The opinion of the court was delivered by: Richard J. Holwell, District Judge:
MEMORANDUM OPINION AND ORDER
Plaintiff Prudential Retirement Insurance and Annuity Co. ("PRIAC"),
brought this action pursuant to sections 409(a) and 502(a)(2) and (3)
of the Employee Retirement Income Security Act of 1974 ("ERISA")
against defendant State Street Bank and Trust Company ("State Street")
on October 1, 2007. PRIAC commenced this suit as an ERISA fiduciary on
behalf of nearly 200 retirement plans (the "Plans") that invested,
through PRIAC, in two collective bank trusts managed by State
Street-the Government Credit Bond Fund ("GCBF") and the Intermediate
Bond Fund ("IBF") (collectively, the "Bond Funds"). This Memorandum
Opinion and Order follows a seven day bench trial on the issue of
whether State Street breached its fiduciary duty to the Plans by (1)
failing to manage the Bond Funds prudently, (2) failing to
manage the Bond Funds solely in the interest of the Plans, and (3)
failing adequately to diversify the Bond Funds' assets.*fn1
The Memorandum Opinion and Order sets forth the Court's
Findings of Fact and Conclusions of Law in accordance with Federal
Rule of Civil Procedure 52. For the reasons that follow, the Court
finds that (1) State Street breached its duty to manage the Bond Funds
prudently, (2) State Street did not breach its duty to manage the Bond
Funds solely in the interest of the Plans, (3) State Street breached
its duty to diversify the Bond Funds, (4) State Street's breaches
caused losses to the Plans, and (5) PRIAC's calculation of damages is
PRIAC is a Connecticut corporation that was established in 2004 when Prudential Financial, Inc. ("Prudential") acquired CIGNA Retirement & Investment Services ("CIGNA") and renamed it PRIAC. (Stipulations of Fact ¶ 1.) PRIAC provides investment options to defined benefit and defined contribution retirement plans; it provides these options to over 7000 organizations and three million participants and beneficiaries.
State Street is a bank organized under Massachusetts law; State Street Global Advisors ("SSgA") is State Street's investment management arm. (Stipulations of Fact ¶ 4.) SSgA provides asset management services to State Street clients.
State Street managed two collective bank trusts that PRIAC made available to its retirement plan clients: the Intermediate Bond Fund ("IBF") and the Government Credit Bond Fund ("GCBF") (collectively, the "Bond Funds"). (Stipulations of Fact ¶ 5.) The Bond Funds were pooled investment vehicles made available to qualified groups of investors. (Id. ¶ 7.) State Street alone held discretion over the management of the Bond Funds' assets. (Id. ¶ 5.) PRIAC had no input into State Street's management of the Bond Funds, in State Street's assessment of the risks the Bond Funds incurred, or in State Street's investment decisions for the Bond Funds. (Id.)
Nearly 200 of PRIAC's retirement plan clients (the "Plans") invested in the Bond Funds through separate accounts maintained by PRIAC. (See id. ¶ 16.) PRIAC's role in this regard was to serve as an intermediary between the Plans and State Street. On August 29, 2007, PRIAC requested that State Street redeem the Plan's investments in the Bond Funds. (Id. ¶ 20.) On October 1, 2007, PRIAC commenced this action on behalf of the Plans seeking to recover losses incurred by the Plans as a result of their investment in the Bond Funds.
Each of the Bond Funds used a Lehman Brothers Index as a benchmark by which the fund's performance could be measured; the GCBF used the Lehman Brothers Government Credit Bond Index, and the IBF used the Lehman Brothers Intermediate Government Credit Bond Index. (See Stipulations of Fact ¶ 17.) As stated in the Amended Fund Declarations for each Bond Fund, the investment objective of each Bond Fund "shall be to match or exceed the return of the" applicable benchmark index. (DX 467 (for the GCBF); DX 468 (for the IBF).) The Fund Declarations also provided, "At the time of purchase, all securities purchased by the Fund will be rated at or above investment grade by either Standard & Poor's or Moody's Investor Services." (DX 467; DX 468.)
