightly different ways. The following exchange took place with Mr Mannix. He was on the trading side of the bank, responsible for cash collateral reinvestment (reporting to Mr Ford and later Ms Rulong) but was the most senior witness called by the bank, apart from Mr Ford, and was a good witness.

The evidence of the bank’s expert, Mr Zimmerhansl, was significant because he presently runs a securities lending operation. In his witness statement he said that BNYM was under no obligation to communicate to AP that the Sigma securities were impaired. He said that in his experience, when information is passed to a client, it is done simply to provide the client with information as a courtesy. In this regard, the following exchange took place in cross-examination.

As this evidence makes clear, so far as securities lending is concerned— unlike in some other investment contexts—it is exceptional for there to be a loss. This reflects the fact, discussed elsewhere in this judgment that a key objective in relation to cash collateral is to safeguard principal. The fact that the situation is unusual fits with the contemporary evidence. As I explain later, the bank regarded the treatment of customers as regards Sigma as “fairly uncharted territory”. The adequacy of its communications with AP during this period has to be seen in that light.

On the evidence, I am satisfied that in performing its obligations under the agreement with its client, it is best practice for a securities lending agent to communicate with its client if there is a serious risk of loss in respect of securities held as collateral for the client’s account. BNYM says that the mere fact that something is best practice does not necessarily mean that it arises as a matter of obligation, with which I agree. BNYM also says that a contractual duty of care such as that found in clause 10(b)(i) of the SLAA requires a party to carry out its obligations under the contract with the requisite level of care and skill, but does not impose additional obligations not otherwise found under the contract. Whilst that proposition may be correct in general terms, in this case it is not disputed that there was a contractual duty to communicate. In performing that obligation, BNYM had to exercise the contractual duty of care, that is, to the standard of a careful and prudent securities lending agent. I am satisfied that this may require communication with its client if there is a serious risk of loss in respect of securities held as collateral for the client’s account. This is not an extravagant obligation, and is one that was recognised by the bank’s own officers. Whether it does so, or not, depends on the facts, as does the scope of the communication. Where (as here) the securities lending agent is not acting in an advisory capacity, there may be no obligation to give a recommendation, but if it does so, it must similarly do so carefully. (For reasons stated below, I do not agree however with AP that these duties arise out of the securities lending agent’s “fiduciary responsibilities”.)

However, given the view I take on the facts, it may not be necessary to decide this. As I describe later, once Sigma was downgraded from AAA in April 2008, and in particular once the indicative price of the securities dropped sharply at the beginning of May, exposing AP to an unrealised loss of US$10.5m, the bank decided to contact AP and other clients about the position. It is not in dispute that it had to do so with all due care. AP’s case is that it should have been contacted earlier. I discuss this issue later.

The commingled fund and “programme approach” issues

As I have described, AP chose a segregated account rather than entering one of BNYM’s commingled funds. The essence of its point in this respect is as follows. BNYM held a large quantity of Sigma securities, and in allocating them to AP’s account, AP says that it was treated as though it was in a commingled fund. This was because BNYM managed its cash collateral investment on a “programme approach”, in other words across its whole range of securities lending clients, instead of treating AP as an individual account. AP says that clients such as itself did not expect to be treated as part of a much wider exposure to securities as part of a "programme approach" or, in effect, as part of a commingled fund, by reason of very large aggregate positions in securities held by the bank.

BNYM says that “programme wide” approach does not mean that BNYM took a decision that all clients would be treated identically regardless of their investment guidelines. It did mean that, insofar as clients were subject to the same investment considerations, there was no reason for BNYM to differentiate between them. That is so whether or not BNYM was subject to any legal duty to act fairly as between its clients, because it follows as a matter of contract and common sense.

In its closing submissions, AP says that what was presented initially as a bank policy, turned out to be nothing more than a resort to common-sense. There is force in that comment, and this issue had less significance by the end of the trial than appeared at the beginning.

After Sigma was approved as an issuer by Mellon Bank in 2005, it acquired a substantial holding of its securities. By the time of the disputed purchases for AP in March/April 2007, its total aggregate holding of Sigma MTNs was approximately US$2.9 billion. Taking AP’s calculations, this corresponded to nearly 6% of Sigma’s total issued debt, and 9.7% of its total US dollar MTNs issued at that time.

The two purchases for AP’s account at issue in these proceedings are:

The merger of Mellon with Bank of New York took effect on 1 July 2007. The latter also had a substantial Sigma holding. The precise figures are hard to pin down, but at this time the merged entity’s combined holding of Sigma was about US$5.5 billion. Taking AP’s calculations, this corresponded to just over 10% of Sigma’s total issued debt, and about 16% of its total issued US dollar MTNs. AP’s case is that this was an enormous holding.

My findings in this regard are as follows. At the time the relevant Sigma MTNs were purchased for AP in March/April 2007, Mellon Bank’s Sigma holding was already large, but I accept Mr Tant’s evidence that it was within its guidelines. After the merger in July 2007, the combined holding of over 10% of Sigma’s total debt was very large, and in excess of the bank’s guidelines. The bank clearly would never have considered acquiring such a large holding but for the merger.

Again however, I accept Mr Tant’s evidence that at the time of the merger this was not expected to cause problems. BNYM restricted further purchases of Sigma until the proportion could fall to below 10%, which was anticipated (I consider reasonably) to happen as issues matured and were repaid. Despite all the problems that ensued, maturities did continue to be met until September 2008.

There were no further purchases of Sigma by BNYM, and the onset of the credit crunch was soon to bring Sigma’s issuance to an end, which happened in October 2007. As to the possibility of reducing its exposure further by disposing of its holding, I accept Mr Richard’s evidence that thereafter it was highly unlikely that the whole of the bank’s holding could have been disposed of, given the state of the market. As liquidity deteriorated in 2008, and Sigma’s position itself deteriorated, I am satisfied that the sale of very large blocks would have become more and more difficult.

It is clear that the size of its Sigma holding taken together with the adverse market conditions placed BNYM in an exposed position, and I reject any contention to the contrary. In a 1 February 2008 email, Mr Ford, who was co-head of Securities Lending at BNYM, and who had a way of describing home truths in plain language, referred to it as the “5 billion dollar problem”. The fact that BNYM would not have been able to dispose of the bulk of its Sigma holding even had it wished to must have coloured its approach.

In fact, Mr Richard’s evidence suggests that Mellon’s holding in itself would have been too big to have been disposed of in the market conditions that prevailed. It follows that the problem existed on the basis of a holding of around 5%, which though large, was clearly not excessive. Further, it is not in dispute that it was possible to dispose of smaller tranches of the size of AP’s holding by sale or ratio trade by which Sigma bought debt back in return for asset sales, and as I shall describe such transactions did take place. This point does not therefore make much practical difference.

AP notes that BNYM has argued that if it exited Sigma MTNs for AP, fairness to all note holders required it to exit for everyone. Without expressing a conclusion on that point, I accept AP’s submission that BNYM was bound to apply the Investment Guidelines agreed in the SLAA regardless of the position of its other clients. I did not understand BNYM to contend differently in its closing submissions.

However, the question arises as to what AP expected application of the Investment Guidelines to involve, and whether it expected to be treated as one of a considerable number of clients, both segregated and commingled, or in effect to be treated as a single client in isolation.

I begin with a point about volume. After Sigma securities had been approved as an issuer, Mr Servis suggested that BNYM should have imposed an initial maximum limit of in the region of US$300 million on Sigma acquisitions across all its clients. In the context of securities lending, I consider that this would be a very small amount. There is no basis in the evidence for this suggestion, and I reject it. I accept the evidence of Mr Tant, who said that US$300 million “is the approximate size of a single purchase of Sigma notes. … There would be no purpose in approving Sigma with such a low limit.”

This gives an indication of the size of transactions that BNYM carried out in its investment function, a point also referred to by Mr Zimmerhansl in his evidence. Large purchases allow the agent to obtain better prices. Once an investment has been made for a client in accordance with its guidelines, the securities lending agent is able to manage it on a common basis with other clients holding the same investment. BNYM says that AP must have known this when it entered into the contract.

In fact, AP’s Investment Guidelines themselves recognised the large potential size even of its own holdings. This clearly appears from section 4a), which provided that the greater of US$100m or 3% of its account could be invested in the obligations of any one issuer.

AP accepted in opening that in order to get the best price for all clients, the lending agent as discretionary investment manager may in practice purchase a block of securities and then allocate these securities to those clients whose investment guidelines permit the purchase of the security. However, it says, this is a matter of administrative convenience for the manager to execute the securities transactions more efficiently, and should not impact in any way on the monitoring or decision-making duties owed by the manager to individual clients.

I do not accept that this reflects the evidence. The tender process sheds some light on the issue. I have referred above to the response which BNYM sent to AP on 30 July 2004. The document was put in cross-examination to Mr Grunditz, who was asked whether it was not clear to him that the way that the bank was conducting its securities lending programme was on a global basis, looking at issuers and issues for clients globally. His answer appeared to acknowledge that it was. He said that, “when I read the document, yes, it’s true that they assess this on a global level, but what we signed was a discretionary mandate”.

He did not however assent to the proposition that there is no inconsistency between looking at securities on a global basis, and a discretionary mandate for a particular account. In fact, on the evidence I do not think that there is any inconsistency, given the nature of securities lending.

It has to be kept in mind that at the time of the tendering process in 2004, AP was contemplating making a considerable proportion of its entire portfolio (nearly a third) available for lending. This led eventually to a potential pool of US$10 billion of AP securities available for lending out. AP was examining four tenders for the custodianship/securities lending role (which clearly went together) from leading market participants, and had the assistance of an independent consultant. Having heard them give evidence, I am satisfied that the relevant officers of AP are people who would have made it their business to understand how the investment of cash collateral would be conducted in practice. Once the contract was operational, they had daily reports which set out what collateral was being held on their account.

I consider that BNYM is entitled to the findings of fact that it seeks in this respect. It must have been obvious to AP in appointing BNYM as its securities lending agent that it would be one of a considerable number of clients, both segregated and commingled. It cannot have expected that BNYM would operate its account as if there were no other clients. This reflects, in my view, the reality of securities lending business carried on for the mutual convenience of financial institutions. Further, it must have been obvious to AP that various aspects of its account management would be centralised. As BNYM says, one of the main reasons for selecting a securities lending agent is for its investment expertise, including in credit analysis of collateral, which would be expected to be shared across clients. So far as it is to the contrary, I reject the evidence of Mr Cirrito on the point.

Did the bank owe AP fiduciary duties?

It is common ground that BNYM owed AP a duty of care both under clause 10(b)(i) of the Securities Lending Authorisation Agreement and in tort. The issue here considered is whether AP is correct to argue that BNYM also owed it fiduciary duties. BNYM says it is not.

A commercial banking relationship does not generally give rise to fiduciary duties (JP Morgan Chase Bank v Springwell Navigation Corp [2008] EWHC 1186 (Comm) at [573], Gloster J, Aff’d [2010] EWCA Civ 1221). But in relation to the duties of a global custodian, entrusted with the safekeeping of its clients’ securities, and the duties of a securities lending agent, entrusted with the lending out of the clients’ securities against suitable collateral, a bank may in my view be subject to fiduciary duties in respect of those activities (see generally Yates & Montagu, The Law of Global Custody, 4th ed, 2013). These are the kind of activities in respect of which fiduciary duties are capable of arising by virtue of their nature. In principle, therefore, I agree with AP that the relationship between it and BNYM was capable of giving rise to fiduciary duties on the bank’s part.

But a party such as a securities lending agent may be in a fiduciary position as regards part of its activities and not as regards other parts. The fact that BNYM may have been a fiduciary in some respects does not mean that it was a fiduciary in all respects. The court has to consider the particular duty alleged to be breached (Saltri III Ltd v MD Mezzanine SA SICAR [2012] EWHC 3025 (Comm) [123], Eder J).

As a matter of fact, in some of the material in evidence BNYM expressly stated that it is subject to fiduciary duties. At a marketing presentation in Stockholm on 28 May 2004, BNYM (then the pre-merger ABN AMRO Mellon) explained its lending philosophy of minimising “… risks to our clients by applying the appropriate risk parameters and maintaining strict adherence to our fiduciary responsibilities”. The responsible BNYM officer did not shy away from this as a current statement in cross-examination. I have no doubt that this reflects the nature of the bank as an institution with a global securities lending business. Such a business has to run to similar standards wherever it is conducted, particularly those of the United States where BNYM is based.

But it does not follow that because the term “fiduciary” is used in documentation, fiduciary duties necessarily arise under English law. The term “fiduciary” is sometimes used to describe a particular type of financial business (as in Germany or Switzerland), and where it is used in a legal sense, as in common law jurisdictions, the term has somewhat different connotations from jurisdiction to jurisdiction, albeit derived from the same roots in equity. So far as English law is concerned, the essence of a fiduciary relationship is that the fiduciary subordinates his own interests to those of his principal’s (Wollenberg v Casinos Austria International Holding GMBH [2011] EWHC 103 (Ch) at [208], Lewison J).

In the English context, it has been said that the law relating to fiduciary duties in its application to the financial sector and to the interrelationship of such duties and regulatory duties is still in relative infancy (Jackson & Powell on Professional Liability (7th ed.) at para 15-046). This may be because primacy is given to the contract. As was said in the leading Australian case, the contract regulates the basic rights and liabilities of the parties and the fiduciary relationship, if it is to exist at all, must accommodate itself to its terms (Hospital Products Ltd v United States Surgical Corp (1984) 156 CLR 41 at 97, Mason J).

This applies with particular force where the parties are both substantial financial institutions dealing on an arms-length basis. The fact that an agency relationship existed between BNYM as lending agent and AP as lender does not change the analysis (Kelly v Cooper [1993] AC 205 at 213H-214A, Lord Browne-Wilkinson).

BNYM denies that it owed any fiduciary duties to AP, and contends that the terms of the contractual relationship are inconsistent with such the existence of such duties. There is in my view force in this point. Particular reference is made to clause 15(o) Global Custody Agreement, clause 17(e) SLAA, and clause 7 SLAA. These appear to acknowledge that conflicts of interest may arise, that BNYM was entitled to enter into transactions in which it had an interest without making disclosure, and that the agreement did not impose an obligation to lend AP’s securities in the event it made a loan of another client’s securities. (As I have said, although AP says that the GCA and SLAA were separate agreements, they were linked.)

The parties did not agree as to how the law treats the relationship between a fiduciary duty and a duty of care contractual or tortious. The case of Bristol & West Building Society v Mothew [1998] Ch 1 is the leading authority in this respect. It concerned the liability of a solicitor who had given negligent advice. He argued that while he had been negligent, he had not been guilty of a breach of fiduciary duty. Upholding this submission, at page 18F Millett LJ distinguished between them as follows:

“The nature of the obligation determines the nature of the breach. The various obligations of a fiduciary merely reflect different aspects of his core duties of loyalty and fidelity. Breach of fiduciary obligation, therefore, connotes disloyalty or infidelity. Mere incompetence is not enough. A servant who loyally does his incompetent best for his master is not unfaithful and is not guilty of a breach of fiduciary duty.”

As a matter of history, in this litigation AP’s claims were based on negligence, the claims for breach of fiduciary duty being added by amendment shortly before trial. They were based on Mr Ciritto’s expert report, but as BNYM says, the question whether it was a fiduciary and if so what its fiduciary duties entailed, were matters of law for the court. The main such claim was an allegation by AP that the size of BNYM’s holdings of the Sigma MTNs had the inevitable consequence of creating a conflict of interest because of the increased probability that selling a position would impact negatively on the price of the instruments across the programme. That was a true claim for breach of fiduciary duty, but during the course of argument, that amendment was withdrawn by AP. Nothing that I have heard during this trial suggests that this was anything other than the right decision. The supposed conflict of interest would have made a securities lending operation impossible.

What remains? To quote AP’s opening, BNYM owed fiduciary duties to it as a result of the investment management discretion and/or the agency relationship and/or as discretionary investment adviser in the management of AP’s Account under the SLAA (1) to act at all times in AP’s best interests; (2) to treat AP fairly; (3) not to prefer any other client to AP in circumstances where the situation called for equal treatment; and (4) to bring any material information about AP’s investments to AP’s attention, and in particular to give full disclosure of material information during the conversations during the telephone and email conversations in May 2008 in relation to the matters relevant to the Sigma MTNs in AP’s Account.

For reasons set out elsewhere, as regards duty (4) (the alleged duty of disclosure), I have rejected the wide duty of disclosure which AP contends for. Otherwise, within the parameters of the contract, none of these duties were in dispute. BNYM accepts that it had to act in AP’s best interests, to treat it fairly (including in communicating with it about Sigma in May 2008), and not to prefer other clients where the situation called for equal treatment. The issue is whether BNYM complied on the facts of the case.

As it has been put, “… equity is proscriptive, not prescriptive. … It tells the fiduciary what he must not do. It does not tell him what he ought to do” (Attorney General v Blake [1998] Ch 439, 455E, Lord Woolf MR). BNYM submits that a fiduciary duty to act in a client’s best interests means that the fiduciary must avoid conflicts and must not make a secret profit. Positive obligations are engaged only where the duty is breached. Thus a fiduciary in a position of conflict must disclose it and seek consent from his principal, and one who makes a secret profit must give an account and disgorge it. But, it says, those positive obligations arise as remedies rather than standalone duties. I agree with this analysis.

My conclusion is as follows. In broad terms, the issues that arise in this case are whether BNYM should have acquired Sigma securities for AP’s account, whether it should have disposed of them when Sigma got into difficulties, and whether it presented the position fairly when it contacted AP about the subject. There are of course other claims such as breach of mandate, but those aside, I agree with BNYM that in substance what is at issue is BNYM’s competence as a securities lender. None of the breaches alleged by AP involve any duty properly categorised as fiduciary. They are, in BNYM’s words, negligence claims dressed up in fiduciary garb. The fiduciary duty analysis adds nothing in my view, and I do not accept AP’s submissions in that regard.

The contractual standard of care

To recap, clause 10(b)(i) SLAA defines the standard of care required of BNYM as lending agent. It has to perform its obligations under the agreement “with the care, skill, prudence and diligence which, under the circumstances then prevailing, a prudent person acting in a like capacity and familiar with such matters would use in the conduct of an enterprise of like character and with like aim”. A number of disputes arose between the parties as to the meaning of this provision.

AP submits that clause 10(b)(i) sets out the appropriately high standard of care attendant upon a securities lending agent acting under an investment management discretion, i.e. acting to invest monies prudently on behalf of another party with the fiduciary responsibilities relating to protecting principal attendant upon that role. (As to fiduciary responsibilities, I refer to the discussion above, but the submission is not dependent on the bank’s position as fiduciary.)

As to whether AP is right to argue for a high standard of care, it submits that the addition of the words “prudence and diligence” must connote a higher standard than had those words not been included. In the context of securities lending, where clients such as AP were told that the bank selected only those credits and investment opportunities that provide a near certainty of full repayment of principal, and superior secondary market liquidity, the word “prudence” connotes caution, in the sense of trusteeship.