A significant amount of the testimony and evidence at trial focused on the meaning of the term "enhanced index fund" and whether State Street presented the Bond Funds to PRIAC as "enhanced index funds." PRIAC's position is that it reasonably understood from its communications with State Street that the Bond Funds were "enhanced index funds," which, in PRIAC's view, are funds that seek to "modestly outperform the benchmark while taking on minimal additional risk." (PX 10; see Tr. 63:17-21 (Blume testimony).) State Street, on the other hand, contends that it did not present the Bond Funds as "enhanced index funds," but rather as "active" funds, and that, in any event, the term "enhanced index" has no fixed industry definition, so any attempt to use it to establish the character of the Bond Funds would be inappropriate.
1.Definitions: Excess Return Target (Alpha), Predicted Tracking Error, Information Ratio
The Bond Funds' risk and return characteristics can be described with reference to two figures: targeted excess return (or alpha target) and predicted tracking error. Both of these figures are measured in "basis points." Each basis point represents a 0.01% deviation from the benchmark. The basis points for targeted excess returns represent the margin by which the Bond Funds strove to outperform their benchmark; the predicted tracking error measured the anticipated deviation from the benchmark. Predicted tracking error is a way of measuring the risk of a particular portfolio. Predicted tracking error is the variable that an investment manager can control, while targeted excess return (or alpha target) is a goal an investment manager seeks within the targeted tracking error parameters. (See Tr. 90:5-7 (testimony of PRIAC's expert Dr. Marshall Blume ("The item that an investment manager can control is the risk's characteristics. The alpha or the expected difference between the returns that you obtain and the index are goals.")); Tr. 112 (Blume testimony ("I would agree that the alpha is a goal of this fund and the tracking error is the variable that you actually can control.")).)
A third figure that is relevant to understanding the risk and return characteristics of the Bond Funds is the information ratio. The information ratio is the ratio of a fund's excess return target to its predicted tracking error. (See DX 10, at 2.) A higher information ratio is indicative of a fund that seeks to achieve alpha without a significant amount of risk. (See DX 10.) A higher information ratio "can distinguish between the skilled portfolio manager, who achieves outperformance with relatively little risk, from the 'cowboy' who achieves outperformance through very high-risk strategies." (DX 10; Tr. 775:11-14 (Reigel).) An information ratio of 0.5 is reasonable in the investment management industry, (see Tr. 817:5-7 (Reigel)); an information ratio of approximately 0.7 is "challenging," (see Tr. 1092); and an information ratio of 1 is "very optimistic" and arguably unrealistic, (Tr. 127:23 (Blume); Tr. 917:24-918:9 (Armstrong)). Each of these figures is relevant in understanding the parties' dispute about the characterization of the Bond Funds.
2.State Street's Understanding and Use of the Term "Enhanced"
There is no well-established industry-wide definition of an "enhanced index fund." (See PX 10.) There also are no well-established industry guidelines for the appropriate alpha target and predicted tracking error of an enhanced index fund. Nonetheless, the evidence at trial established some general principles.
An enhanced index fund generally seeks to "modestly outperform the benchmark while taking on minimal additional risk, thus achieving a high information ratio." (PX 10; see Tr. 63:17-21 (Blume).) "The investment management industry generally defines enhanced index strategies as investment approaches within predetermined risk and return parameters." (PX 10; see also Tr. 1187-88 (testimony of State Street's expert Dr. Andrew Carron).) Predicted tracking error for enhanced index funds typically ranges between 25 and 75 basis points. (Tr. 67:7-14.) Predicted tracking error for passive funds generally ranges between 5 and 25 basis points, and predicted tracking error for active funds typically is greater than 75 basis points. (Id.) Thus, enhanced index strategies are considered to fall somewhere between passive and active strategies. (See PX 46.) In addition, a fund whose stated investment objective is to "match or exceed" its benchmark fits "the generally accepted definition of what an enhanced index fund is." (Tr. 460 (Fischel testimony).)
b.State Street's Description of an Enhanced Strategy
A State Street document from the mid-1990s reflects the industry norms discussed above. (See PX 61.) The document is undated, but footnote references suggest that it was produced around 1996. (See id.) The document does not mention either Bond Fund by name nor does it represent a contractual agreement about how the Bond Fund would be managed, but it is relevant in assessing State Street's understanding of the term "enhanced."