The standards of “prudence” and “diligence”, AP submits, also connote a high level of procedural care, requiring documented policies and procedures, and an emphasis on regulatory compliance and compliance with market practice. As to this, I am satisfied from the evidence and material disclosed that BNYM’s procedures were adequately documented. No separate point arises in this respect.

BNYM submits the standard of care imposed on BNYM under clause 10(b)(i) SLAA is equivalent to the usual professional negligence standard of reasonable care and skill. In particular, the use of the word “prudent” does not connote a higher standard, whether by comparison to the standard of care owed by a trustee or otherwise. Further, the standard of care required of BNYM is limited to taking reasonable care and skill in performing its obligations under the SLAA, and cannot give rise to the additional obligations.

I have touched already on the last contention. Clause 10(b)(i) provides for the standard of care to be observed by BNYM in performing its obligations under the agreement. BNYM says that it does not in itself give rise to, as it was put, a raft of additional duties. I agree that the clause cannot extend the ambit of BNYM’s obligations as securities lending agent, but in performing that role, BNYM had to meet the contractual standard of care. BNYM accepts (for example) that the purchase of a permitted type of investment can be challenged if the bank fails to use reasonable skill and care to select an appropriate asset of that type. It accepts that in putting the Sigma risks to AP in May 2008 it was under a duty do so fairly, though these points are not spelled out in the agreement. I do not think that anything turns on this contention. So far as this case is concerned, if the bank falls short of the contractual standard of care, liability results not by way of imposing additional obligations, but of giving effect to the existing ones.

The main difference between the parties at trial was whether the clause imposed a “high” standard or the “usual” standard of reasonable care. I doubt there is much difference of substance between them. BNYM is the industry leader in securities lending, and it would be surprising if it did not seek to uphold high standards in that role. In their evidence, its witnesses stressed their commitment to client care and observing best practice.

As a matter of wording, AP has pointed out that the words of clause 10(b)(i) track the language of the US Employee Retirement Income Security Act of 1974 (ERISA), a federal law that sets minimum standards for pension and health plans. However, no particular construction was advanced at trial based on this fact. My preference is for AP’s contention, because BNYM was entrusted with lending out the portfolio of a pension fund which had important public duties (see the similar point made in Bartlett v Barclays Bank Trust Co Ltd (No 1) [1980] 2 WLR 430 at 534 by Brightman J). However, it makes no difference to the outcome. I would have reached the same conclusion on either interpretation.

AP’S CLAIMS RELATING TO THE ACQUISITION OF THE SIGMA MTNs

There were two tranches of Sigma medium term notes acquired by BNYM for AP’s account, being US$5.5m (maturing 12 March 2009) acquired on 12 March 2007, and US$30m (maturing 30 October 2008) acquired on 26 or 30 April 2007 (the date of acquisition being in dispute).

AP’s case in respect of acquisition as set out in one of the schedules attached to its closing submissions is (in summary) that:

AP’s pleaded case is that the Sigma MTNs were not a suitably conservative investment for AP’s portfolio, and were not an investment that would have been made by a prudent person having regard to AP’s core objective and its low appetite for risk. This broadly corresponds with AP’s issues (1) and (7) above and raises a number of questions. First, what was AP’s appetite for risk, and in particular, was it a conservative securities lending client as it has asserted? Second, were the Sigma MTNs a “suitably conservative investment” as collateral for AP? Third, does “suitability” as a concept apply, given that the parties had agreed detailed Investment Guidelines?

Was AP a particularly conservative securities lending client?

In its opening submissions AP contended that, as BNYM knew from the tender process and reflected in the contract, AP was a conservative and risk-averse securities lending client. This was an important part of its case. It said that whether or not Sigma MTNs were acceptable as collateral to securities lenders generally, they were not acceptable to AP. In fact, it is not in dispute (as Mr Zimmerhansl put it) that participants in the business are conservative. As BNYM expressed it, the issue on the pleadings and the evidence is whether AP’s Investment Guidelines marked AP out as a particularly conservative securities lending client. BNYM’s case is that they did not, as shown by the wide range of asset classes and ratings permissible for investment.

The evidence given by Mr Grunditz was to the effect that AP relied on BNYM to provide it with draft guidelines that were appropriately conservative. The facts in this respect are as follows. In the tender process which I have described above, the response of 17 May 2004 that BNYM sent to AP attached the investment guidelines for the Delaware-based Mellon GSL Pennsylvania Fund for USD investments, in other words a commingled fund. AP knew that it was a commingled fund, and had yet to decide whether to go that route. It is not in dispute that AP had the opportunity to amend these guidelines if it wanted to apply them to a segregated account. Mr Grunditz said that he read them, but that he “didn't have any views on adding or deleting anyone from there”.

Although it did not in the event go into the Pennsylvania Fund, preferring a segregated account, the investment guidelines which AP went on to adopt were much the same. There were some changes, and in her statement, Ms Thomasson-Blomquist said that they made the guidelines more conservative. This assertion requires careful appraisal.

The changes made reduced the weighted average maturity from 45 to 30 days, inserted a limit of US$1.5bn on the cash collateral programme (which was removed in January 2005), and deleted money market funds as permitted investments. An overall limit of US$7bn on lending was imposed, which was about 30% of AP’s total assets at the time (it was increased to US$10bn in May 2007).

I agree with BNYM that so far as investments were concerned, the changes that AP made went primarily to interest-rate risk, and not credit risk. This was something which Ms Thomasson-Blomquist largely accepted in cross-examination. The deletion of money market funds, which was the change made to permitted investments, also went to interest, since they were relatively low yielding.

Further, as Ms Hammarlund accepted, the reason for the segregated account was not to do with any supposed conservative appetite as regards investments. It was to protect AP in the event of the failure of BNYM (in other words the institution holding the collateral). There is no evidence that AP selected a segregated account because of concerns over the credit risk of investments permitted in the commingled funds.

Further, although AP has emphasised that it was one of very few legacy Mellon clients which had a complete restriction on illiquid securities, the restriction came from the Pennsylvania Fund guidelines. It also has to be seen against the evidence of Mr Mannix, who said that the bank’s practice was that it would not purchase any security unless it was deemed liquid at the time of purchase. (He came from the BoNY side of the merger.)

AP’s witnesses suggest that it entered the SLAA in the expectation that so far as the cash collateral reinvestment securities were concerned, there would be little risk to principal. I accept that, and it is reflected in BNYM’s material sent in the tender process quoted above talking of “preservation of principal”, as well as “safeguarding principal” in the agreed objectives. I am satisfied that AP had this expectation in common with all funds that permit securities lending.

As to what the term “safeguarding principal” means, AP relied on certain bank documents (and the comments of the witnesses on them) as showing that BNYM would select only those investments which provided a “near certainty of full principal repayment”. An example comes from the cross-examination of Mr Mannix, who agreed that the idea of preservation of principal is that the investments that are purchased, at the time of purchase, are deemed to represent a near certainty of payment. As Ms Demmler put it, were the investments “… essentially what we call ‘money good’. Were they going to pay out at maturity".

But as Mr Fort said, "Yes, the expectation is that there is little to no risk. Again, as I stated earlier, it doesn't mean that there is a guarantee". AP criticised Mr Fort as a witness who stonewalled, prevaricated and dissembled. I do not accept that, and consider him to have been competent and professional (though as appears later I have not accepted part of his evidence where it is inconsistent with the contemporary material). It is important in my view to be clear that the use of phrases such as “near certainty” did not amount to a guarantee by BNYM that the securities in which the cash collateral was invested would never default, nor would they have been understood by prospective lenders as such. That would have been a different transaction from the one the parties entered into. AP does not suggest the contrary, but its emphasis on the language of “near certainty of repayment” at times has come close.

As the bank’s presentation to AP in May 2004 put it, “No lender would ever state they offer an un-conservative cash collateral re-investment programme”. But at the same time, it is obviously not possible to achieve a return without risk. I am satisfied that AP was fully aware that embarking on a programme that included for the first time the investment of cash collateral was not a risk-free course of action. The truly conservative course was to avoid securities lending altogether, or to avoid the reinvestment of cash collateral, either of which courses was open to AP. Ms Hammarlund realised, she said, that its choice meant more risk—more risk that is than in AP’s previous programme with Citibank.

In the schedule to its closing submissions, AP did not in the event invite any finding to the effect that it was a particularly conservative and risk-averse securities lending client. I would not have made such a finding. The evidence is that securities lending clients like pension funds are conservative by nature, but there is no evidence that AP was any more conservative than others. AP’s witnesses were clearly dedicated to the well-being of the fund, but the fact is that they were prepared to lend out a substantial part of their portfolio on terms which were not materially different from those of the commingled fund with which they had been provided by the bank. They did so for what I infer is the same reason that other funds enter into the activity, namely because it provided some extra return, at a risk that was perceived to be low, but a risk nevertheless.

Were the Sigma MTNs a “suitably conservative investment” as collateral?

AP maintains the Sigma MTNs were not a “suitably conservative investment” as collateral for AP. In this respect, it relies principally on the evidence of Mr Servis. Among the points he takes are that SIVs as a form of securities were “among the most complex structured finance vehicles ever devised”. He draws attention to the fact that Sigma’s business model relied on it rolling over its debt in order to redeem existing liabilities. As regards the underlying asset portfolio, he draws attention to the lack of transparency, and the risk of subordination in relation to repurchase transactions. He also draws attention to the vulnerability of the funding model, if a SIV were to lose its AAA rating. A further point is that Sigma, unlike some other SIVs, lacked the support of a financial sponsor, such as Citibank or HSBC.

Despite BNYM’s submissions to the contrary, viewed from today’s perspective, I consider that most of these points are in substance correct. Although it is straightforward to describe the basic model of a SIV such as Sigma, on analysis the package for an investor was a complex one. The most obvious example concerns the terms of the Private Placement Memorandum, and the nature of the right of recourse to the portfolio of assets, as is shown by the difference of opinion in the courts about how Sigma’s assets were to be distributed after the appointment of the receivers (see Re Sigma Finance Corp [2009] UKSC 2). (I do not however agree that SIVs were “among the most complex structured finance vehicles ever devised”.)

Further, although the credit rating agencies apparently had CUSIP level information about the securities in the portfolio (the CUSIP is the code which identifies a specific security), investors acquiring Sigma MTNs only knew about asset classes. The Sigma MTNs and this form of investment vehicle generally were (in substance) securities in securities, the identity of which was unknown to the investor, and which themselves included securities in securities, giving rise to a further level of non-transparency. They were, in a word, opaque.

Further, the subordination of senior creditors to repo counterparties proved in the end to be a major problem. Finally, I am satisfied that (as Mr Servis says) the model depended on the AAA rating, partly to make it acceptable to investors like money market funds, and partly because the AAA rating enabled Sigma to offer a comparatively low interest rate upon which it founded its arbitrage. I doubt whether Sigma was especially vulnerable to downgrades engendering a "cliff effect" plunging the security straight to junk, as AP suggested. But it does not matter, because I accept the importance of the AAA rating to its business model.

The ratings point merits some further attention. Mr Jonathan Dawid, junior counsel for BNYM, took me through Moody’s Rating Methodology for International Structured Finance dated 28 January 2004, which is focused on SIVs as an asset class. The Moody’s model involves sophisticated mathematics, and as he pointed out, included a scenario in which the SIV did not issue new debt, but was obliged to rely on its asset portfolio to fund maturities. The implication was that if even the credit rating agencies’ model did not predict a loss for investors on most scenarios, there was no reason to criticise BNYM for having reached the same conclusion.

In that regard, BNYM also placed reliance on a Standard & Poor’s report of 12 September 2008, to the effect that it had considered five different scenarios, the first assuming a continuation of Sigma’s performance to date and the other four assuming progressively more stressed environments. In the first three scenarios, S&P’s modelling found that Sigma’s MTNs would be repaid in full. The fourth predicted recoveries in excess of 90% of par. Only the last, which involved Sigma selling assets at an accelerated rate to pay debts before they fell due, involved a recovery of below 90%.

The tone of what is a brief report is positive (perhaps not surprisingly since, though the ratings were lowered that day, they were still A ratings). AP makes various criticisms of the methodology which I need not resolve. Suffice it to say that in my view, it is very surprising that Sigma still had investment grade ratings at this time, since for reasons I have explained elsewhere, by now it was hugely vulnerable. Lehman Brothers collapsed three days after the report was published, but so much of its portfolio was tied up in repos by now, that a much lesser shock would have been sufficient to cause its demise. This report has no credibility in my view, and I do not think it carried any weight with BNYM at the time.

Inevitably, the well-publicised flaws in the ratings given by the credit rating agencies as laid bare by the financial crisis cast a shadow over the AAA rating given to structured investment vehicles such as Sigma, and cast doubt on the mathematical models used to justify them. Despite its submissions to the contrary, the evidence suggests that BNYM itself lost faith in their reliability long before September 2008. A revealing email sent on 13 March 2008 by Mr Timothy Robison, who was the chief operational risk officer at BNYM, and I infer not only senior but close to the market, says that, "The rating agencies have been useless throughout this structured finance fiasco from the beginning. Nice to see they are keeping that record intact." At that time, Sigma was still rated AAA, despite the fact that it had been effectively defunct for months, and was using various techniques to raise money on its assets to meet maturities as they fell due.

BNYM was certainly not alone in this view. At the beginning of April 2008, Sigma finally lost its AAA status. A piece from an analyst at JP Morgan (who Ms Demmler thought was “usually pretty on-point”) dated 7 April 2008 said that he thought that the late downgrade of Sigma “says much more about Moody’s than it does about Sigma”. Investors were “painfully aware” of the problems, including reliance on repurchase agreements: “For many of them, the rating agencies lost credibility on this name long ago”.

However, so far as AP’s claim relates to the purchase of the Sigma securities for its account, the question is how matters looked in March/April 2007. AP cites a report published by the FDIC in its Supervisory Insights on 30 May 2008, which says that “Two significant factors in the recent market turmoil have been the over-reliance on credit ratings and a misunderstanding of what those ratings mean”. But if, at the time of acquisition the previous year and before the credit crunch blighted these securities, BNYM did place over-reliance on the AAA rating, this report is an indication that it was in the company of others in the market.

In fact, the evidence was that BNYM did not place exclusive reliance on the ratings. There were other factors that recommended the issuer to BNYM as well. Sigma was the largest single SIV, and was seen at the time to be particularly well run. BNYM’s witnesses gave evidence as to the standing of Sigma prior to the onset of the crisis. Mr Mannix described it as “… a very popular investment, well respected in the market places and carried the highest credit ratings. It was the most reliable SIV, as it was the largest and the oldest, and had the right people running it”. Ms Demmler, who was responsible primarily for Sigma being added to the list of Approved Issuers, said that “Sigma was well-known and highly regarded. It was one of the most widely held programs in the market”. Mr Ford said that Sigma “… was perceived as the most successful and most highly regarded of any similar entity” because it was better capitalised, and more conservatively managed. Mr Fort said that “Sigma was one of the most well respected issuers in the market for … securities lending-enhanced cash investors”. Mr Tant described Sigma as “a well established operator with a good track record”.

These were experienced bank officers, I have heard them give evidence at some length, and I accept it in this respect. As Mr Tant (who was head of the bank’s Credit Analyst Group at the material time) said, he was aware that there were certain risk factors specific to Sigma and SIVs generally in 2005—they were buying longer duration assets and arbitraging shorter duration funding. That he acknowledged created a certain risk element, but it was also what created the return. This is, in my view, a fair description. He added that in the case of Sigma, the way the vehicle was structured in terms of the capital and liquidity and how those factors were managed was considered to be “particularly robust”. Again, I accept this evidence.

It is borne out by the evidence as to the nature of the parties that were acquiring the securities at this time. It is not in dispute that on the eve of the credit crunch, Sigma, and SIVs generally, were widely held both by securities lending agents and by money market funds. According to Mr Richard’s evidence, a 2005 report by JP Morgan identified that 70% of investors in SIV senior debt were money market funds, with another 20% represented by securities lending cash collateral investments.

The defendant produced a schedule that suggested that at the time of Sigma’s default, a number of well known securities lending agents were still holding Sigma MTNs, and Mr Fort gave evidence as to others. As described above, Mellon Bank added Sigma to its list of Approved Issuers in 2005. It is also significant in my view that Bank of New York, itself a major custodian, had independently approved Sigma for its securities lending clients, and had acquired a large holding. AP’s case implies that both Mellon Bank and Bank of New York were negligent at the time in approving Sigma.

AP appeared to accept that it is correct that many money market funds purchased Sigma MTNs. It also, I think, accepted that they are conservative funds, since investors treat them as a pool of liquidity, and it is essential that the net asset value (NAV) of the funds does not fall below one dollar per share (known as “breaking the buck”).

AP gives a number of reasons why this does not show that the Sigma MTNs securities were suitable for conservative investors. It points to the fact that the funds were permitted to have 5% of their investments in illiquid securities, and that after the onset of the crisis, sponsors entered into support agreements to ensure that investors did not suffer loses. These points may be correct in themselves, but the fact is that money market funds acquired large holdings of the SIV issues prior to the onset of the crisis. Mr Richard, who makes this point, also says that Sigma’s portfolio of assets was reputed to be of high quality, diversified and contained minimal exposure to US sub-prime mortgage securitisations.

The global financial crisis was a seismic event, and in judging investment decisions made before its onset, it is important to avoid the benefit of hindsight. As BNYM says, the indisputable fact is that Sigma MTNs were widely purchased and regarded by the market as suitable instruments for cash collateral reinvestment. On the above evidence, I am satisfied that at the time it acquired the Sigma MTNs for AP’s account, it was reasonable for BNYM to consider the Sigma MTNs as investments appropriate for AP. Further, at the time of acquisition, I find that these securities satisfied the “safeguarding principal” objective in the Investment Guidelines.

Does “suitability” apply, since the parties agreed Investment Guidelines?

For reasons given elsewhere, I have rejected AP’s breach of mandate case to the effect that “only investments which when purchased had a near certainty of payment were within the mandate and objectives agreed”.

However, in its written closings, BNYM sought to take its submissions in this regard one step further, contending that these clauses “define completely” the level of credit risk which AP was willing to accept. It submitted that the effect of the Investment Guidelines was that BNYM is granted a discretion, with the extent of that discretion being delimited by the guidelines.

In the course of argument on this point, Mr Ali Malek QC, leading counsel for AP, asked hypothetically whether BNYM would have been entitled to buy Sigma MTNs in (for example) early 2008, at which point they were still rated AAA, and fell within the guidelines. The answer is obviously not—by that time Sigma had ceased altogether to issue new debt in the wake of the credit crunch. In its written submissions BNYM accepted this, but on the basis that by early 2008, the securities had become illiquid, and so fell outside the criteria in section 4h) of the Investment Guidelines which would have applied to the notional acquisition.

This in my view is only a partial answer to AP’s hypothetical question. Even if the Sigma securities had been technically liquid in early 2008, a prudent securities lending agent would not have acquired them for the account of AP. A purchase might have been within BNYM’s authority, in the sense that the security complied with the guidelines, but nonetheless have resulted in a contravention of the contractual standard of care in clause 10(b)(i) SLAA. (This, it is to be emphasised, is an entirely different question to whether, having properly acquired the securities, the bank should have sold them then, or could legitimately have taken the view that it was better to continue to hold them to maturity. The present discussion has to do with the bank’s duties on acquisition.)

In oral closing, Mr Mark Hapgood QC, leading counsel for BNYM, put the bank’s contention differently. He submitted that if a type of security is within an asset class expressly permitted by investment guidelines, there can be no argument about the suitability of that asset type. In the present case, he says, any argument that SIV medium term notes were unsuitable for AP’s portfolio is hopeless. The only ground on which the purchase of a permitted type of investment can be challenged is if the bank fails to use reasonable skill and care to select an appropriate asset of that type.

I agree with that submission. In the securities lending field, as in the case of other banking activity, it is necessary to distinguish between a bank’s mandate, which defines the authority given to it by its customer, and the standard of care which it must observe in acting pursuant to that authority, whether arising under the terms of a contract, or in tort. An acquisition of securities which falls within the parameters of the customer’s investment guidelines is an authorised investment made in accordance with the bank’s mandate. As asset-backed securities rated AAA, the Sigma MTNs fell into that category. It cannot be challenged on grounds of lack of authority.

However, in making the investment, the bank remained subject to its duty of care, in this case as defined in clause 10(b)(i) SLAA, which required it to perform its obligations as Lending Agent with due care, skill, prudence and diligence. AP’s expert evidence made general criticisms of SIVs as an asset class. But in agreement with BNYM, I do not think that the duty of care can be invoked in support of an argument that SIVs as such were unsuitable for AP. It had expressly mandated BNYM to acquire AAA asset backed securities, and SIVs were such securities. The issue is whether there was some aspect of the Sigma MTNs which made the acquisition of those particular securities a breach of duty.

There is only one credible candidate in that respect. AP submitted that while some of the SIVs were “sponsored” by banks, Sigma was not sponsored. The question arises whether, on the basis of this distinction, the Sigma MTNs were unsuitable as collateral albeit within the Investment Guidelines.

It is necessary to appreciate that bank “sponsorship” did not amount to a guarantee, and there was at least one instance of a bank allowing its SIV to fail (Standard Chartered in the case of Whistlejacket). However, Sigma was the largest single SIV, and was seen prior to the onset of the crisis to be well run. So while I accept AP’s submission that the fact that it did not have a bank sponsor was (contrary to BNYM’s case) a disadvantage, I find that the overall credibility of the issuer made up for that. This is demonstrated by the widespread holdings of the Sigma securities which I have described above.

Further, the evidence is that it was perceived by the market that Sigma had an advantage in that (unlike the SIVs that were taken over by their bank sponsors) it had no market value or ratings-based enforcement triggers. These had been removed in 2003 in order to give Sigma more flexibility to respond to difficult market conditions and avoid it being forced to liquidate assets at distressed prices. This is one of a number of attributes cited by Mr Richard which he says compare favourably with other issuers in the SIV sector. On the evidence, I am satisfied that the lack of bank sponsorship did not make it unreasonable for BNYM to consider the Sigma MTNs as investments appropriate for AP.

AP’s breach of mandate case has other limbs to it as follows.

AP says that to the extent that the acquisition of the Sigma MTNs was part of a programme approach which in effect joined AP's investment in Sigma MTNs to the aggregate position of all of the other securities lending clients of BNYM which held Sigma MTNs, this was not authorised under the SLAA and was in breach of AP's key objectives "to keep all cash collateral and related investments in a segregated account in the name of [AP]" and "to safeguard principal".

I have dealt with the commingled fund /programme approach issue above, and refer to what is said there. But given my conclusion that the Sigma MTNs were suitable investments for AP, it does not matter whether they were allocated to its account applying a “programme approach”. There was no breach of mandate and no negligence.

AP says separately that there was a breach of mandate in that BNYM had no authority under the SLAA to include AP as effectively part of a "commingled" fund or programme approach for purchase of Sigma MTNs as the aggregate position resulted in AP's Sigma MTNs being "illiquid securities" within clause 4h) even if they were not when viewed in isolation.

In my view, this is an artificial point. It is not in dispute that AP’s holding could have been sold. In fact, this is a plank in AP’s second and third heads of claim. There was never any question of selling BNYM’s entire holding (or all of Mellon’s holding), even if such a sale would have been possible. There is no basis for arguing that the securities were illiquid within the meaning of the AP Investment Guidelines on this ground.

AP has an allied point, saying that there were no formal or documented internal concentration limits applied by the Credit Analyst Group within Mellon Bank, and in this the bank’s systems were seriously deficient. It says that had Mellon “applied appropriate concentration limits to the Sigma MTNs, from the date of the Initial Review (or the date on which the Sigma MTNs were admitted to the Approved List) then the Sigma MTNs would not have been purchased for AP’s Account at the dates of acquisition”.

I reject this contention. I accept Mr Tant’s evidence that Mellon did apply an internal limit 10% of total issued debt, and the evidence of Mr Richard and Mr Zimmerhansl that this was a reasonable limit. In substance, it was not in breach of that limit. As was put to Mr Servis:

A breach of mandate case is also based on the liquidity requirement in the Investment Guidelines. Section 4h) of the guidelines provided that, “The Account may not be invested in illiquid securities. For the purposes hereof, an illiquid security is defined to be an issue that is not sellable without the sale proceeds being materially impacted due to a lack of other trading activity in the security. …”.

AP contends that the Sigma securities were illiquid, and that their acquisition on its account was consequently in breach of mandate, and the bank’s contractual duty of care, given its key objective of safeguarding principal. It also contends that BNYM failed to assess the liquidity of the securities against the requirement in its guidelines.

In support, AP contends that as it and BNYM both understood at the time of the SLAA, liquidity is a significant risk characteristic of an investment. The purpose of a prohibition on illiquid investments in a conservative portfolio is to safeguard the capital in the account by prohibiting the acquisition of securities which cannot easily and quickly be sold and converted to cash without loss, particularly in times of market volatility and stress. An investment is typically considered by the market to be a liquid security, AP says, if it is able to be sold in the market at short notice and with minimal price impact (i.e. discount). The importance of holding liquid investments is particularly acute in the context of management of cash collateral accounts because of the well understood feature of the industry that meant that there may be a need to liquidate, quickly, any position held in order to return borrowed securities at short notice.

BNYM’s case is that the Sigma MTNs were liquid securities, within the meaning of section 4h) of the Investment Guidelines, at the time of purchase for AP in March and April 2007. It accepts that the securities became illiquid following the onset of the credit crunch later in the year.

I have described above the circumstances in which the Investment Guidelines were agreed, and though AP says that it was one of very few legacy Mellon clients which had a complete restriction on illiquid securities, there is no evidence that it placed any particular emphasis on liquidity. The same provision was contained in the Mellon GSL Pennsylvania Fund. Nor does it need to show such evidence. Unless the Sigma MTNs satisfied the liquidity requirement at the time, their acquisition for AP’s account was a breach of mandate.

So far as I am aware, the term “illiquid securities” has no fixed legal meaning. For present purposes, what matters is the contractual definition in section 4h) of the Investment Guidelines (see above). AP cited what I found to be a helpful comparison from the SEC rules applying to money market funds, which in Rule 2a-7 define illiquid securities as "securities that cannot be sold or disposed of in the ordinary course of business within seven days at approximately the value ascribed to them by the money market fund". On this basis, the idea behind the concept of liquidity in this context is not merely that the securities can be disposed of, in the sense that there is a market for them, but that the price on sale will approximate to their ascribed (or mark to market) value.

AP makes a number of points in support of its case, which I deal with below, but BNYM makes a particular point, which it says, if correct, decides the issue. AP’s case, it says, is based on the lack of a secondary market for Sigma MTNs. Whilst BNYM disputes that there was such a lack, it submits that the size of the primary market for these securities at the time of acquisition itself satisfies the liquidity requirement.

BNYM submits that prima facie evidence of such liquidity comes from the sheer amount of Sigma’s issuance, which in 2007 stood at some US$55 billion with a weighted-average life of just over a year, meaning it was issuing debt at a rate of some US$50 billion per year. This factual point is not as I understand it in dispute.

There was, therefore, BNYM submits, clearly an enormous demand for Sigma debt in the market. While most of that demand would have been satisfied by new issues, from a buyer’s perspective there is no distinction in terms of security or coupon between a newly issued note and one offered on the secondary market, and it is common ground that the same dealers acted both as primary placement agents and secondary market brokers. To suggest that notes for which there was such demand in the marketplace were illiquid is, it submits, absurd.

A number of points are raised by AP to displace this conclusion. Mr Cirrito’s evidence was that the MTNs were illiquid, because they were private placements issued under Securities & Exchange Commission Rule 144A. This is a rule that restricts certain securities from being purchased other than by qualified institutional buyers (QIBs). However, any support that this point may have given for AP’s case was lost when in cross-examination Mr Servis did not agree with it. None of the other experts agreed with it, and given the number and purchasing power of QIBs it is far from self evident that this restriction reduces liquidity, and I do not accept it.

AP was on stronger ground in its submission that there was not an active secondary market for the MTNs. Mr Richard identified five sales of Sigma MTNs by BNYM between October 2006 and 30 April 2007 totalling US$544m executed at prices between 99.98 and 100.02 (in other words at par). In a second list, he identified 16 transactions over the 2 years and 3 months between May 2005 and August 2007. BNYM also disclosed bid lists relating to sales of Sigma MTNs in the period January-May 2007, which show that when it sought to sell Sigma notes for clients in this period, it received multiple bids close to or exceeding the original purchase price.

There was some significance in the latter point, because while other participants in the market would not know the actual price agreed for the trade, they would see the “cover”, i.e. the second-highest bid, which gave at least a degree of transparency to Sigma pricing. However, in my view, although Mr Richard expressed the opinion that the secondary market for Sigma securities was a liquid market, the figures he produced suggest that Sigma MTNs were thinly albeit regularly traded on the secondary market.

I infer that this was because (as the evidence also showed) purchasers bought them to hold to maturity. In fact, AP accepts (subject to its other submissions) that it “might have been a reasonable decision to acquire the Sigma MTNs in March and April 2007 on the basis of a ‘hold to maturity’ policy”. These securities were not acquired with a view to secondary trading. Such evidence as there is does however support the view that the securities were sellable without the sale proceeds being materially impacted due to a lack of other trading activity. That would mean that they were liquid within the contractual definition.

AP contends that the evidence shows that no-one assessed the suitability of AP's Sigma MTNs position against its requirement not to be invested in illiquid securities under section 4h), either at acquisition, or at any stage during the period after acquisition and up to default in October 2008. I do not agree with that contention. It is accepted by BNYM that the liquidity of the MTNs was compromised because of the credit crunch. After that happened, the status of the Sigma securities was (as I describe below) kept under close review by BNYM.

AP also contends that the section 4h) restriction does not appear to have been coded in to BNYM's systems as a restriction mandated by AP's guidelines. Instead, BNYM coded in for all clients a minimum overnight liquidity of 20%. The four annual compliance reports dated 28 February 2005, 31 January 2006, 31 January 2007 and 31 January 2008 show, AP says, that BNYM had not coded anything into its systems for the section 4h) restriction.

This contention may be factually correct, in that there is no mention of section 4h) in the compliance reports. But this does not advance AP’s case, because I have concluded for reasons set out above that the section 4h) restriction applied at the time of acquisition, but not on a continuing basis. If anything, the fact that it was not coded in to BNYM's systems tends to support that conclusion.

There is some support for the view that BNYM itself regarded SIVs generally as liquid in March/April 2007, in the sense that the position changed at the onset of the credit crunch. Slides of a Risk Committee Meeting held on 8 October 2007 under the heading “Impact of August/September 2007 Market Conditions” stated, "Investments in SIVs, short duration AAA rated MBS and ABCP declined in market value and became virtually illiquid”.

On balance, I accept BNYM’s case as to liquidity. Mr Richard’s evidence is that over the period 2006–7, the amount of SIV debt issued increased very substantially. As I have said, in retrospect, this can be seen as part of the bubble in asset-backed securities that ultimately burst. However, there was plainly a strong demand for Sigma debt at the time that BNYM acquired it for AP’s account. On balance, I agree with BNYM that the strength of the primary market is sufficient to establish liquidity.

So far as the secondary market is concerned, I consider that the evidence justifies the conclusion that had a holder sought to sell even substantial amounts of Sigma MTNs into the market in March or April 2007, the sale proceeds would not have been materially impacted due to a lack of other trading activity in the security. That is the liquidity requirement as defined in the Investment Guidelines, and I consider that it was satisfied. Even if Mr Servis is right to say that much of the secondary liquidity was supplied by Sigma itself, this does not affect the conclusion.

While (as is clear from the slide I have quoted) BNYM accepts that the securities became illiquid relatively shortly after acquisition, I find that this was because of the exceptional conditions associated with the onset of the credit crunch.

The remaining limb of AP’s breach of mandate case is as to the limit of 40% placed on investments in floating rate securities. Section 4e) of the Investment Guidelines provided that “No more than forty percent of the Account may be invested in floating rate … securities” (such as floating rate notes or FRNs). The Sigma MTNs fell into this category.

AP’s case is that the acquisition of the US$30m Sigma MTNs maturing 30 October 2008 for AP’s account in April 2007 breached that requirement. For reasons I shall explain, the issue depends on whether the notes were acquired on 26 April 2007 (as BNYM submits) or 30 April 2007 (as AP submits). It is not in dispute that if AP is correct, the purchase was in breach of mandate, and BNYM must make good AP’s losses in respect of the notes.

The parties agree that the relevant date is the date of purchase of the AP Sigma MTNs for AP. The bank’s total purchase in this transaction was US$300m. The purchase was made via Lehman Brothers, and the trade ticket shows the date and time of the trade, namely 26 April 2007 at 16.56 pm. BNYM says, and I accept, that the trade details were incorrectly entered on the bank’s systems, showing the counterparty as JP Morgan and the trade date as 30 April 2007 and the time as 10.17 am. (The document also suggests that the issue date was 30 April 2007.)

The documents AP relies on are as follows. The AP Holdings Report shows both purchase and settlement dates as 30 April 2007. In the document allocating the securities to AP and sixteen other clients, the settlement and trade dates are 30 April 2007. AP submits that these are the documents that show the allocation and effective purchase date for each of the clients to whom the US$300m bulk order was allocated. This shows conclusively, AP says, that the securities were purchased for AP on 30 April 2007. The trade ticket, it submits, concerns only the aggregate order.

The point depends on the total of floating rate securities in AP’s account on the two dates. In this regard, BNYM says (and I do not think that this is in dispute) that on 26 April 2006 its total FRN concentration was 38.33%. Its total portfolio size at that time was US$2.066bn. The impact on AP’s account of purchasing US$30m of Sigma MTNs was to increase the total FRN concentration from 38.33% to 39.78%, which was within the 40% limit.

What happened according to BNYM (again not I think disputed) is that by 30 April 2007, the overall size of AP’s account had decreased slightly which resulted in the relative proportion of FRNs increasing to 41.4% on that day. The nature of securities lending is such, it says, that it is impossible to predict in advance what the size of a portfolio will be on any day. That is because the amount of cash collateral to be invested is determined by the value of securities that the client has on loan, and at any time a borrower may return a security or the client may recall it. I accept this analysis as factually correct.

AP’s portfolio manager at the time (Mr Julian Anantarow who was based in Pittsburgh) did not give evidence, but there was some explanatory evidence. Mr Mannix said that if a security is appropriate, the relevant portfolio managers will have discussions regarding the purchase of amounts for their portfolios. All of the orders are then combined, and one large block order is placed with the entity that is offering the security. Once made, the purchase is allocated across individual portfolios.

Mr Fort agreed, saying that the process at Mellon was fundamentally the same prior to the merger with BoNY. The securities lending cash reinvestment team held a daily conference call at about 8.30am every morning. During these calls, they discussed opportunities for investments. Decisions as to what securities should be purchased would be made by the entire securities lending cash reinvestment team jointly. Following the call, and again as a team, they executed orders to buy securities and allocated these across the relevant accounts.

Their evidence, therefore, is that the bank’s practice is to allocate securities at the time of the purchase order. BNYM says that it would make no sense to put in an order for US$300 million of securities without knowing in advance what could be allocated to each client. There is in fact a manuscript note on one of the bulk purchase tickets which is consistent with this evidence, but without explanation from the maker, who appears to have been Mr Anantarow, I cannot give it any weight. However, in view of the other evidence, I find that it is likely that the decision to allocate US$30m to AP took place at the time of the bulk purchase on 26 April 2007, and not on 30 April 2007.

AP’s case is based on the allocation date as shown in the documents, and has considerable force. But I consider that BNYM’s case makes more commercial sense. This is a breach of mandate claim, so that in effect absolute liability attaches to the bank if it gets it wrong. The bank has to know at the time of the bulk purchase whether the amount it is contemplating purchasing can be allocated among its clients, and this can only practically be done by reference to the percentage limits as they are at that time. On the facts of the present case, I consider that the date on which the MTNs were acquired and allocated to AP’s account was the bulk purchase date, that is 26 April 2007, and that this was the relevant date for the purpose of assessing compliance with section 4e) of the Investment Guidelines. Accordingly I reject the asserted breach of mandate on this ground.

At trial, AP did not develop a late contention added by amendment that the purchase of the Sigma MTNs for AP as part of the programme of nearly US$3bn Sigma MTNs contravened the bank’s duty of best execution to AP under COB 7.5 in that the bank would not, on its own case, be in a position to obtain the best price available for AP’s tranche of US$35mm Sigma MTNs on a sale if those Sigma MTNs had to be sold as part of a much larger US$3bn aggregate exposure.

It was right not to do so, since the point is not arguable. The duty of best execution has to do with the mechanics of acquiring or selling securities, not the merits or otherwise of the trade. AP’s contention is in effect a suitability point. As BNYM says, the duty of best execution is a duty that, by definition, applies only on the execution of a client order. It has nothing to do with the underlying investment decision.

For reasons stated above and elsewhere in this judgment, I reject each of the components of AP’s case. I do not consider that BNYM owed AP fiduciary duties, nor would it have made any difference if it did. In summary, I find that there was nothing wrong with the acquisition of the Sigma securities for AP’s account. They were seen as appropriate cash collateral investments for securities lending programmes at the time, including for AP. The problems emerged afterwards with the onset of the credit crunch.

AP’S CLAIMS RELATING TO THE RETENTION OF THE SIGMA MTNs

AP’s case in respect of acquisition as set out in one of the schedules attached to its closing submissions is (in summary) that:

For reasons set out above, I have rejected submissions (1), (2) and (6). Those aside, the essence of AP’s case is that BNYM should have sold the securities, or should have informed AP as to the risk of default, and should not have preferred other clients over AP in that regard.