The document states, "The objective of enhanced management [sic] is to add value over the index while mirroring its risk profile. As a result, our strategy combines the predictable strengths of passive management with the repeatable aspects of active management." (Id.; see also Tr. 226 (Hatfield testimony) (stating that "enhanced index" was a term State Street used to describe its products).) The document also states that enhanced index strategies usually seek an excess return of 20 to 30 basis points over their benchmark index. In addition, according to the document, "the annual guideline for maximum total tracking error is 75 basis points" for enhanced index strategies. (PX 61.) The document describes tracking error as an "ex-ante number based upon a variance/co-variance matrix developed at Lehman Brothers." (Id.) This notion of predicted tracking error as the variable that the investment manager can control is consistent with the testimony of PRIAC's expert, Dr. Marshall Blume, who opined, "The item that an investment manager can control is the risk's characteristics. The alpha or the expected difference between the returns that you obtain and the index are goals." (Tr. 90:5-7; see also Tr. 112 ("I would agree that the alpha is a goal of this fund and the tracking error is the variable that you actually can control).)
State Street understood the term "enhanced" to refer to a type of strategy that fell somewhere between purely passive and fully active. State Street's 2004 Fixed Income Procedure Guide stated: "At SSgA, we view the active risk-return spectrum as continuous and we have strategies with varying degrees of risk that span from passive, to enhanced to fully active strategies." (DX 571; see also PX 461 (graph showing "Risk (Tracking Error)" on the x-axis, "Value added (Excess return)" on the y-axis, and "Enhanced Indexing Strategies" falling between "Passive Management" and "Active Management").)
In an October 2006 presentation to business students at the Kenan-Flagler Business School, State Street described its "goal" for its "Global Fixed Income" division. The goal was "[t]o provide benchmark oriented, highly risk-controlled investment strategies at very low cost." (PX 498.) State Street's presentation also referred to each Bond Fund and its "Enhanced Index Benchmark." (Id.)
C.State Street's Description of the Bond Funds
1.State Street's Descriptions of the Bond Funds to PRIAC
In 1996, prior to the Plans' investment in the Bond Funds, State Street faxed a letter to Amy Hatfield of PRIAC that used the term "Enhanced" in the name of each Bond Fund. (See PX 82.) The investment management agreement that the parties later signed, however, did not use the term "Enhanced" in the name of either Bond Fund. (Tr. 238-39 (Hatfield); DX 470.)
In a 2002 email to Robert Frascona of PRIAC,*fn2 a State Street employee informed PRIAC that "SSgA is willing to offer a lower fee schedule for the enhanced bond funds within the Cigna [PRIAC's predecessor] Program. We believe that the schedule below is an accurate fee that recognizes the role of enhanced funds between active and passive strategies." (PX 46.)*fn3 The fee schedule provided that State Street would charge a fee of 10 basis points for the first $50 million of assets in the fund, a fee of 8 basis points on the next $50 million, and a fee of 6 basis points on any amount thereafter. (Id.; Stipulations of Fact ¶ 8.) By contrast, the average fee for a State Street active fund in 2007 was 52 basis points. (See PX 625.)
In an October 2002 presentation to CIGNA, State Street presented its fixed income funds as falling into three categories: "Active," "Enhanced," or "Passive." (PX 406.) The GCBF was listed in the "Enhanced" category. (See id.)
In 2003, PRIAC employees conducted a site-visit at the State Street offices. At a presentation during the visit, State Street categorized the Bond Funds as part of an "Active Core Bond Strategy." (PX 63.) State Street indicated that both Bond Funds had an excess return target of 30-40 basis points and a predicted tracking error of 40-50 basis points. (Id.) State Street's presentation also described the Bond Funds as employing a "risk-controlled process [that] ensures consistent, steady performance." (Id.)
Following the visit, Robert Frascona, a PRIAC employee, wrote a summary of the visit, which stated, "Although SSgA characterizes the [Bond Funds'] strategies as actively managed, they truly fall between passive management, where TE is expected to be 0-10 bps and active management, where TE is expected to be in excess of 70 bps." (PX 48.) A State Street employee similarly noted in a written summary of the visit that both Bond Funds "reside on Cigna's platform of products in an enhanced bond category. They are available for selection if plan sponsors choose the particular style offered by the strategies-low tracking error fixed income." (PX 635.) State Street's summary indicates that it understood the Bond Funds to employ a "low tracking error" strategy, which is consistent with the 1996 description of an enhanced strategy as one that seeks to "add value over the index while mirroring its risk profile," and with the 2003 description of the Bond Funds' "risk-controlled process." (PX 61; PX 63.)