BNYM’s case is that it reasonably believed, at all times, that the most advantageous course for clients such as AP was to continue to hold the Sigma MTNs, given that (i) the most likely outcome was considered to be that the notes would be repaid in full; (ii) even in the unlikely event of a default, analysis indicated that noteholders would recover at or close to par; and (iii) the illiquidity of the credit markets meant that market prices available for Sigma MTNs were well below that justified by the quality of Sigma's assets and the expected recovery on the MTNs if held to maturity.

Over the litigation, AP has put its case as to when the securities should have been sold in different ways. The originally pleaded case was that the notes should not have been retained after January 2008. At the beginning of the trial, there was some emphasis on the view of Mr Servis that the sale should have happened when Citigroup bailed out its comparable SIVs in December 2007. In oral closings, AP suggested that the sale should have been in the autumn or late summer of 2007. In written closings it is said that, “In conclusion, AP contends that by January/February, it was clear that AP's Sigma notes should have been sold”. In the schedule to the closings, as I have said, AP seeks a finding that BNYM should have sold the MTNs by 30 September 2007, or 29 February 2008, or May 2008 or June to August 2008.

This issue as to timing raises the question as to what price AP’s Sigma MTNs could have been sold at the various times which AP contends for. A useful document in this regard is Schedule 2 to the Re-amended Particulars of Claim. This is a table showing a summary of AP's best estimate of the prices for which its holding of Sigma MTNs could have been sold in the period from August 2007 to September 2008. Although it is not formally agreed, AP’s reliance on it was not challenged by BNYM, and I do not understand the figures to be disputed. They were effectively endorsed by Mr Richard and I accept them. (For the avoidance of doubt, it is not a contemporary document, but an extrapolation from available data for trial purposes.)

It shows that while AP’s holding of securities could have been sold without an appreciable loss on par (US$35.5m) up to 31 October 2007, thereafter the price declined showing a loss of about US$11m by the end of April 2008. Thereafter the price recovered considerably, before Sigma’s final default in September. BNYM suggested that this was because of returning confidence in Sigma. It is correct that there is some indication of more bid activity, but the most likely cause in my view is that the risk was perceived to decline the closer it got to maturity. The last available number is US$29,775,000 on 31 August 2008. It is common ground that prices would have been much lower had BNYM sold (or tried to sell) its entire holdings (or much larger blocks) had that been possible at all.

BNYM criticises what it says is AP’s failure to provide a comparator price, that is, what AP contends would have been the expected recovery on the MTNs if held to maturity at the various times at which it says they should have been disposed of. There is some force in this, although expected recovery is something which the market would be expected to price in, along with the possibility of default. But it does not arise as a major issue on the facts as I have found them.

BNYM’s closing submissions contain a useful analysis of the key metrics as regards the Sigma MTNs as seen from BNYM’s perspective. This highlights Sigma’s monthly business reports, the price available for Sigma MTNs on the market, and their ratings. On the basis of such information, BNYM submits that a reasonable securities lending agent in its position could properly have concluded that a sale at the depressed market prices being offered (whether on a single client or programme-wide basis) would represent a worse outcome than that expected from continuing to hold the securities.

It is to be noted however, that although he gave indicative prices, Mr Richard’s view was that a sale on a programme-wide basis (in other words of BNYM’s entire client holding or even a large part of it) would have been impossible in the market conditions that prevailed after the onset of the credit crunch. I accept that evidence, and reject BNYM’s submission that the fact it continued holding the Sigma MTNs, not only for AP, but for “the vast majority of clients … is the strongest evidence there is that the Bank’s overall view, at all times, was that the MTNs would be repaid in full”. Its options in that regard were limited.

A significant aspect of AP’s case is that BNYM’s credit monitoring was defective. AP says that the credit analysis carried out by the Credit Analyst Group (and Ms Demmler in particular) was very limited indeed. Ms Demmler performed no liquidation analysis for Sigma MTNs at any stage. Such credit analysis as was carried out did not make realistic assumptions about key variables. This was because, AP says, the credit analysts did not understand the impact of repo collateral leaving the Sigma vehicle, and was far below the standard to be expected of a securities lender such as BNYM.

Had BNYM’s credit analysts understood the impact of the repo financing, AP says, and carried out a proper liquidation analysis using realistic assumptions for repo haircuts and for mark to market prices on the Sigma underlying asset portfolio, and made reasonable assumptions to consider what would happen if falling prices in the market caused repo counterparties to trigger margin calls, the analysis would have shown risks of serious losses given default (LGD) to Sigma MTN holders from May to September 2008.

The particular point made by AP, therefore, goes to the effect of the use by Sigma of repo financing. I have noted above the increased use of repos as a source of funding after the onset of the credit crunch. Based on the evidence of Mr Servis, which was itself based on Sigma’s monthly business reports, AP relied on modelling which was the subject of much discussion at trial.

It is necessary in my view to look back at the position at the outset of the credit crunch. Sigma’s repo funding position stood at US$980 million as at 30 April 2007, which was approximately 2% of its liabilities, in other words negligible. By 29 August 2008, the position had altered completely, and Sigma’s repo funding had reached US$17 billion, which was approximately 60% of its liabilities. AP says that:

I accept these figures, and though BNYM downplayed them, they show clearly how repos expanded as a source of funding for Sigma over the relevant period. Some of BNYM’s witnesses referred to repos as a “non-traditional” source of funding, but that is not in my view an apt description. Repos were used because access to funding via the CP and MTN markets which Fitch in September 2007 had described as “critical for SIVs to maintain normal operations” (see above) had dried up.

In view of the way the argument progressed, it is necessary to say something about the nature of repos, which is relevant in one particular respect. In simple terms, a repo is a transaction in which one party sells an asset (such as fixed-income securities) to another party at one price, and commits to repurchase the asset at a different price in the future. Although a repo is structured legally as a sale and repurchase of the securities, it behaves economically like a secured loan, with the securities acting as collateral (see e.g. DCC Holdings (UK) Ltd v Revenue and Customs Commissioners [2011] 1 W.L.R. 44, SC; In the Matter of Lehman Brothers International (Europe) (in Administration) [2010] EWHC 2914 (Ch) at [79], Briggs J; and the definition in Directive 2002/47/EC of 6 June 2002 on financial collateral arrangements, Art 2).

The collateralisation is by way of haircuts/initial margins, which are usually set such that the purchase price of a repo is less than the market value of the collateral—that is, they typically result in over-collateralization. This is taken from a paper cited by BNYM, Haircuts and initial margins in the repo market, European Repo Council, by Richard Comotto, published by ICMA on 8 February 2012. (There is no need in the present context to distinguish between haircuts and initial margins, and the parties did not do so in their submissions.) At paragraph 3.6, the ICMA paper explains that on default, “Any excess collateral/cash remaining after the sale or purchase of collateral securities has made whole the position of a non-defaulting party cannot be retained by that party and so does not represent contingent additional compensation for bearing default risk”.

In my view, as a matter of law the position as regards excess collateral in the case of a default in a repo transaction depends, in the usual way, on the applicable terms, such as those in the Global Master Repurchase Agreement (“GMRA”). But as a matter of principle, I consider that BNYM is correct to submit that a repo counterparty is only entitled to keep collateral to the value of its outstanding debt, and that any excess must be returned. This is because the purpose of repo margin (like that of any security) is to provide lenders with a means of repaying what is owed in the case of default, not to give them a windfall profit. This point has consequences for the modelling performed by Mr Servis, and relied on by AP, as I shall now explain.

During the trial, considerable reliance was placed by AP on modelling produced by Mr Servis in his supplemental report of 14 March 2013, in response to a supplemental report produced by Mr Richard on credit monitoring. Partly on this basis, it was maintained that BNYM had misunderstood the impact of the repo financing in the event of a default by Sigma, and in particular did not appreciate the effect of repo haircuts. Taken at face value, the tables showed the financial position deteriorating according to the size of the haircuts.

The evidence as to Sigma’s repo position came mainly from the rating agency reports, and did not contain much detail. I also consider that it is unclear whether the agencies relied on information provided by Sigma, or had access to the agreements themselves. On 18 March 2008, S&P reported haircuts of between 2 and 10% as follows:

“Since July 2007, Sigma has had to turn largely to the repurchase (repo) market to refinance its debt maturities, and it now finances $11 billion of its asset portfolio through those markets. These repo agreements, which have maturities between three months and a year, enable Sigma to defer selling in current markets. Each repo requires overcollateralization ranging from 2%-10%. These overcollateralization levels could exhaust a large portion of Sigma’s capital since they are posted away from Sigma and in favour of the repo counterparties. If liquidity conditions worsen, these additional margin posting requirements could cause Sigma to default in the event of a missed collateral posting.”

A higher figure of 12.4% can be calculated on the basis of Sigma’s accounts to 30 April 2008, though BNYM points out that this is based on par, not market value. BNYM emphasises that the Private Placement Memorandum stipulated (in effect) that initial haircuts should not exceed 10%.

However that does not reflect what happened in practice. A report of 30 September 2008 says that Sigma told Moody’s that the total notional amount of their assets was US$27bn, with US$17.4bn of repo liabilities and overcollateralization of approximately US$8bn (US$4bn initial margin and US$4bn variation margin). This indicates that haircuts in excess of 10% were being negotiated, and probably significantly so.

Whatever the size of the haircuts, Mr Servis assumes in calculating loss that, in a liquidation scenario, repo counterparties would retain all of their pledged collateral, including the haircut. However as noted above, any excess over the value of the loan would have to be returned. His tables show that the value of assets available to meet senior noteholders’ debts would reduce with each increase in haircut even though the underlying asset price remains constant. As BNYM says, this cannot be right.

Mr Servis’ response was to the effect that in practice, repo counterparties have flexibility to value the assets and will select a price so as to retain all the collateral pledged to them. That is not in my view right either. Under the GMRA, the value of securities held as collateral is required to be determined as “the amount which, in the reasonable opinion of the Non-Defaulting party, represents their fair market value”. This is not in the counterparty’s discretion. (I add that I do not accept BNYM’s further submission that the tables actually support its case.)

BNYM’s case is that far from being a negative factor, as AP submits, Sigma’s reliance on repo financing was a positive factor. It seeks a finding that BNYM reasonably considered that Sigma’s ability to rely on repo financing was a benefit having regard to the long repo terms and low margins being reported by the ratings agencies and the relatively short maturities of the Sigma MTNs held by the bank.

I shall come to the evidence as to what the bank in fact thought about repos later. There are some other points to deal with on the evidence first. As to the length of repo terms, the passage from the S&P report of 18 March 2008 cited above refers to repo agreements which have maturities of between three months and a year. To show that Sigma had in fact negotiated long-term repos extending, for the most part, beyond the maturities of the MTNs held for AP, BNYM relies on S&P’s report of 12 September 2008. This shows, it says, that of Sigma’s repo obligations, US$4.5bn fell due in the third and fourth quarters of 2009, with further amounts maturing in 2010, 2011 and 2012.

The report makes clear that it was based on what Sigma had told S&P. It says that Sigma was able consistently to extend repos with approaching maturities. But the numbers mentioned for late maturing repos on which BNYM bases this point fall well short of Sigma’s repo liabilities, reported as US$17.4bn on 30 September 2008. BNYM’s contention is contested by AP, and I do not consider that it is proved. In any case, it is common ground that with falling asset values, Sigma was at risk of margin calls, as eventually happened.

In submissions by Ms Catherine Gibaud, junior counsel for AP, it summarises the effect of the repo financing as follows. By August 2008, it was clear on Sigma's monthly reports that the repo financing had reached US$17bn—which BNYM accepts. Even with a modest haircut of 15%, this meant, AP says, that the starting point for any credit analysis should have been that US$20bn of assets had left the Sigma vehicle as repo counterparty collateral. Broadly I accept this, and the general picture is in my view clear. (I should say that I have placed no reliance on Appendix 3 to AP’s closings, which BNYM says does not take account of a fall in outstanding MTNs.)

In its closing submissions, BNYM referred to an email sent to Mr Tant by Ms Saisselin of Sigma on 18 September 2007, which suggested that not all the margin was in fact posted to counterparties. This was not explored in the hearing, and while no doubt there will have been variations in the terms of repo agreements, it is not in my view sufficient to alter the picture as established on the rest of the evidence.

It is necessary to be clear as to what the repo transactions involved. A repo would enable Sigma to repay maturing liabilities, without an outright sale of assets, but on the security of the assets. The rating agencies considered that Sigma had managed this process well. For example, all the outstanding commercial paper (which was short term) had been repaid by July 2008. However, when the repo itself matured, unless Sigma was able to roll it over (and it seems that Sigma had some success in rolling repos), the only source of repayment was the assets. The process in effect bought Sigma time, in the hope that the market would return to a pre-crisis position, enabling it to resume funding by issuing securities. But it never did.

The fundamental point about the repos, in my judgment, is that they subordinated the senior note holders like AP to the repo counterparties. This is not in dispute, and followed from the terms of the Private Placement Memorandum, but was also inherent in the repo transactions themselves which transferred the assets to the counterparties. Even if BNYM is right that much of the repo lending extended beyond the maturity of AP’s MTNs, Sigma was at risk of margin calls on a falling market. If it could not meet a call, the whole of its repo exposure risked going into cross-default. This was in the event what happened, and it was not in any sense unforeseeable in my view.

The result was, as AP says, that when Sigma failed, most of its assets went to repay the repo counterparties. I reject BNYM’s submission, therefore, that the repo transactions were beneficial to AP. They were in my view very disadvantageous.

BNYM puts its case as follows. In hindsight, it says, following the failure of Lehman Brothers and the subsequent collapse of asset values, it is easy to say that repos represented a high risk and that Sigma should instead have focused on selling assets. But absent foreknowledge of Lehman, BNYM says, it was plainly reasonable to anticipate that marks on Sigma’s assets would increase as they approached maturity. Long-term repo represented an attractive option, because it avoided a need to sell assets at prices that appeared to be temporarily depressed. Once prices had recovered sufficiently, Sigma could either sell the assets and repay the repo, or re-repo them at a lower margin to par.

I agree that it is important to avoid hindsight, and there is evidence (which I accept) that Sigma managed its repo-funding with some skill. But I do not accept that long-term repo was an attractive option. Whether and if so when prices would recover was unknowable, and Sigma was at risk of further falls prompting margin calls in the meantime. I agree with AP that this did not depend on foreknowledge of the collapse of Lehman Brothers, which in any case was a part of the crisis that came to a head in September and October of 2008.

AP criticises BNYM for not carrying out a liquidation analysis for Sigma MTNs at any stage which considered the repo factor. In response, Mr Richard said that this would have been impossible because BNYM had no access to the CUSIP details of the securities held in Sigma’s portfolio, despite asking for it. The frustration that this caused BNYM appears from an email of 1 May 2008, in which Mr Robison said, "In terms of the overall recovery value of our debt, as you know we have tried and failed many times to get the complete listing of CUSIPS from Sigma that we would need to make such a determination. Without that we would just be taking a guess".

Nevertheless, I agree with AP that BNYM does not appear adequately to have analysed the impact of repos in a liquidation. Mr Ford maintained that BNYM thought that Sigma’s “pay as you go” enforcement procedures would be protective of short-term noteholders such as AP. In the event, they were not, though BNYM says that was partly due to the premature sales of assets pursuant to a misunderstanding by the lower courts in the Sigma litigation of the effect of the relevant documentation.

I do not accept BNYM’s case in this respect, because Mr Ford’s evidence is not supported in the contemporary material. In any case, the effect of the repo transactions over time was to reduce the remaining reason for confidence in the Sigma notes, namely their backing by good quality assets. Further, I do not accept (as it submits) that it was reasonable for BNYM to rely on the analyses of the rating agencies in this regard. I have referred above to the email from Mr Robison, chief operational risk officer at BNYM, saying that, "The rating agencies have been useless throughout this structured finance fiasco from the beginning”.

AP goes further, submitting that BNYM’s officers did not understand the effect of the repos. I doubt that this is the true picture. It is correct that there is some support for this in some of the answers given in the Oklahoma depositions, but these were experienced people who would have understood the general effect of Sigma’s repo funding strategy.

I have dealt with the repo-financing issue in some detail, because it was an important area of debate at trial. However though I have accepted many of AP’s submissions, overall I do not think that it has a decisive impact on the points which I have to decide in this case. In my view, the effect of repo funding is part of the overall factual picture.

Repos aside, I accept Mr Richard’s conclusion that BNYM monitored the condition of Sigma closely as the financial crisis continued to worsen in a way that was reasonable in the circumstances. In fact, there can be no doubt that it monitored it very closely indeed. The contemporary material makes that clear. Further, I do not accept the recovery figures proposed in AP’s closing submissions, which AP says a proper liquidation analysis would have shown. These are based on Mr Servis’ tables which are flawed for reasons I have stated.

I must now deal with the evidence as to what happened within BNYM as regards Sigma during the period from August 2007 particularly up to May/June 2008, when AP itself decided to hold on to the securities. I shall deal with this in periods, as AP did. AP says that had its known and agreed investment objective been applied, its discretionary portfolio would not have been exposed to the risk of failure of the Sigma MTNs. This part of the case raises in particular the allegation that at some point BNYM should have sold the MTNs, or at least ought to have informed AP of the risk before it finally did so in May 2008.

Both parties have given their own interpretation of the contemporary material, including large numbers of internal BNYM emails. AP in particular relies on what is said in these emails about Sigma in support of its case. In oral evidence Mr Tant (head of the Credit Analyst Group at BNYM) pointed out that views within a financial institution may differ about investments, and it is important that people should be able to express them openly. It does not follow that what they say in emails represents the only available view of the investment in question, or the view of the institution. I accept that general approach in seeking to evaluate the material.

I have set out the factual background above. What is striking in retrospect is how quickly the loss of confidence in the financial system set in. As early as 5 September 2007 a Moody’s report said that:

“SIVs in particular have to mark their portfolios to market, presenting difficulties when, as currently, the market in question is not really functioning. As our banking group stated in its global teleconference on 23 August, ‘…the blow to confidence of the global financial system means that what was once liquid is now illiquid, and good collateral cannot be sold or financed at anything approaching its true value’.”

As I have explained above, another SIV rated AAA called Cheyne Finance went into enforcement on 28 August 2007. From that time at the latest, I infer that Sigma was a major concern for BNYM not least because of the size of its holding. However, the bank also appears to have taken comfort from the distinction between Sigma and other SIVs, since Sigma did not have market value or NAV-based enforcement triggers requiring it to go into wind down on a drop in prices.

According to AP, the most important document showing why AP's exposure to Sigma should have been eliminated in September 2007 is the presentation to BNYM's Risk Committee attached to Mr Tant's email of 7 September 2007. I agree that it is an important document, and it was circulated to senior management within the bank. However, the parties are in disagreement as to what it shows. AP says that the presentation referred to the strategy of reducing exposure to SIVs and recommended the continuation of this strategy. What is critical, it says, is the analysis of the default risk for SIVs (including Sigma) of "medium". AP should never have been exposed to a default risk of medium.