Through the middle of 2005, PRIAC created quarterly Fact Sheets about each of the Bond Funds to send to the Plans. Before sending the Fact Sheets to the Plans, PRIAC sent them to State Street for State Street's review and approval. The Fact Sheets used language substantially similar to the language used in State Street's 1996 description of its enhanced index strategies (PX 61). For example, the Fact Sheet for the IBF from the fourth quarter of 2004 describes the Fund as "employ[ing] an enhanced bond indexing strategy which seeks to add consistent value over the benchmark through a disciplined, risk-controlled investment process that combines qualitative and quantitative portfolio management techniques." (PX 77.) State Street approved these Fact Sheets without commenting on the use of the term "enhanced bond indexing strategy" or on the description of that strategy; it did, however, provide comments unrelated to those issues. (See Flinn Dep. 48-49, 54-57; DX 94; PX 411; PX 549; PX 557.) PRIAC sent materially identical Fact Sheets to State Street again in mid-2007, and State Street again approved the Fact Sheets without commenting on the language describing the Funds' strategy as "enhanced," although there is no indication that PRIAC specifically asked State Street to do so. (See DX 42; DX 49.)
In some later communications with PRIAC, State Street referred to the Bond Funds as "active." Commentaries and characteristics reports sent to PRIAC in January and June 2007 referred to the "Active Core U.S. Government/Credit Fund" and the "Active Intermediate Bond Fund." (DX 215 (January 2007); DX 170 (June 2007).) At the same time, State Street sent other communications to PRIAC that continued to describe the Bond Funds as enhanced. (See PX 456 (January 2007 communication referring to the "Govt/Credit Enhanced Bond"); PX 555 (same in July 2007).)
The references to "active," however, do not alter State Street's prior and more specific descriptions of the Bond Funds as having "low tracking error," (PX 63), and using "a risk-controlled" strategy, (PX 498). Nor would the references to "active" necessarily have put PRIAC on notice that the Bond Funds had undergone any material change in strategy because PRIAC always understood that there was an "active element" to the Bond Funds, in the sense that the Bond Funds sought excess returns by taking positions that did not simply mirror the Funds' benchmark indices. (See PX 280.) Thus, PRIAC reasonably understood that State Street was offering the Bond Funds as investments that employed a "low tracking error," "risk-controlled" approach that sought to "modestly outperform the benchmark while taking on minimal additional risk." (PX 63; PX 10.)
2.State Street's Description of the Bond Funds that Did Not Directly Reach PRIAC
In a March 2004 internal presentation, State Street described CIGNA as a client who used State Street's "Enhanced Fixed" strategies. (See PX 644.) Gregory Mulready, a State Street client services representative, testified that he understood this presentation to indicate that CIGNA offered State Street's enhanced index funds-including the Bond Funds-to its clients. (See Mulready Dep. 51.)
Similarly, during the 2006 presentation to business school students described above, State Street made reference to each Bond Fund and its "Enhanced Index Benchmark." (See PX 498.)
D.The Bond Funds' Alpha Targets and Targeted Predicted Tracking Error over Time
When the Plans first invested in the Bond Funds in the mid-1990s, the Bond Funds sought excess returns of 20-30 basis points over their benchmark indices. (See PX 61, Tr. 302-03 (Frascona testimony).)
When PRIAC conducted its 2003 site-visit at the State Street offices, State Street informed PRIAC that both Bond Funds had an excess return target of 30-40 basis points and a predicted tracking error of 40-50 basis points. (See PX 63.)
In February 2005, Robert Frascona of PRIAC sent an email to Kallie Hapgood of State Street inquiring about the performance of the Bond Funds. (See DX 147.) Frascona indicated that from 2002 to 2004 the Bond Funds intentionally had tracked their benchmark indices so that, net of fees, the Bond Funds actually underperformed their benchmarks. (See id.) Frascona suggested that three years was a long time for an "ENHANCED INDEX FUND" simply to replicate the performance of its benchmark. (Id. (capitalization in original).) Frascona's email suggests that PRIAC expected the Bond Funds to generate at least some excess returns over their benchmarks.