BNYM says that the report’s recommendation was to “retain on approved list more established SIVs with strong sponsors” and to “continue to reduce to within limits approved programs” (i.e. to continue to allow the concentration to reduce to below 10% through maturities). There was a recommendation in the report that BNYM “Actively seek opportunities to reduce Restricted SIV’s while minimizing loss impact on clients”, but this was not a reference to Sigma.

BNYM correctly paraphrases the recommendation in my view. In distinction to “Restricted SIV’s”, Sigma was bracketed with “more established SIVs with strong sponsors”. BNYM says that in describing the default risk of SIVs as “medium”, the report was referring to the SIV sector as a whole, rather than to Sigma specifically. Again, I agree with that interpretation, and it was supported by the oral evidence of Mr Tant.

Overall, the presentation appears to give a fair view of the sector as reasonably seen by BNYM at the time. There was no recommendation to sell in the case of Sigma, as opposed to some of the other SIVs, and so far as I am aware this did not happen at this time.

AP cites a number of reasons why its Sigma holding should nevertheless have been sold at this stage. First, it says that a person whose objective is to “safeguard principal” takes a very small hit and cuts his losses. However, I reject this. There was no good reason to take a “small hit” at this stage. For reasons explained, I have rejected AP’s case that it was more conservative in its investment approach than other securities lenders.

Second, AP says that Sigma was at that stage proposing that BNYM enter into a trade to reduce its exposure to Sigma. However, the evidence shows that the bank considered this, but it rejected it on grounds that were reasonable. Third, it says that while a “hold to maturity” policy might have been reasonable at the time of acquisition, there was no reason to follow it in September 2007. However, I accept BNYM’s evidence to the contrary, because at this time it was reasonably anticipated that the notes would be paid in full. Fourth, AP says that it should have been informed of any concerns relating to its investment. I reject that submission as well for reasons which are essentially the same as the above.

I am satisfied that BNYM was under no duty to dispose of AP’s Sigma holding at this point in time, whether under its duty of care, or under the investment guidelines, or otherwise. It had no way of knowing that the crisis was going to deepen. In his evidence Mr Tant said that the market expected things to normalise in 2008: “…the biggest concern in the fourth quarter of 2007 by all the players in the marketplace … was ‘just get us into the first quarter 2008, things will normalise’".

Further, the crisis caused problems right across the financial market. As investment professionals, AP would itself have been aware of the situation. It does not appear, however, that it took any particular steps to scrutinise the securities which were being held on its account in place of that part of its portfolio that had been loaned out.

As noted above, by the end of October 2007, Sigma had stopped issuing MTNs. On 26 November 2007, HSBC announced that it was bailing out its SIVs. On 13 December 2007, following a reduction in their capital NAV to below or near 6o%, Citigroup decided to consolidate its SIVs on to its balance sheet.

According to BNYM, the latter development was regarded as good news, because of the bank’s very large exposure to the Citigroup SIVs. I do not, however, accept its submission that because it meant that there would be less competition for liquidity in the marketplace, it was also positive news for Sigma. AP did not press in closing the view expressed by Mr Servis in his report that this was the moment to eliminate, wherever possible, all exposure to independent SIVs.

The evidence of Mr Tant and Ms Demmler is that they were comfortable with Sigma as an investment at this time. This has some support in the contemporary material. When asked by the Managing Director of BNYM Asset Servicing why the bank continued “to feel good” about Sigma, Mr Tant responded by email of 19 December 2007 that, “Sigma continues to have strongest capital position of any SIV and has considerable capital cushion. … They are working on deal to strengthen funding profile but according to … Sigma the banks keep going back into a huddle. Asset quality is solid and suffered no material downgrades”.

At year end 2007, unlike some other SIVS, Sigma’s ratings had not been downgraded. It had however been placed by Standard & Poor’s on “Negative Outlook”. The Wachovia Securities article dated 18 December 2007 I referred to above seems to me to reflect perceptions as at the end of 2007 so far as the evidence in this case is concerned. It said that “Even though Sigma Finance is the largest question mark on the street, it remains one of the strongest as well, with a Net Asset Value of capital at November month-end of approximately 79”. BNYM says that this means that Sigma’s assets were sufficient to repay all its senior debt (which included AP) and 79% of its capital notes.

The evidence suggests that in early 2008, the view about Sigma became distinctly less sanguine. One reason for the change in sentiment may be that in October 2007, Citigroup, Bank of America, and JP Morgan Chase had announced an attempt to raise a US$100bn superfund, which was known as the Master Liquidity Enhancement Conduit or M-LEC, to help support the orderly unwinding of many SIVs. On 7 January 2008, it was reported that M-LEC would not be launched after all.

Whatever the reason, this change is demonstrated in what Ms Falchetti of Standish called a “dramatic change” in Sigma’s responses. In a report which she prepared in February 2008, she says that, “of all SIVs in business at the start of 2007, Sigma is the oldest and has always enjoyed a reputation of being managed very well and holding high quality assets. Sigma is the only non-bank sponsored SIV in which Standish has invested”.

However, she goes on to refer to numerous conversations with Sigma since the start of the SIV crisis. She notes that Gordian Knot (the manager) considered that funding problems were a temporary condition, and that all discussions with them up until 15 January 2008 were relatively reassuring. Following its decision to take its own SIVs on to its balance sheet however, “Citi apparently walked away from its efforts to help Sigma”. (This was presumably a reference to M-LEC.)

Ms Falchetti then describes calls which she had with Sigma at the end of December, and then on 15 January 2008:

The "your guess is as good as mine" remark is of relevance, since the US$5.5m tranche of AP’s MTNs also had a March 2009 maturity. AP says that it is the clearest of evidence of a belief by Gordian Knot that there was a significant risk that the Sigma MTNs maturing in March 2009 would not be paid on maturity.

The report goes on to say that “Gordian Knot is now scrambling to find additional repo counter parties”. It says that an ad hoc creditor group had been formed, including BNYM. Once “May maturities hit, Sigma has the potential to go into enforcement”.

In meeting this report, BNYM has emphasised the operational independence of Standish. However, the evidence shows considerable contact between Standish and securities lending. Ms Falchetti’s report itself includes details of its holdings, together with maturities, which securities lending are said to have provided. In any case, it seems unlikely that two arms of the bank would have taken a different view on such a large and potentially risky exposure.

Although it is not suggested that Mr Tant or Mr Fort saw this report, it appears that Mr Tant was in on Ms Falchetti’s call with Sigma of 15 January 2008. He certainly spoke to Ms Saisselin of Sigma himself and reported it to Mr Robison (who was the bank’s chief risk officer on the reinvestment side) the next day.

In fact there is a chain of emails. On 15 January 2008, Mr Robison says of the Sigma notes “…They should be OK until at least May and June when the last of the ICR fund debt matures. I would be more worried if we had late 2008 or 2009 paper here”. Late 2008 or 2009 paper was of course what AP held. BNYM says that the concern related to the risk of a mark-down or downgrade of Sigma MTNs, rather than a risk of loss on maturity, but I do not accept that interpretation.

The ICR (Institutional Cash Reserve) fund referred to here was the liquidity vehicle for the legacy BoNY segregated account clients. As such, it had to be able to operate at a dollar in, a dollar out. This meant that the NAV of the fund had to be maintained above the 0.9975 threshold, which made it susceptible to fluctuations in the value of its underlying assets. It had a holding of Sigma securities of more than US$500m, the last of which, as Mr Robison’s email indicates, matured in May/June 2008. It is clear from this and other material that the potential effect of Sigma on the ICR fund was of particular concern to the bank, including Mr Tom Ford who himself came from the BoNY side of the business.

In his response to Mr Robison of 16 January 2008, Mr Tant reporting on the conversation with Gordian Knot said that “Juliette [Saisselin] let the air out of the balloon in terms of potential fixes near term, rather indicating that they are casting far and wide to look for more repo availability. … No ceiling on repo, however I suggested it is not a long term fix”. For reasons explained above, he was right that repos were not a long term fix. He went on, “what is plan B if markets don’t normalise. They are not ready to have that discussion at least publicly”.

BNYM says that nothing in this undermines Mr Tant’s evidence of his view as to Sigma’s unlikelihood of default. I do not accept that. Nor do I accept that, as BNYM puts it, the fact that Fitch stopped rating Sigma at the end of January 2008 was a “complete non-event”. (The explanation appears to be that Gordian Knot had not renewed its contract with Fitch and would not continue to provide it with information on Sigma.) There were negative signs in the market by now. Asked whether there were any offers to buy Sigma paper, Mr Mannix reported on 30 January 2008, “The only indication we got from the Street was at 75”. This of course was far below par.

A meeting was arranged between the bank and Gordian Knot for 4 February 2008. BNYM says, and I accept, that Mr Ford, who was Co-Head with Ms Kathy Rulong of the Securities Lending Group, was keen to ensure that the bank was as well prepared as possible. On 31 January 2008, he emailed Mr James Palermo, who is a Vice-Chairman of the bank, saying, “We have a senior level meeting on Monday with the management of a AAA-rated, independently managed SIV called Sigma which we believe may be moving closer to default because of current market conditions and the difficulty it is currently experiencing in funding itself”. Mr Ford added that he thought that it would be a good idea for senior officers to discuss the client implications with him.

BNYM’s submission (and indeed Mr Ford’s evidence) is that this email did not reflect an actual belief that Sigma would default, but rather a concern about what he might be told by Gordian Knot in the meeting. I reject this submission. Mr Tant said that he was being “overly dramatic”, but I reject that too. The email was to one of the top people in the bank, would have been carefully drafted, and speaks for itself. Any doubt is dispelled by some late disclosure in the form of another email that Mr Ford sent to Ms Rulong the same day: “I'm hoping against hope,” he said, “that we can do something to reduce our positions before its [Sigma’s] demise”.

Contrary to BNYM’s case, I consider that this is what he thought at the time. Mr Ford retired in June 2008, though he continued to assist the bank on various matters. He was clearly a loyal as well as a respected officer. I have little doubt that he declined to come to London to give his evidence, giving it instead from New York over the video link for an afternoon only, because he realised that he would be unable sensibly to sustain the bank’s interpretation of these emails in cross-examination.

I was not taken to a note of the meeting of 4 February 2008, but it is not in dispute that the bank was unable to persuade Gordian Knot to provide CUSIP level of the Sigma portfolio. However, according to Mr Ford’s witness statement, “I recall leaving the meeting satisfied with Sigma’s financing strategy”. According to his witness statement, he was “90% confident” that the notes would be paid at maturity.

Again this assertion of confidence made in his witness statement for this trial has to be measured against the contemporary material. The bank was discussing with Sigma the possibility of a deal which would take the ICR fund out of its May/June Sigma holdings by way of a repo transaction. There were some difficulties with this, but Mr Ford said in an email of 7 February 2008 that, “i'd like to reiterate my interest in getting something done because i continue to believe our unsecured notes may ultimately represent a risk level of at least 15% of their face value and the repo transactions [with Sigma] represent less [risk]”.

BNYM says that this simply reflects the particular importance of maintaining the value of the ICR fund. That may have been the motivation, but I am satisfied that it makes no difference to Mr Ford’s view of the Sigma securities. As AP points out, not only does this email foresee a risk of loss on Sigma holdings, but the risk of loss on AP’s holdings—which unlike those of the ICR fund did not mature until October 2008 and March 2009—would likely be higher.

On 21 February 2008, a client which had itself made contact with BNYM required it to sell its holding of US$6m at a loss of US$1m, which (Mr Mannix said) suggested a price of about 85. Mr Ford says that this represented a loss of about 20%. In an email at the time, he said that he “would sell [Sigma] to whoever was buying”. At about this time Mr Fort reported that they had sold Mellon Capital mutual fund's US$27m Sigma bonds maturing on 12 March 2009 at 87.5. On 12 March 2008, BNYM sold US$53m maturing 15 September 2008 at 89 on behalf of a client.

Given this evidence, AP submits that BNYM was at fault in reporting in its holding reports to AP near-par prices for Sigma MTNs in the period up to 30 April 2008 when it knew, from its own experience, that market prices were well below this.

This submission is a substantial one, and my view is as follows. These securities were not exchange-traded, and as indicated above, the secondary market was thin. A pricing source was necessary not only for reporting to the client, but to enable BNYM to mark the securities to market. Since securities lending agents generally acquire collateral on a hold to maturity basis, actual market transactions might not happen at all. For this reason, BNYM relied on an independent pricing vendor, IDC.

The position was clearly disclosed to AP in the holding reports:

“Prices of securities reported herein are provided by pricing vendors and rating agencies (‘pricing vendors’) used by The Bank of New York Mellon (BNYM) in the ordinary course of business. Such prices are not independently verified by BNYM. Some reported prices are not updated by pricing vendors on a regular basis (and therefore are considered stale) and in some cases, pricing vendors do not provide prices. For securities where BNY Mellon's pricing vendors do not provide prices, BNYM may use prices (which may be indicative bids) provided by one or more dealers in such securities or by the Depository Trust Company or other security depository (generally par value or the price paid on the last trade settled through such depository) until such time that BNYM's pricing vendor provides prices.

There was also a disclaimer of liability by BNYM. Though the above evidence shows that the IDC pricing (like the ratings) lagged the market, I am satisfied that using a pricing vendor was normal and reasonable practice. Despite sales which achieved much less than par, I consider that BNYM acted reasonably in continuing to report the IDC prices at this time. What eventually prompted BNYM to contact AP about Sigma was a sharp drop in the IDC pricing in April 2008.

Although Mr Ford and Mr Tant suggested that BNYM had continuing confidence in Gordian Knot (i.e. Sigma’s management), again this is inconsistent with the contemporary material. An email of 17 February 2008 shows Ms Rulong telling Mr Ford about a meeting of Sigma investors, who were "very concerned about the arrogance of SIGMA's management and that they were not doing enough…. that they cannot continue to protect their capital note holders from losses there is also considerable concern about the rating agencies and possible imminent downgrades”.

On 13 February 2008, Mr Mannix reported that S&P refused to sanction the proposed repo deal with the ICR fund, which was also AAA rated. He said that S&P "decided that it would be inappropriate to increase exposure to a distressed issuer, and as a SIV it would be considered distressed". Mr Mannix said that S&P might revisit the repo proposal if Sigma would agree a 150% margin with BNYM. (This margin may be compared with those discussed earlier in this judgment.) There is other evidence to the effect that Sigma was at this time dealt with by the “distressed desks” of securities traders, and I agree with AP that this is significant as showing the market view. In the end, the proposed ICR deal with Sigma came to nothing.

Sigma continued to be AAA rated at this time. Moody’s put it on review for downgrade on 27 February 2008, indicating that it was considering a downgrade to the AA range. However, overall it remained positive, stating that:

“The company’s portfolio is of high credit quality, with 45% Aaa-rated, 43% Aa-rated, 9% A rated, 2% Baa rated, and 1% rated Ba-B. The company has limited exposure to ABS CDO and monoline wraps, and has no direct exposure to US subprime RMBS.

Since mid-2007, the company has had limited access to the commercial paper and medium-term note markets. However, the company has successfully tapped alternative funding sources such as repurchase agreements (over $22 billion transacted since July 2007), asset for debt exchanges, opportunistic debt buy-backs and asset liquidations. The notional value of the company’s asset portfolio declined from USD 57 billion on July 27, 2007 to USD 41 billion on February 15, 2008, reducing the company’s leverage from over 13 times to approximately 9 times.

Despite these positives, the overall market price deterioration, continued inability to issue senior, and reliance on repos have increased the company’s risk profile. Asset prices have continued their unprecedented decline. The average price was 100.2% of par on July 27, 2007, and was 96.99% on Feb 15, 2008. To date assets have been liquidated close to their marks.”

Sigma’s monthly report for the end of February 2008 showed that by that date, the par value of its asset portfolio was US$39.2 billion. According to BNYM, based on an asset price of 96.99, that implies a market value of US$38 billion, compared to outstanding debt of US$36.1 billion.

The report seems to have given some reassurance to Mr Ford. This is demonstrated by an email to Mr Tant of 7 March 2008:

“whenever I see this monthly report I get a sense of higher, not lower confidence about sigma. obviously the program is 4 billion smaller (feb vs jan); with capital roughly the same (-21 million). a major portion of the portfolio decline ($2.9 billion) seems to be the liquidation of 2-5 year assets, generally from the ugliest categories mbs/clo&cbo’s/other abs. while they lost 1.1 billion of short maturities (0-1 year); that doesn’t seem so surprising under the circumstances. do you see it similarly, or do you think I’ve got my rose-colored glasses?”

A reply email (if there was one) is not in evidence, though Mr Ford says that so far as he can recall, Mr Tant did agree. Mr Tant says that the email is consistent with his general thinking at this time.

On 18 March 2008, Sigma issued its quarterly report setting out its results to 31 January 2008, which were confirmed by its auditors, PwC. Page 5 of the report included a statement that “The directors, having taken into account market considerations subsequent to 31st January 2008, still expect sufficient funding to be made available to Sigma for the next 12 months”.

AP contends that by January/February, it was clear that AP's Sigma notes should have been sold. It says that it is clear that the notes were wholly inappropriate having regard to AP's investment objectives. Prudent practice also required BNYM to speak to AP about its investments (in other words well before it did so on 16/19 May 2008).

BNYM says that the end February monthly report shows that for all the concerns being expressed at the time, Sigma remained a strong and well-financed entity. It considered the situation carefully, reaching out both to Sigma and to other creditors. The fact that it did not sell out of Sigma at prices well below those justified by Sigma’s assets is no grounds for criticism.

As I have explained, I consider that perceptions of Sigma took a turn for the worse at the beginning of 2008. On the other hand, securities lending client sales at this time (there were only one or two) seem to have been at the clients’ initiative. An underlying issue was whether the strength of Sigma’s asset portfolio meant that the securities could be held to maturity with a reasonable prospect of repayment in full for the senior noteholders. Mr Tant says that this was his view at this time. Mr Ford clearly did not agree, but by the end of February 2008, even his perception of Sigma appears to have improved (that was not to last).

The bank’s witnesses are entitled to be judged on the basis of what they knew at the time, and on balance I consider that BNYM is right to submit that there was no obligation on it to sell AP’s Sigma notes, whether under its duty of care, or under the Investment Guidelines, or otherwise, at this time. A ratio trade would have got rid of the Sigma holding without an immediate loss, but it would have involved AP very substantially increasing its overall market exposure, even assuming the trade could be done within its guidelines (which BNYM disputes).

It is important to appreciate, as I explain later, that in practice such transactions would not have been performed by BNYM without consulting AP first and getting its consent. In reality, this is the prior question, therefore. As a matter of fact, BNYM was not contacting its clients about Sigma at this time. I accept that AP has an arguable case that by now this duty had arisen. On the other hand, this was a difficult exercise, as it proved to be when it was eventually done. It ultimately was a judgmental matter for the bank whether to do so at this point. Matters were soon to change, but on balance I do not consider that the BNYM’s appreciation of the risks associated with Sigma was such that it was in breach in failing to speak to AP about its Sigma holding in January or February 2008, particularly given the more positive note that appears to have existed at the end of the month.

By early March 2008, the documents show that there was a move to a firm-wide solution of the Sigma problem. I infer that it is from this time that securities lending and Standish began to adopt a cooperative approach. There is a considerable amount of analysis within the bank over this period, and it shows that monitoring of Sigma was vigorous.

The problem was pressing, because as S&P noted on 18 March 2008, Sigma had approximately US$15bn in senior debt and repo maturities that would come due between then and June 2008. (In fact in due course these liabilities were either met or rolled over.) S&P moved Sigma from negative outlook to credit watch negative, implying that a downgrade from AAA was now considered more likely.

Sigma itself was trying to persuade the bank to enter into a ratio trade. Mr Mannix reported on 2 April 2008 that, “Sigma called to tell us that essentially, things are getting worse for them, not better, with the passage of time. They cannot access funding, and maturities continue to drain liquidity. They suggested we ‘strongly’ consider a debt for asset swap at some multiple”. Mr Mannix says that Ms Saisselin (BNYM’s main point of contact with the managers, Gordian Knot) had a tendency to overplay things if she thought this was required in order to get things done. Whilst I agree that it was a considerable achievement that Sigma managed to continue to meet its obligations at this time, I also consider that it is very likely that things were indeed getting worse for it not better with the passage of time, as Mr Mannix was told.

On 4 April 2008, Moody’s downgraded Sigma from Aaa/P-1 to A2/P-2, and S&P followed on 7 April 2008. BNYM minimises the importance of this, pointing out that this still left Sigma with a mid-investment grade rating, and Moody’s write-up pointed to numerous positive features. It submits that both Moody’s and S&P’s reports therefore “chimed with the Bank’s overall view that Sigma remained unlikely to default”.

This is an unduly optimistic submission, in my opinion, particularly since the rating agencies were perceived at the time to have seriously misread Sigma’s true position. I consider that losing its AAA rating was an important moment in the history of Sigma. An article from the Financial Times of 10 April 2008 was circulated within BNYM. It said that, “Its management fought hard but it looks as though Gordian Knot's $40bn structured investment fund, Sigma, faces a slow death". The reason given was that even if the market revived, the loss of its AAA rating would prevent Sigma from funding itself at the prices which would make the operation profitable. The significance of this is that a market recovery would no longer provide Sigma with a return to normal business.

BNYM submits that various presentations including those on 28 March 2008 and 7 April 2008 show that though the bank was considering capital support for various of its funds, this was not linked to an expectation of a Sigma default, but rather the need to protect its relationship with clients (and the resulting revenue stream) from the impact of Sigma being downgraded and marked down in price. The documents in my view do not support this interpretation.

The bank’s case is that it considered that the problem was one of liquidity, and that it considered that Sigma would pay on maturity, and so believed that it was best to continue to hold the securities. In closing, it submitted that there was wide room for different views, and that it was at least a reasonable view. There is some support for this submission in a briefing which Mr Tant and Ms Demmler gave to relationship managers on 7 April 2008. A note records them as saying:

However I am satisfied that this view did not prevail with more senior officers in the bank, as is shown by the contemporary material. On 10 April 2008 Mr Ford circulated an email to senior bank officials about a conference call that had taken place with the fund manager Fidelity. He reported that Fidelity “… are more pessimistic than we are regarding Sigma's life expectancy. They don't expect them to survive though 6/30/08. … Based on the results of this conversation, I intend to immediately return to our strategy of utilizing ratio trades (at 3 to 1) with Sigma to the greatest extent possible to reduce our longer maturity positions and increase the likelihood of the payment of our shorter maturities.”

Shortly afterwards, BNYM conducted a ratio trade taking out a US$300m position in Sigma securities for what I infer was a major Middle East client of securities lending. (In accordance with the convention adopted by the parties, I shall avoid naming BNYM clients.) What seems to have happened is that BNYM approached the client for instructions first, but the client responded saying that it was for BNYM to take the decision. BNYM says that it had to take action because the clients’ guidelines restricted holdings to AAA securities. However, this is not the full story. An internal email from the relationship manager refers to a “significant concern that Sigma will fail in Q3 or Q4”.

BNYM says the relationship manager’s concern arose because “he had misunderstood the situation”. Its case is that it was the bank’s considered view, based on the analysis performed by Mr Tant and Mr Demmler, that the most favourable outcome for clients was to continue holding their Sigma MTNs. It cites a Credit Suisse report of 6 April 2008 which was relatively optimistic about Sigma held by money market funds.

Again, the ratio trade for the Middle East client was part of a picture. AP cites ratio trades carried out for other clients in the sum of US$137m. BNYM says that between April and June 2008, it concluded ratio trades on behalf of numerous commingled and segregated account clients. In total, the bank conducted ratio trades for Sigma MTNs with a face value of US$674.5m, which it says, at a 3 to 1 ratio, raised almost US$2bn in liquidity for Sigma.

BNYM points to the raising of liquidity as a positive feature, but the fact is that for these clients it was judged better to take on three times the par value of the Sigma securities in other assets rather than continue to hold the securities themselves. BNYM says that this course was not available to AP because of its guidelines and the nature of the assets on offer which were predominantly longer than two years. But BNYM did not at this time put to AP the option of loosening its guidelines, which could have been done, and I infer happened in the case of other clients.

On 22 April 2008, Mr Ford and Ms Rulong made a joint presentation to the bank’s Collective Investment Funds Committee with the title “Securities Lending – SIGMA Investment Alternatives”. Since these were the co-heads of securities lending at BNYM at this time, what they said at the time is clearly significant as representing the view of the bank. They pointed out some of the strengths of Sigma, as noted elsewhere in this judgment.

However, they also explained the risks as they saw them at this time. These were that:

I consider that this is clear evidence of the gloomy nature of the bank’s analysis by late April 2008. BNYM sought to explain the document in various ways none of which was convincing. Although, as it points out, it is premised on the lack of additional funding, it makes clear the problem that such funding gave rise to, namely the disappearance of suitable assets if other note holders got to them first by way of ratio trades or “locked them up” in repo agreements. (Ms Saisselin (of Gordian Knot) told Mr Mannix on 23 April 2008 that its funding needs to the end of September 2008 were US$18bn, averaging about US$3.5bn per month, which gives some indication of the scale of the problem.)

Mr Ford’s suggestion in cross-examination that what was said about the risk of loss was “true” but it did not include “all the protective liquidation and bankruptcy payment structures” does not stand with the terms of the presentation (and is inconsistent with the terms of an email sent by Mr Mannix the following day).

In particular, I reject Mr Ford’s suggestion that Ms Rulong was “casting Sigma in the most negative possible truthful light in order to get this committee to be as responsive as we both thought they needed to be regarding the type of trading we needed approved”. Mr Ford avoided the language of misrepresentation, but that was the effect of what he was suggesting. The idea that Ms Rulong would misrepresent the position to the bank’s Collective Investment Funds Committee in order to get it to take action, or that Mr Ford would allow her do so, does not bear scrutiny. (Ms Rulong did not give evidence because she no longer works for the bank.) I am satisfied that the presentation stated their view as heads of securities lending at the bank of Sigma at the time.

BNYM points out that the ratings agencies gave Sigma A ratings right up to the collapse of Lehman Brothers, and says that ratings were historically good indicators of default probabilities. It points to what Moody’s was saying as late as July 2008 about the market value of Sigma’s portfolio, which it says supports the proposition that Sigma’s failure was the unpredictable result of the Lehman collapse.

However, I am satisfied that by early 2008, Sigma’s rating was not regarded in the market as a reliable indicator of the strength of the company. Further, as evidenced by comments which are set out elsewhere in this judgment, it was not regarded as such by BNYM. By now, that is, April 2008, it is clear that the bank placed little or no weight on Sigma’s rating (or the agencies’ analysis supporting their ratings) in assessing risk. In summary, whilst the view that Sigma would pay in full at maturity may have been an available view, it was not by this time the view which prevailed in BNYM. The next transaction involving client “I” shows this.

This client was one of a number of significant US clients of the bank. At the end of April 2008, it was discovered that US$50 million of Sigma MTNs had been acquired for the client in breach of its guidelines, and the question arose how the bank should handle this. To some extent it is, in my view, a barometer of the bank’s beliefs as regards Sigma, since any loss would be for its own account. The options were to hold the MTNs to maturity on its own balance sheet, or do a ratio trade as principal, which would involve Sigma buying them back at par, in return for the purchase by BNYM of US$150m of Sigma assets.

Commenting on the option of retaining the securities to maturity, Mr Ford said in his witness statement that, “Again, I did not think this risk [that is, the risk of Sigma defaulting] would materialise, but it was a factor to take into account. Had I thought default was likely, I would not have suggested it as a possibility at all”.

This evidence that he did not think that default was likely was completely contradicted by the bank’s late disclosure. As originally disclosed, the email chain does not include Mr Ford’s email of 30 April 2008 at the head of the chain. Of the options, what he in fact says in that email is that:

“I think b is the best way to go as well. Sigma is likely to default by 9/15 September 2008 which would eventually lead to a realized loss and probably some disclosure and reserves before that. We need to keep this moving or the proposed abs securities may disappear.”

This email is inconsistent with BNYM’s case. In cross-examination Mr Ford asserted that he exaggerated his concerns in order to get approval for the deal that he wanted done, which was a ratio trade. I do not accept this. He forwarded his email to Ms Rulong and copied it to Mr Palermo (who as I have said was a Vice-Chairman of the bank). As I said in relation to the similar suggestion in relation to the presentation to the Collective Investment Funds Committee, the idea that he would express a false opinion to get things done does not bear scrutiny. I am satisfied that he said that Sigma was likely to default by 15 September 2008 eventually leading to a realised loss because that was what he thought at the time.

In the event, his prediction as to the time of default was close—he was out by a week or so—and he was right in predicting that the default would lead to a realised loss. By now, it was entirely predictable that Sigma’s predicament would sooner or later have this result. The risk was not in any way dependent on the unanticipated collapse of Lehman Brothers as the bank has suggested. It is regrettable that this email was not disclosed.

In fact, BNYM decided not to retain the MTNs to maturity, but disposed of them by way of ratio trade. The obvious conclusion, as AP says, is that this shows that a ratio trade was seen as a better outcome than holding on to them.

AP submits that “between March and 14 May 2008, BNYM senior management embarked on a strategy of sales and ratio trades to reduce and eliminate Sigma exposure for clients, including segregated and commingled clients”, a strategy in which it says AP should have been, but was not, included. Although as stated above there were sales, and this is an important part of the evidence in the case, I do not think that the evidence supports the submission that there was a “generalised strategy” of the kind alleged over the March to May 2008 period. Nor do I think that AP has established that BNYM preferred other clients over AP, since the sales that did happen were relatively limited compared to BNYM’s total Sigma holdings. By May 2008, when BNYM began contacting clients about Sigma, AP was included. I do not accept, therefore, Mr Cirrito’s view that it was treating AP as a “third class client”.

However, although I have rejected AP’s retention claim as regards the 2007 period with little hesitation, it is plain that the argument that the securities should have been disposed of, along with the alternative claim that it should have been contacted by BNYM about Sigma’s problems earlier, becomes much stronger with the passage of time.

BNYM objects that AP does not anywhere say what it believes the bank should have assessed as the intrinsic value of the MTNs to serve as a comparator to their market price. It says that it is impossible to see on what basis it can be considered unreasonable for the bank to have reached a view that the most likely outcome was that the Sigma MTNs would be repaid, in full, on maturity.

However, for reasons stated, I am satisfied that by April 2008, this was not the view which prevailed in BNYM. By April 2008, it was recognised that Sigma was in a very vulnerable position. As I shall explain shortly, what eventually prompted the bank to contact its clients about the situation was the action by IDC, the pricing vendor, at the beginning of May 2008, sharply reducing the indicative price of the securities. At that point, I infer that the bank had little option but to explain the ramifications to its clients.

The issue is as to the extent of BNYM’s obligations. The bank submits (with force in my view) that the size of its Sigma exposure is directly relevant to the reasonableness of its actions. There is obviously a difference between how a bank will manage an exposure of US$35m and one of over US$5bn, particularly in the market conditions of August 2007 onwards. What might appear to be a reasonable course of action for one client in isolation, as it says, may take on a very different complexion when considered in the context of hundreds of clients.

The most significant point appears to me to be as follows. It is common ground that a sale at this time could only have been accomplished at a substantial loss to AP. A ratio trade would have got rid of the Sigma holding without an immediate loss, but it would have involved AP substantially increasing its overall market exposure.

Further, BNYM says (and I do not think that this is in dispute) that a ratio trade could not have been concluded within AP’s existing Investment Guidelines. Absent an express instruction from AP therefore, the bank could not have been under any obligation to execute such a trade, which would have amounted to a breach of contract.

More broadly, I am satisfied on the evidence that in practice transactions involving such substantial implications for the client would not have been performed by BNYM without consulting the client first and getting its instructions. If a recommendation was requested by the client, it would doubtless be given. Mr Cirrito opined to the contrary, saying that, “It should just have been a case of BNYM calling AP to let AP know”. I do not accept that evidence. Whilst it is correct that the ratio trade for the Middle East client was done without its express consent, this probably says more about the importance of the client, than the extent of the bank’s duties as securities lending agent. Without such instructions, I do not consider that BNYM as lending agent was legally required to exercise its discretionary mandate and dispose of AP’s Sigma holding, whether under its duty of care, or otherwise. While there may be exceptional cases, this would be a major requirement to impose on a securities lending agent, and I do not consider that the evidence justifies it as regards Sigma.

For that reason, I consider that the key issue is whether BNYM was under a duty to contact AP about the Sigma situation, which is AP’s alternative case, and if necessary obtain its instructions. As a matter of fact, up until the beginning of May, BNYM did not contact its clients generally about Sigma, though it did thereafter. It would be a difficult exercise, as it proved to be when it was done. It would also be a sensitive exercise, because a wholesale disposal of the securities was not possible for reasons explained above. It was ultimately a judgmental matter for the bank whether to contact clients, and if so which, and at what point in time. Despite the force of AP’s contentions, on balance, I do not consider that the position was such that BNYM was in breach of duty in failing to communicate with AP about Sigma before it actually did so in May 2008.

If the above conclusion is wrong, I do not consider that any duty to dispose of the securities and/or contact AP arose until April 2008, when Sigma’s position was perceived to have substantially deteriorated as described above. For these reasons, I find against AP on its retention claim.

AP’S CLAIM RELATING TO COMMUNICATIONS BY BNYM IN MAY 2008

I have mentioned elsewhere that the prices given by BNYM’s pricing vendor, IDC, were much higher than were being realised in the relatively few actual sales of Sigma securities of which there is evidence (the IDC prices were relatively close to par). It seems that on about 22 April 2008 Standish sold US$50m December 2008 maturities at 68, which Mr Fort described as a “little desperate”.

On 1 May 2008 IDC reduced its indicative prices on Sigma notes to 70 (for October 2008 maturities) and 68 (for March 2009 maturities). These were the maturities held by AP. I am satisfied that this reduction represented a steep and sudden adjustment by IDC, and left noteholders such as AP potentially facing a large loss.

For May/June 2008 maturities, which would fall due soon, Mr Robison felt that the IDC price was too low. It is clear from the context that his concern was the effect that the IDC prices would have on the NAV of the ICR fund. The bank substituted 97.20 for May 2008 and 95.8 for June 2008 (which were the remaining ICR maturities) on the basis of a probability of repayment in full of 90% and 85% respectively and a recovery in the event of default of 72%. (In fact, these notes were paid by Sigma at maturity.)

At this time, the bank produced calculations as to LGD (loss given default) in respect of the various Sigma maturities in a spreadsheet Mr Tant sent to Mr Browne (Chief Risk Officer) on 1 May 2008. Mr Tant said he could not recall who prepared it, but there is no reason to suppose that he was not comfortable with the numbers. For the maturities from September 2008 onwards, which included those held by AP, the analysis predicted a loss of 28.5% based on a 75% chance of default and an estimated loss of capital of 38% in the event of default resulting in projected NAVs of 71.5%. This is a very different calculation from that applied to the May/June 2008 maturities.

BNYM says that it involved considering the lowest prices at which the Sigma MTNs could be valued consistent with maintaining the commingled fund NAVs. It was, BNYM says, a “hypothetical downside scenario” not intended to be a realistic estimate of the expected loss. However, in the light of the other evidence I do not accept this, and agree with AP that the document is self-explanatory. As AP says, it undermines BNYM's case that it still believed that the AP Sigma MTNs were on track to being paid in full on maturity, and is inconsistent with any belief by this time that there would no shortfall on principal in the event of default.

The steep reduction in IDC’s pricing of Sigma’s securities had the consequence that in May 2008 BNYM made contact with clients who were exposed to Sigma, or at least some of them. In this case, attention has centred on two telephone calls between BNYM and AP on 16 May and 19 May 2008. Before coming to them, I must deal with the evidence as to the contact which was made with other clients.

From early May 2008, Standish began to write to its own clients about Sigma. The letters began by identifying the client’s Sigma holding. The letters continued:

“Based upon our continuing consideration of the recent market events and the liquidity challenges and related financial difficulties Sigma is experiencing, and with no certainty of relief being realized in the near term, we believe that there is a significant likelihood that Sigma will not be able to continue to meet its commitments in respect of the Sigma notes in the medium term. A default in its obligations under the Sigma Notes will cause an enforcement event which will likely result in a prolonged liquidation period as well as diminished principal repayment.”

In an effort to mitigate this risk, the letters described an opportunity to conduct a ratio trade with Sigma (at the usual 3 to 1 ratio for such transactions). It expressed the belief that the substitute securities “are of a better credit quality and prospects than the Sigma Notes and will result in an over-all improvement of the portfolio”. Because some of the substitute securities did not conform to client investment guidelines, a waiver of those guidelines was requested if the approach outlined was acceptable to the client.

In substance, the letters proposed that clients agree to a ratio trade to dispose of their Sigma securities because of the “significant likelihood” that Sigma would default. The draft was approved by outside lawyers, who made a number of amendments before it went out. I reject BNYM’s suggestion that this meant that its terms did not necessarily reflect the view of the bank officers concerned.

BNYM says that these letters, which were disclosed late and are inconsistent with its case, are irrelevant, because Standish (a BNYM asset management company) was a separate corporate entity from the securities lending group of BNYM, and took its own credit and investment decisions. Whilst this is correct so far as it goes, I repeat my finding that from early March 2008 different parts of the bank began to adopt a cooperative approach as regards Sigma. This is borne out by the similarity in letters sent by securities lending, even if its officers were not aware of the Standish letters at the time they were sent. The Standish letters are therefore relevant. (A suggestion in BNYM’s written closings that there was insufficient context to assess such letters by reason of limits on disclosure was not correct, and not pursued in oral closings.)