Hapgood responded to Frascona's email with updated alpha target, predicted tracking error, and information ratio figures. Hapgood's response indicates that the Bond Funds' alpha target currently was 40-60 basis points, that their predicted tracking error was 50-75 basis points, and that their targeted information ratio was greater than 0.5. (See id.) These figures are consistent with State Street's description of the Bond Funds as "enhanced" index funds that utilized a "low tracking error," "risk-controlled" approach that PRIAC understood sought to "modestly outperform the benchmark while taking on minimal additional risk." (PX 63; PX 10.)
Hapgood's figures for the Bond Funds' alpha target (40-60 basis points) and predicted tracking error (50-75 basis points) yield an ex-ante information ratio of 0.8. This is a challenging information ratio, but it is consistent with an enhanced index fund's general goal of "achieving a high information ratio." (PX 10.) In any event, since the predicted tracking error of a fund is the variable the investment manager can control, (Tr. 112 (Blume)), Hapgood's figures are more indicative of an overly optimistic alpha target than they are of an understated predicted tracking error. Accordingly, Hapgood's figures would not have given PRIAC reason to believe that the Bond Funds' risk targets exceeded the 50-75 basis points stated in the email.*fn4
Hapgood's email was the last time that State Street reported predicted tracking error figures to PRIAC. (See Tr. 1198:5-11 (Carron testimony).)
By May 2006, State Street had increased the Bond Funds' alpha target to 70-80 basis points. (See PX 142; Tr. 813 (Reigel testimony).)*fn5
State Street offered its clients, including PRIAC, access to certain account information through State Street's Client's Corner website. State Street made PRIAC aware of the existence of the website through emails. Every month, State Street emailed Robert Frascona of PRIAC a copy of monthly performance reports for the Bond Funds, and in the same email, State Street wrote,
You can access your reports on SSgA's Client's Corner, our password protected website. Client's Corner provides secure access to your organization's investments with SSgA, including: Performance Information and Quarterly Investment Commentary; Appraisals and Trade Summaries of commingled funds;
SSgA Research and Commentary; Direct link to your SSgA Client Service Team to submit questions and requests. (DX 392.) In 2007, the Client's Corner website for PRIAC listed the Bond Funds as "active" and indicated they each had an alpha target of 50-75 basis points. The website did not indicate a predicted tracking error for the Bond Funds.
In 2007, only two individuals from PRIAC accessed the website. Michelle Cappalla, a back office employee, accessed the website in March, and Matthew Dingee, a PRIAC analyst and a direct report to Robert Frascona, accessed the website in August and September. (See DX 89; DX 830.)
State Street's Head of Risk Management, Patrick Armstrong, testified that in 2007 the Bond Funds were managed according to an ex ante information of 0.5, and that, in his opinion, an information ratio greater than 0.5 was "overly ambitious, overly heroic, because I don't see it empirically."*fn6 (Tr. 916:19-21, 917:24-918:9.) An information ratio of 0.5, along with an alpha target of 50-75 basis points, means that the Bond Funds in 2007 were being managed to accept a predicted tracking error of 100-150 basis points. (Tr. 916:22-917:3 (Armstrong testimony).) This figure is double the predicted tracking error range disclosed to PRIAC in 2005 and is materially greater than the previously disclosed range. Nonetheless, State Street never disclosed the figure to PRIAC. Although State Street did post the Bond Funds' alpha targets on a rarely-used web portal, that disclosure was insufficient to notify PRIAC of the change in risk target that occurred between 2005 and 2007. To conclude from the Client's Corner website that the Bond Funds had increased their predicted tracking error to 100-150 basis points, PRIAC would be required to assume that the information ratio of 0.8 implicit in Hapgood's 2005 email was unrealistic and that the website's alpha target reflected State Street's willingness to accept a predicted risk-range that was double what it previously had disclosed to PRIAC. Based on the evidence, that is not an assumption that PRIAC reasonably should have made. At most, the Client's Corner website could have put PRIAC on alert that the Bond Funds had increased their predicted tracking error range to approximately 62-93 basis points, which is the predicted tracking error gleaned from the 50-75 basis point alpha target and the 0.8 information ratio implicit in Hapgood's email. In either event, the use of the rarely-viewed Client's Corner website was not an adequate way of making PRIAC aware of the significant increase in predicted tracking error.