Up to trial, there was one letter disclosed by BNYM sent by securities lending, which was to a US client (client “W”) dated 29 May 2008 and signed by Mr Fort. With the necessary adaptations to the securities lending context, the letter is the same as the Standish letters. It contains the same reference to “significant likelihood” of Sigma’s default, the likelihood of a diminished principal repayment (i.e. a loss), and expresses the same belief that the substitute securities are of a better credit quality and prospects than the Sigma notes.

Given BNYM’s case, this letter plainly required some explanation. Mr Zimmerhansl said in his report that, "With respect to the comments expressed by one individual from BNYM to one client, this did not represent BNYM's official position and therefore was not a conflicted view, rather it was a client accommodation effort and not a deviation from BNYM's neutral approach". By “conflicted view”, Mr Zimmerhansl was referring to the fact that the letter is inconsistent with the position taken by BNYM in this case.

Mr Zimmerhansl was unaware of the Standish letters when he made his report. However, even leaving aside the Standish letters, further disclosure shows that the client “W” letter was not the only one. The Oklahoma depositions refer to a letter dated 19 May 2008 to another US client (client “K”) in the same terms. In his deposition, Mr Fort said that this was “… a document that was provided for many clients in doing ratio trades on a separate account basis, so this was more of a standard format than a letter that I handcrafted”.

In cross-examination on this point, Mr Fort said that he thought that there may have been one other, but that he did not recall how many letters there may have been. At any rate, it is clear that Mr Zimmerhansl’s factual premise was incorrect in this respect.

BNYM’s case on these letters is as follows. First, it submits that read closely, it can be seen that the letters are merely putting a different “spin” on what is, at heart, the same message, namely that the bank believed the most likely outcome to be repayment in full at maturity, but that the risk of a worse outcome existed and could not be discounted. I reject that. The bank’s case has been (to quote from its closing submissions) that it “reasonably concluded that Sigma was unlikely to default prior to March 2009”. These letters express the opposite conclusion.

BNYM submitted that not too much should be read into the word “significant” when attached to the likelihood of default. In re-examination, Mr Fort, who was the signatory, said that “above 5 per cent would be significant” (Mr Gordon agreed also in re-examination). I reject that as an implausible interpretation of the term “significant likelihood” of default. Clearly, a much greater probability of default is envisaged, and hence the reason for the ratio trade.

In that regard, a proposed motion that was prepared for client “W’s” investment committee recited that the client “faces a substantial loss of principal, possibly up to 100% of the Sigma notes”. I accept BNYM’s submission that the timing and content of the various emails does not make good the contention that this information came from BNYM. However, the proposed motion was emailed to Mr Fort in draft, and if he disagreed he could have challenged it. He said in cross-examination that it was likely that he would have telephoned to say that he disagreed with this language, but there is nothing in the contemporary material to support this. I do not accept his evidence on this part of the case unless supported by the contemporary material.

In fact, the evidence suggests that the terms of the letter were much clearer about the risks associated with Sigma than the impression given by Mr Fort in phone conversations. In an email of 20 May 2008 to Mr Fort, the client commented (in relation to a draft of the letter):

I further reject BNYM’s suggestion that the ambit of “significant likelihood” of default was restricted to “medium term” maturities later than AP’s maturities. This is contrary to the terms of the letters which are in respect of “Sigma Notes” with stated maturities. Insofar as Mr Zimmerhansl supports this suggestion, I do not accept his evidence in this respect.

However, the primary explanation given by BNYM to explain the letter is as follows. It is said that it was provided at the client’s request to ease the passage of the ratio trade, and that its contents did not in reality represent BNYM’s view. To quote its closing submissions, the client “… sought a document in writing that could be used to justify expanding their investment guidelines so as to permit a trade that involved not merely reducing their Sigma exposure, but at the same time taking on an increased exposure to a variety of long-dated assets during a period of severe illiquidity in the credit markets. Given that there was already a letter in existence that had been approved by external counsel, and which conveyed the message sought by the client, it is hardly surprising that this was the document that was used, nor that Mr Fort did not feel it appropriate to make changes to the already approved language”.

However carefully it is put, this explanation, as AP said, amounts to a suggestion that Mr Fort (and by extension the bank) was involved in a deception of the client by sending a letter which he knew to be false for the purpose of getting the client’s investment committee to approve a ratio trade. No bank would contemplate such a course, and I regard the explanation as unsustainable.

In any case, this was not an isolated letter, because apart from the letter to another securities lending client, Standish was writing to its clients in the same terms at the same time. Further, the explanation does not find support in the other material. Neither the client’s email of 20 May 2008 that I have quoted above nor Mr Fort’s reply suggests that the letter had been sought by the client to be used to justify expanding its investment guidelines.

I agree with AP that the more plausible explanation is that Mr Fort believed in the matters stated in the letters. That is my finding on this issue. I do not agree with BNYM’s suggestion that the letters are “unrepresentative outliers”. On the contrary, I consider that they are consistent with most of the other contemporary documents at this time.

By this point in time, I am satisfied that Mr Tant agreed with this view. In cross-examination he said that he largely agreed with the letter, but would have taken out the word "significant". On that basis, the letter would have read, “… we believe that there is a likelihood that Sigma will not be able to continue to meet its commitments in respect of the Sigma notes in the medium term”. As to the words “diminished principal repayment”, he would have been happy with “potentially diminished”. This evidence is consistent with the spreadsheet he circulated on 1 May 2008.

BNYM submits that it was its considered view, as communicated to other clients, that the current market prices for Sigma MTNs did not reflect their intrinsic value but were a reflection of market illiquidity, and that clients would be better served by holding them to maturity than by selling the notes at such prices. The documents it relies on are as follows.

On 13-14 May 2008, a draft "representative letter" was circulated by BNYM to be sent to clients participating in one of the Mellon cash collateral funds signed by Ms Rulong. It was envisaged that there would be a different standard letter for each Mellon commingled fund, and none for the ICR fund (which was a legacy BoNY fund), or for separate accounts. The draft letter noted the recent drop in Sigma’s IDC pricing, and stated:

There is no evidence that this letter was sent, but it received attention in London, where the AP client relationship function was based. (BNYM’s credit analysis and market functions were carried out in the US.) There was a suggestion that the draft letter should be attached to monthly statements, presumably for clients generally, but on 15 May 2008, Mr Nigel Taylorson who was Head of Relationship Management (Pensions) said that the letter only applied to a particular fund, and that “…we will not send [the letter] since we will be making calls instead. I will discuss individual client plans with you individually”.

Later that day, Mr Taylorson emailed saying that “…I am not proposing to send letters and instead call just [client X] and AP, but nobody else. In all these cases we will be reactive only but we are ready to answer the questions if they arise (the chances are very strongly that they will not!)”. Among the reasons he gave for not contacting other clients were that their exposures were tiny, and that “we believe that the situation is gradually improving and if left to maturity in all likelihood no loss will be incurred”, and that contact was “more likely to create the problem we are trying to avoid”. (This seems to be a reference to clients pulling out of the programmes.)

On 16 May 2008, a Q&A document was circulated dealing with particular Mellon funds (AP of course had a segregated account), but BNYM says that it was to be used by relationship managers on client calls about Sigma. The document says that BNYM does not believe that the reduced prices would “result in an actual impairment of the Fund that will be ultimately realised”:

“Q. What is the outlook for SIGMA?

It is extremely hard to make concrete predictions. Analysis by Moody’s, which downgraded SIGMA by five notches from AAA to A2, shows the vehicle has sold off $9.5bn of assets in the open markets to repay debt, while persuading some note holders to take about $4.4bn of underlying assets in lieu of repayment.

At the same time, SIGMA has been working its banking contacts to arrange as much funding as possible in the form of repurchase, or repo, agreements, raising about $14bn from 17 different lenders.

Q. What is the outlook for the Funds?

We do not believe that the application of the above mentioned market prices to the SIGMA senior notes, and the resulting sharp decline of the Fund’s shadow NAV, will result in an actual impairment of the Fund that will be ultimately realised.

In fact, we believe that the Fund will receive higher values for its SIGMA senior notes by utilizing a ‘hold to maturity’ or other tactical disposition strategy.

This is consistent with the cash collateral reinvestment strategy BNY Mellon Asset Servicing has followed on your behalf to navigate through these difficult market conditions.

Accordingly, BNY Mellon Asset Servicing continues to be able to operate the Funds for all daily loan and maintenance activity on an amortized cost, $1/€ NAV per unit basis.”

In the light of the other BNYM material from this time set out above, I do not accept its submission that the draft client letter and the Q&A reflected the considered view of the bank as to the risks attendant on holding the Sigma MTNs at this time. At best, they gave a partial view. However, the documents were tailored to commingled funds, and for clients in funds (unlike Sigma) the important point was probably the last one made in the Q & A, that BNYM would continue to operate the funds for all daily loan and maintenance activity on a $1/€ NAV per unit basis.

In any case, neither document was sent to AP. Though AP was to be called by phone, no letter was sent to it. Clearly, a letter could not have been sent that reflected the statement in Mr Taylorson’s email that “we believe that the situation is gradually improving and if left to maturity in all likelihood no loss will be incurred”. This would have stated the opposite conclusion to the letters that securities lending were sending to other clients at the same time.

The call to AP was made on Friday, 16 May 2008 to Ms Hammarlund by Mr Gordon and Ms Jobling, the Client Service Officer who dealt with AP in relation to securities lending (she is no longer employed by BNYM and did not give evidence). No advance notice of the call was given, as might have been expected, and Ms Hammarlund described it as coming out of the blue. It was a short conversation lasting about five to ten minutes.

It was the first time that Ms Hammarlund had been contacted about Sigma (or any security) before. On the BNYM side, Ms Jobling seems to have known something about the Sigma issue, but Mr Gordon only knew what he had learned the day before. I doubt either was in a position to give any real analysis of the position. They clearly knew nothing of the background as set out above.

There is no dispute of substance as to what was said. Mr Gordon says in his witness statement that, “We informed Ms Hammarlund about the recent repricing and expressed the view that although the Sigma MTNs had been marked down there was no cause for panic as the Bank believed that the Sigma tranches would be paid in full at maturity. Ms Hammarlund appeared calm and said that she would pass the information on to Ms Thomasson-Blomquist”. Though she was calm, Mr Gordon said that Ms Hammarlund understood the seriousness of the pricing reduction.

Ms Hammarlund made a note of the call because she needed to keep Ms Thomasson-Blomquist (who had responsibility for the financial performance of AP’s securities lending programme) informed. The note (in English translation) records that she had had a discussion with BNYM about the balance in AP’s cash reinvestment portfolio. In relation to Sigma “there is a balance that they are intending to keep till maturity … which equals about 2% of the cash reinvestment program”.

She reported that, “They had a maturity recently which was processed without any problems. Sigma pays all its interest due without any problems and has got good reputation. At the same time the pricing in that case for the cash reinvestment portfolio was based on the rate of 80, since another player left the position prior to maturity. Mellon says they are completely confident that Sigma is going to fulfill its obligations, but cannot obviously guarantee anything. They want to inform their customers, but believe that one should sit still for a while. They monitor it daily.”

Except for the part that records that BNYM was “completely confident” that Sigma was going to fulfil its obligations, it broadly accepts this note. Perhaps those precise words were not used, but I am satisfied that the message was a reassuring one. This follows from his witness statement, in which Mr Gordon says that the bank’s view was that “if held to maturity the Sigma MTNs would pay with no loss to our clients”. He accepted in cross-examination, obviously rightly, that he would not have been speaking to Ms Hammarlund in the terms he did if he had been aware of the letters that BNYM was sending to other clients at the time.

I do not accept the bank’s submission that the purpose of the call was simply to alert AP to the situation and ask if it wanted a fuller discussion. Mr Gordon did not regard this as a preliminary call, to be followed by a more detailed briefing. He says that the request for a further call came from AP. This was prompted by Ms Hammarlund’s note to Ms Thomasson-Blomquist saying that, “If you need more information, I can arrange a teleconference with Steve and Verity on Monday. Verity can also send more information about Sigma”. A further conference call was requested by Mr Grunditz and was fixed for Monday, 19 May 2008.

Mr Gordon said in cross-examination that a “more balanced” picture was given on 19 May 2008, and this is the bank’s case. The call took place at 10.30 am, and AP was represented by Ms Thomasson-Blomquist and Mr Grunditz (Ms Hammarlund could not attend), and BNYM was represented by Mr Gordon, Mr Tant and Mr Fort, (Mr Skeel stood in for Ms Jobling). The only record is in the form of notes jotted on an email by Mr Gordon. That email was from Mr Skeel to Mr Fort and Mr Gordon setting up the call, and said that it “is just to provide a little more information and speak to additional people at the client who were not present on Friday”.

Mr Grunditz's evidence was that, “To the best of my recollection we were informed by BNY Mellon that although the price of the Sigma notes was 80% during a recently performed transaction, Sigma was a company which had recently met its payment obligations and BNY Mellon were confident that Sigma would be able to meet its future obligations. BNY Mellon’s view was that we should not rock the boat and that we should continue to hold the notes – which were set to mature in October 2008 and March 2009”. BNYM points out that this is similar to Ms Hammarlund’s note of the earlier call. He accepted that he could not remember “verbatim” what had been said.

Ms Thomasson-Blomquist’s evidence is that the purpose of the call was for BNYM to provide additional information, and that BNYM informed AP that it was not their intention to sell the Sigma MTNs. She said that, “Based on the information they gave us both in Verity Jobling’s call to Gun Hammarlund and during our telephone call, I did not think there was any reason to question their decision to hold onto the Sigma notes”.

She said in cross-examination that they were told that they had no reason to worry about Sigma. She did not accept that they were told that, “No SIV can really exist going forward”, which is a statement in Mr Gordon’s note. Since the note is the only contemporary record, I accept that this was said. However, it is of limited relevance, since the issue was not whether SIVs could go forward (the market understood that they could not), but whether Sigma would nevertheless pay on maturity. Ms Thomasson-Blomquist said that the purpose of the call was not, as she understood it, that BNYM wanted further instructions or a decision from AP. I accept that evidence, and find that AP was not asked to and did not take a decision. She did however agree in cross-examination that there came a point in the conversation when the bank, having explained its position, asked if AP was happy to go along with it, and AP said something like, “We agree with you and there is no need to sell or to exit immediately”.

On BNYM’s side, Mr Fort said in his witness statement that he does not remember the call. In his oral evidence, he said:

Mr Tant said that he did not recall the call in detail, but that it did not focus solely on Sigma, but also addressed how the portfolio might be managed in the prevailing market environment. He recalls that AP was informed that Sigma was facing challenges, but that BNYM retained confidence in it because it had been successful in finding other sources of liquidity and doing the right things against a difficult market backdrop, such as boosting liquidity and de-leveraging.

He says that, “I also believe that during the call, the possibility of a ratio trade was mentioned. While I cannot recall this specifically in relation to the AP call, I recall that generally on client calls regarding Sigma, Bob Fort would explain that there was an opportunity for a client to enter into a ratio trade”. The possibility of a ratio trade is not mentioned in Mr Gordon’s note and is not supported by the other witnesses, and my finding is that the subject of ratio trades was not mentioned in these phone conversations.

The most significant participant on BNYM’s side was Mr Gordon, who was Relationship Manager for AP, and the only person to be on both calls. In his witness statement, he refers to his note, and says that, “… we repeated that there was no need to panic. Mr Tant and Mr Fort explained that the price reduction was the result of the sale of a Sigma note by a hedge fund that had led IDC to drop all of its prices, however they remained confident that Sigma would continue to perform and pay out at maturity on the October 2008 and March 2009 notes held by AP. They explained that Sigma’s underlying assets remained of high quality and that a recent maturity had been paid out in full. AP agreed that there was no need immediately to exit Sigma but asked if we could think of ways to reduce its risk without sacrificing revenue”.

Leaving aside the unexplained reference to sale by a hedge fund, BNYM says, I consider rightly, that Mr Gordon had a much clearer recollection than Mr Tant or Mr Fort. However, I do not think (as it submits) that “the telephone call involved an in-depth discussion of Sigma”. Nothing in the evidence supports this, and I consider that it is more likely that the discussion was superficial.

As I have stated, Mr Gordon’s witness statement says that Mr Tant and Mr Fort told AP that “they remained confident that Sigma would continue to perform and pay out at maturity on the October 2008 and March 2009 notes held by AP”. When the contemporary bank documents with a different message were put to him in cross-examination, he qualified this saying that AP was told that, “This asset type is finished, but for 5 months, for 10 months, we’re hoping it's going to be okay for you”. In re-examination, he agreed with the passage from Mr Fort’s evidence that “significant likelihood” of default meant a 5 per cent chance (an interpretation that I have rejected).

I accept that the bank did not give the impression that there was a guarantee. Otherwise, I found this further evidence of Mr Gordon unconvincing, and do not accept it. I find that his witness statement gives an accurate summary of the call, and that AP was told, not that BNYM was “hoping it was going to be okay”, but that it remained confident that Sigma would pay out at maturity on the notes that AP held.

Based on another note that he found, Mr Gordon thinks that there was another call between himself and Ms Hammarlund on 20 May 2008. This was only produced after AP’s witnesses had given their evidence, but in any case I do not think that it takes matters much further. The note records the fact that at current prices, AP faced a US$10.5 million loss on the Sigma MTNs, which is not in dispute.

In an internal email of 21 May 2008, Ms Jobling said that she had found out that AP was a “bit scared by the Sigma holding”. She had to go back to AP with some additional information about reducing risk, but “they’re not looking to force a sale of any assets that are under water but now they are aware of the situation they are looking to react quickly”.

On 23 May 2008, Ms Jobling sent a lengthy email to Ms Hammarlund and Ms Thomasson-Blomquist dealing with the management of risks, as they had asked, and making various suggestions. The email goes on to deal with Sigma, which it describes as an “‘impaired’ asset”. In the course of the email, it raises the possibility of an “asset swap”:

“We have undertaken a review of your cash collateral investment pool and I can advise that we’re currently running very high liquidity within the fund (around 32%). … We believe this to be a strong portfolio, with SIGMA being the only ‘impaired’ asset. We have reason to believe that we will see improvements of this asset at May month end, however this does depend on any trading activity that may take place in the market of this security between now and then.

As you know, our strategy is a buy and hold one. We’re intending to hold these positions until maturity. SIGMA is still making timely interest payments, and there was a full maturity on the 14th May 2008 of one line which we hold within our other funds. One option available to us would be to do an asset swap for a portion of the SIGMA position within your portfolio. This is something we’d need to look into further if you were considering this option. The main issue with this is that the assets we would receive would have a longer duration than those currently permitted within your reinvestment guidelines. We would only be able to do this for those assets that we are comfortable with and of issuers that we have previously completed our due diligence on. Can you let me know if this is something you’d like us to take further for you?

…

I hope that I have given you some food for thought. Depending on which route(s) you wish to go down, we will have to undertake some additional analysis. Please feel free to contact me if you have any questions on any of the above points and I hope to hear from you soon.”

BNYM submits that it thereby suggested an asset swap (i.e. ratio trade) to reduce AP’s Sigma exposure. This would require AP to take assets outside its guidelines, and it asked AP for guidance as to what it wanted the bank to do and whether it required any more information. However, it says, AP fell into a torpor. It never responded to Ms Jobling’s request that it contact her if it wanted more information about a possible asset swap. The reason is (it submits) that AP had no interest in concluding such a trade. Nor did it ask BNYM for a recommendation. It did not regard selling the Sigma MTNs as an attractive option, because this would involve crystallising a loss. AP appreciated full well that there was a risk that the prices of the Sigma MTNs would fall further and that the only way to avoid any exposure to further price falls was to sell them. In choosing to keep them, BNYM submits that AP accepted the risk of any further falls in their value, including to the point of default.

AP’s evidence was as follows. In cross-examination, Mr Grunditz said that for AP to have taken action there would have had to have been a clear analysis as to why BNYM made the proposal, so as to give it something to base its decision on. A sentence in an email about a swap, he said, was not enough. He said that rather than explore an asset swap, AP decided to tighten its investment guidelines. When it was put to him that ratio trades had a disadvantage in that AP would in effect be buying parts of Sigma's portfolio at the par value when in fact they were not trading at par, he said “… if there would have been a better result in the long run or overall to do that, and if we had good documentation to support it, we could have done so. Everything is based upon the fact that you have good documentation with a good analysis and a proper recommendation from Mellon that we can actually go with”. He said that if BNYM had given a recommendation to sell, AP would have considered doing so.

In cross-examination, Mr Rahmn said that, “If a situation arises whereby we are expected to take a decision or where the bank considers that we should take a decision, then, of course, there had to be adequate information with analysis and proposals for what … it is we have to take a decision on, and that's happened in the past where there has been … a situation… whereby [BNYM] has requested that AP takes a decision and they have also presented us with various proposals to decisions with the recommendations and also with supporting explanations as to why”.

Similarly, Ms Thomasson-Blomquist denied the suggestion that AP had a general approach of seeking to avoid crystallising losses. Her evidence was that “in view the information we had been given concerning Sigma, there seemed no need to swap all or part of AP’s position in Sigma (whose price BNY Mellon expected to recover) for assets of an unknown quality and which would require us to relax our investment guidelines”. AP’s witnesses were generally reliable, their evidence on this point is credible, is not contradicted in the contemporary material, and I accept it.

I have set out above my finding that a securities lending agent is expected to communicate with its client if there is a serious risk of loss in respect of securities held as collateral for the client’s account. I consider that such a duty arose in this case in May 2008 after the reduction in the pricing of the securities by IDC. However, this is academic, because BNYM did in fact contact AP about Sigma at this time. Having done so, BNYM accepts that it was under a duty to do so fairly. This was a correct and inevitable concession.

The parties’ respective cases in summary are as follows. BNYM submits that AP was given a balanced picture of Sigma. AP submits that representations were made to AP in the phone calls that were false or misleading, and that BNYM did not take reasonable care in its communications.

I begin by noting that client calls following the IDC price drop must have been difficult, and I infer that the securities lending officers had little or no experience of this kind of exercise. As explained earlier, collateral in this business was (in Ms Demmler’s words) expected to be “money good”. Something of the flavour is caught in Mr Ford’s email to Mr Palermo of 31 January 2008, in which he says that “… the treatment of clients in the situations that may possibly present themselves will be in fairly uncharted territory”. That was a fair comment, in my view.

BNYM submits that the bank was not, as it put it, dealing with ingénues. It is correct, and pointed out elsewhere in this judgment, that AP was an investment professional. Ms Thomasson-Blomquist accepted that she knew that SIVs had problems in the market. However, AP clearly could not be expected to have performed the kind of scrutiny of Sigma that had been undertaken by BNYM over the previous months. There was no specific alarm bell ringing until the beginning of May, since Sigma prices were reported by BNYM as close to par. In the circumstances, it was reasonable, in my opinion (as its witnesses said) for AP not to question BNYM’s expertise.

I am satisfied that the view conveyed to AP in the May communications was that BNYM was confident of repayment, and that the notes should be retained. Although BNYM may not have been under any obligation to give a recommendation (because it was not acting in an advisory role), it did in substance do so. AP would probably have asked if it had not.

What was said has to be placed against the other evidence. I have set this out in detail above, and need not repeat it. It is sufficient to compare the position of client “K” which like AP had Sigma holdings maturing on 30 October 2008 and 12 March 2009. The letter to client “K” is dated 19 May 2008 (the copy in evidence is not signed), that is the same day as the second call to AP. As explained above, the letter says that BNYM believed that there was a significant likelihood of a default by Sigma which would likely result in a prolonged liquidation period as well as diminished principal repayment. This client was told that it was better to have substitute securities than the Sigma notes, and encouraged to waive its investment guidelines so that a ratio trade could be performed. Such a trade would have the disadvantage, as BNYM points out, of the client taking on an increased exposure to a variety of longer-dated assets. But that was deemed preferable to continuing to hold Sigma.

These client communications were not, as BNYM submitted, different versions of the same theme, but are opposite, and irreconcilable. The difference in the impression made on the clients is (in my view) shown by the email of 20 May 2008 from client “W” to Mr Fort (see above). It is clear from that email that after talking to Mr Fort, the client’s impression was that the “wise move was to wait and see how things progress”. This was similar to the impression that was given to AP. But then the client read the letter, “and the way that was worded made it seem much more likely that Sigma's position was in trouble … I didn’t think things were that dire”.

The bank’s view of Sigma can be seen demonstrated in the ratio trades it was carrying out, in particular, by reference to client “I”. As noted above, because of the breach of that client’s guidelines, any loss arising on the Sigma holding would be for the bank’s account. It decided not to retain the MTNs to maturity, but disposed of them by way of ratio trade. This reflected its belief at that time that Sigma was likely to default by September 2008 eventually leading to a realised loss. So far as the evidence in this case is concerned, that was essentially the view expressed in the client letters, which, whatever differences of opinion there may have been internally, and whatever was being said on the phone at the time, must in my opinion be taken to state the bank’s position.

In communicating with AP about the Sigma notes, the bank had to give a fair view of the risk of default and loss, and what the alternatives were to holding the securities to maturity. On the facts, I conclude that BNYM’s communications with AP about Sigma at this time did not give a fair view. I make it clear that I am satisfied that there was no deliberate attempt to mislead (and reject any case based on falsity), but that was the unintended effect. By telling AP that it remained confident that Sigma would pay in full, when its analysis at that time was that there was a significant likelihood of default, BNYM seriously played down the risks. A balanced picture would have made it clear that there was real concern within BNYM about the creditworthiness of Sigma and the possibility of default and loss of capital to MTN holders such as AP in a liquidation.

Further, AP was not told in the phone calls that there was an alternative to either a sale, or holding to maturity, in the form of a ratio trade. There was a reference to the possibility of an asset swap in the email of 23 May 2008. However this has to be seen in the light of what AP had been told about the likelihood of default. The fact that AP’s guidelines precluded such a trade, as BNYM has emphasised, is not an answer, because other clients were being invited to relax their guidelines, and AP could have done so too, had the matter been put to it properly. I should add that the case of both parties in oral closing submissions was that at this time there were assets of sufficient quality left in Sigma’s portfolio to make a ratio trade possible. (It appears that no ratio trades were carried out between BNYM and Sigma after June 2008 because Sigma no longer had assets of sufficient quality available to meet the requirements of securities lending clients.)

There was nothing in the email of 23 May 2008 to modify what AP had been told on 19 May 2008 in the phone call. BNYM confirmed that its strategy in relation to Sigma was a buy and hold one, and that it intended to hold the positions until maturity. The reference to an asset swap had none of the detail contained in the letters to other clients as described above. In those circumstances, I am satisfied that AP cannot be criticised for not pursuing the matter. Given what it had been told about BNYM’s view as to repayment of the notes, it is also not surprising that AP did not instruct BNYM to sell them, thereby incurring a large loss. For the same reasons, nothing in my view turns on the fact that on 17 June 2008 AP did in fact instruct BNYM to tighten its guidelines.

That being the factual position, I turn to the legal analysis. AP’s case is that BNYM made representations to AP to the effect that (a) there was no need to be concerned about the Sigma MTNs but that the fall in their price was attributable to lack of liquidity in the market rather than problems with Sigma or the MTNs themselves, and that (b) though there were no guarantees, there was no reason to believe that the Sigma MTNs would not pay out in full at maturity. AP submits that BNYM (1) did not take reasonable care in making the representations and/or did not have reasonable grounds for making them, (2) alternatively, what was said was misleading in that relevant facts were omitted in circumstances where BNYM had a duty of full disclosure of material facts to AP by reason of its agency and/or fiduciary relationship, (3) the representations constituted negligent misrepresentations and/or misleading partial representations, and (4) BNYM did not make full disclosure in relation to the option of a ratio trade in its email of 23 May 2008.

In written closings, BNYM submitted that (1) the views expressed by the bank in the telephone calls of 16 and 19 May 2008 regarding the prospects for the Sigma MTNs were non-actionable statements of opinions that, in any case, were reasonable and reflected the bank’s considered view of Sigma at the time, (2) it was not negligent of the bank to provide such views to AP, nor was the bank under any obligation to provide additional information regarding the range of internal (or market) views about Sigma, (3) BNYM owed AP no fiduciary duty to give “full disclosure” of information regarding Sigma, nor of the steps it had previously considered in relation to the ICR fund, nor, in the absence of a request from AP, did BNYM owe AP any duty to provide AP with additional information or a recommendation regarding the ratio trade suggested in the email of 23 May 2008, and (4) As a result of these communications, from 16 May 2008 (or, at the latest, 19 May 2008), AP was aware of, and accepted, the risk of further declines in value of the Sigma MTNs.

Mr Hapgood QC rightly accepted that if I found that the bank’s presentation of the Sigma risk to AP was not a fair one, it would incur liability to AP, and that no points arose on the pleadings. In the light of that concession, discussion as to the legal analysis in closing was relatively limited. That is my finding, and except for its submission that it did not owe fiduciary duties to AP, which I accept, I reject its points set out above. I reject the suggestion that the statements made were non-actionable statements of opinion, since BNYM factually misstated what its view of Sigma was, and I find that it was negligent of the bank to tell AP that it remained confident that Sigma would continue to perform and pay out at maturity on the October 2008 and March 2009 notes. While I agree that BNYM owed AP no fiduciary duty to give “full disclosure” of information regarding Sigma, it accepts that it had to give sufficient information to put the position fairly, and I am satisfied that it did not do so. Whilst it is correct that AP continued to hold the securities, it did so on the basis of the defective communications made by BNYM. For negligent misstatement purposes, it was not in dispute that this was a relationship of proximity within the Hedley Byrne line of case law.

It follows that I accept AP’s submission that in communicating with it in May 2008, BNYM did not take reasonable care in making the representations it did about the Sigma MTNs, and that what was said was misleading in that relevant facts were omitted which were necessary to give a proper picture. I accept its submission that the representations made constituted negligent misrepresentations and/or misleading partial representations. I consider that BNYM did not adequately put to AP the option of a ratio trade in its email of 23 May 2008. I accept AP’s case on negligent misstatement/negligent misrepresentation.

However, I consider that the preferable analysis is based on breach of duty. As stated above, by clause 10(b)(i) of the Securities Lending Authorisation Agreement, the bank as lending agent had to perform its obligations under the agreement with the care, skill, prudence and diligence which a prudent securities lending agent would use. In my opinion, where a securities lending agent contacts its client about risks concerning a security held for the client’s account, it must present the risks in a fair manner, both under its contractual duty of care, and under the equivalent duty of care that arises in tort. In the case of the Sigma notes, the bank had to give a fair view of the risk of default and loss, and what the alternatives were to holding the securities to maturity. Effectively, the bank’s witnesses accepted that this duty arose as a matter of good practice, and I consider that it arises equally as a matter of law. In performing this task in May 2008, BNYM fell well short of the standards which it rightly set itself. I consider that AP has made good its case in this respect.

The parties’ submissions on these three topics were understandably brief. As to causation, BNYM contended that there are two aspects to consider. There is factual causation, that is, whether as a matter of fact BNYM’s breach caused the loss for which AP claims. Then there is what it calls “hypothetical causation”, which is whether, if the bank had sought instructions from AP whether to sell the Sigma notes or enter into a ratio trade, AP would have given such instructions.

The point which BNYM makes on factual causation is that from either 19 May 2008, alternatively from 17 June 2008 when AP instructed BNYM to tighten its guidelines, AP knowingly accepted the risk that an asset which it knew to be “impaired” issued by an entity which it knew to be in wind-down would fall further in value. It accepts that if I find that it misrepresented the position to AP or put it in a way that fell short of its duty of care, the point does not arise, since the acceptance of risk would have been induced by BNYM’s fault. I have made that finding, and so need say no more on this point.

The point which BNYM makes on “hypothetical causation” is on the basis that in the run up to trial Mr Grunditz and Ms Thomasson-Blomquist put in supplemental witness statements to the effect that if BNYM had expressed views in the same terms as Mr Servis and Mr Cirrito, AP would have instructed BNYM to exit AP’s Sigma position. BNYM describes this evidence as opportunistic.

However their evidence was not posited only on AP’s expert evidence, but also on the facts as then known by the witnesses. The disclosure given by the bank of the Oklahoma depositions and the further documents referred to earlier in this judgment strengthened AP’s case as to what BNYM thought about the Sigma risks at this time. This is basically a factual matter. I do not accept, therefore, that AP’s case on “hypothetical causation” cannot succeed unless the court prefers the evidence of AP’s experts to that of BNYM’s experts (this suggestion was not pursued in oral closing).

I have referred to the evidence of AP’s witnesses above. In his evidence Mr Magnusson, who became Managing Director of AP in March 2008, said that in view of the serious situation affecting Sigma, and had BNYM given it full information, he would have been concerned to protect AP’s assets and he believes his decision would have been to exit the Sigma exposure as quickly as possible and in the way that best protected AP's interests. He was a good witness, and I accept his evidence, as well that of AP’s other witnesses in this respect.

I accept AP’s submission that had AP received a letter in similar terms to that addressed to client “K” on 19 May 2008 or client “W” on 29 May 2008, AP would have exited its Sigma exposure, whether by ratio trade or by sale at the price available, realising proceeds of the order of US$25-26m for its Sigma MTNs. I am satisfied that it has proved its case on causation.

As to foreseeability, BNYM’s contention is that the actual risk that materialised was not within the contemplation of the parties. It submits that “… the losses claimed by AP result from a combination of exceptional circumstances: the sudden loss of liquidity in credit markets from August 2007, and the financial tsunami that resulted from the failure of Lehman Brothers thirteen months later. Both were unprecedented, and neither can be said to have been likely or within the reasonable contemplation of the parties at the time the GCA and SLAA were agreed in 2004”.

I doubt that the assertion that the “financial tsunami” resulted from the failure of Lehman Brothers is correct. This was part of the crisis that came to a head in September and October 2008. As Rix LJ said in Rubenstein v HSBC Bank plc [2012] EWCA Civ 1184 at [118], “… although the Lehman Brothers collapse was both a symptom and a contributory cause of market turmoil, the underlying causes of that turmoil went infinitely beyond Lehman Brothers' difficulties. It stretched to a failure of confidence in marketable securities in which there had previously been greater confidence. And what is new about that?” As he said about the issue as it arose in that case, “The insolvency of Lehman Brothers may have been unforeseeable, but Mr Rubenstein was not invested in Lehman Brothers” (also at [118]).

The facts are different, but it seems to me that this approach is applicable here too. It was within the reasonable contemplation of the parties (see The Achilleas [2008] UKHL 48) that an issuer of securities in which cash collateral had been invested would get into difficulties, even if the credit crunch and subsequent financial crisis had never happened. It was also within the reasonable contemplation of the parties that if BNYM failed to present the position about an issuer in difficulties fairly to a noteholder like AP, when it had a duty to do so, a loss would be incurred which was otherwise avoidable. The issue does not depend on the foreseeability or otherwise of Lehman’s collapse, and I consider that AP was justified in not cross-examining Professor Hubbard on that subject.

BNYM has sought to draw a distinction between foreseeability of default, which I think it accepts, and foreseeability of resultant loss. The suggestion is that the bank believed that even on a default, Sigma’s assets (taking into account the capital cushion) would be sufficient to meet the claims of senior noteholders like AP in full. This point fails on the facts. As I have said elsewhere in this judgment, the failure of Sigma and resultant loss to holders of its debt was entirely foreseeable. The evidence is that it was foreseen by February 2008, and the position thereafter deteriorated. There is ample evidence that BNYM was well aware that loss to noteholders could and probably would result.

In short, the failure of Sigma with resultant loss to holders of its securities was not only foreseeable, but was foreseen by BNYM. As AP said in its closings, there is “copious evidence” to the effect that senior management of BNYM were very concerned indeed about the significant likelihood of default of Sigma and the serious risk of loss of capital to Sigma MTN noteholders, particularly those holding post September 2008 maturities like AP. They were right to be, as its assets disappeared in repo loans, and ratio trades. For these reasons, I reject BNYM’s case on foreseeability.

As to contributory fault, the main criticisms made by BNYM were that AP (1) failed to monitor its portfolio, a criticism aimed in particular at Mr Grunditz, (2) failed to include SIVs in the list of prohibited investments at the end of clause 2 of the Investment Guidelines, (3) failed to appreciate the potential impact of the “credit crunch” on its portfolio, (4) failed following its communications with BNYM in May 2008 to take any steps independently to investigate the creditworthiness of the AP Sigma MTNs (BNYM says that it could have asked BlackRock as a specialist external manager for US corporate bonds for advice), (5) rejected a ratio trade without seeking further information from BNYM about such a trade, and (6) instructed BNYM in June 2008 not to sell any assets at a loss, despite being aware of the increased credit risk being run by the AP Sigma MTNs.

As to (2), this could only apply to AP’s acquisition claim, which I have rejected. As to the others, AP has succeeded on the third claim (communications made to it in May 2008), because of the deficiencies in the way that BNYM put the risks inherent in the Sigma holding when it contacted AP in May 2008. I am satisfied that AP cannot be criticised on any of the grounds asserted by BNYM, given the way the position was presented to it. Mr Zimmerhansl’s evidence on this point does not give sufficient weight to this. I refer to the discussion above in this respect which I shall not repeat. I only add that the criticisms I have made as to AP’s monitoring of the securities lending operation do not apply in this context, and nothing was said by BNYM in the May communications that indicated the need for a second opinion on the Sigma MTNs from BlackRock. In those circumstances, I accept AP’s submission that its loss was not caused by or contributed to by any of the factors identified in BNYM's submissions.

The parties are agreed that if I find in favour of AP on liability, as I have on its third ground, they will be able to agree quantum. If for any reason that should not be possible, I will give a further ruling.

I find against AP on its claims relating to the acquisition and retention of the Sigma MTNs, but find in its favour on its claim in relation to how the position was explained to it in May 2008. The overall result is that AP is entitled to judgment. I am grateful to the parties for their assistance, and will hear them as to any consequential matters.