State Street's failure to adequately disclose the increase in predicted tracking error of the Bond Funds to 100-150 points is significant because the evidence shows that PRIAC relied on the predicted tracking error figures in monitoring the Bond Funds. Following PRIAC's 2003 site visit at State Street, Robert Frascona of PRIAC made the following comment in his written "Site Visit Debrief,"
Comments made during the meeting implied that SSgA is continuing to develop and test various strategies. This gave us the sense of a learn while doing. Given the risk constraints (TE band) we did not sense that the portfolio's [sic] are at risk, but as with all managers, SSgA's approach is continually evolving. (PX 48.) This comment suggests that PRIAC was aware that State Street might increase the risk in the Bond Funds' portfolios by some amount, but more importantly, it suggests that PRIAC relied on the Bond Funds' stated predicted tracking error range in assessing the continued viability of the Bond Funds as investments on its enhanced index platform.
Based on the foregoing, PRIAC was justified in understanding that the Bond Funds were "enhanced index funds" that employed a "low tracking error," "risk-controlled" strategy that sought modest returns over their benchmarks without taking on significant additional risk. With respect to risk metrics, PRIAC was justified in understanding that the Bond Funds would be managed to accept a predicted tracking error range of roughly 50-75 basis points.
E.The Portfolio Managers' and the Fixed Income Executives' Understanding of the Bond Funds
The portfolio managers and the fixed income executives at State Street (most of whom did not interact with clients with any regularity) testified that they would not have described the Bond Funds' strategy as "enhanced," and were unaware that others within State Street had presented the Bond Funds to PRIAC as such. (See Tr. 810-12 (Reigel testimony); Pickett Dep. 824-26 (portfolio manager for the IBF agreeing that both Bond Funds employed predicted tracking errors and alpha targets that were higher than those of an enhanced index fund, and that he would not describe the Bond Funds as enhanced index funds); Wands Dep. 104-05 (State Street's Head of North American Fixed Income noting that the "enhanced" description "would have been inconsistent, in 2007, with what the portfolio group thought of as their products").)
F.The Context of the Increase in Alpha Target and Predicted Tracking Error
In 2006, State Street adopted a business strategy to "leverage fixed income," (PX 139), and one of its goals was to "Establish SSgA as a Global Leader in Active Fixed Income." (PX 184; see Tr. 816 (Reigel testimony).) In August of that year, Paul Greff, State Street's Head of Global Fixed Income, prepared an outline of bullet points that he would later use in a video presentation made available to State Street employees addressing State Street's strategy of gaining recognition as an active fixed income manager. Under the heading "How will we do it?", Greff indicates that State Street should "[t]ake more active risk," and "[g]enerate higher returns for clients in existing products." (PX 141.)
Susan Reigel, the portfolio manager for the GCBF (and the back-up portfolio manager for the IBF), testified that State Street's business strategy had no impact on her management of the Bond Funds. (Tr. 807-08.) Sean Flannery, the former Chief Investment Officer of SSgA, also testified that the "leverage fixed income" initiative had no impact on the Bond Funds. (Tr. 594:17-596:6.) However, Reigel also testified that in 2006 Michael O'Hara, State Street's Head of Active Global Fixed Income, told her that the alpha targets of the Bond Funds should be increased from 50 basis points to 70-80 basis points. (Id. at 816.) O'Hara was one of the senior State Street managers that orchestrated State Street's initiative to gain recognition as a manager of active fixed income. (See PX 141.) This is the only direct evidence that PRIAC has presented of a connection between the "leverage fixed income" initiative and the decision to increase the alpha targets for the Bond Funds. But it sheds little light on the rationale behind the decision. A direction from O'Hara to increase the alpha target of the Bond Funds is insufficient to establish that the decision was the result of State Street's business objectives.
G.The Bond Funds' Alpha-Seeking Investments
The Bond Funds sought to achieve their excess return targets by investing in off-index securities. The Bond Funds' off-index investments were comprised largely of an investment in another State Street fund called the Limited Duration Bond Fund ("LDBF"), and in derivative securities backed by subprime-based assets.
1.The Limited Duration Bond Fund
The LDBF was created to provide a source of "portable alpha" for other State Street fixed income funds. (See PX 176.) As of June 30, 2007, over 85% of the $1.3 billion invested in the LDBF was invested by other State Street funds. (See PX 640.)
The LDBF did not perform well compared to its peer funds. Its five-year annualized returns through March 2007 put it in the bottom sixth percentile of 128 peer funds. (Tr. 850-51.) While the LDBF did generate positive returns, the returns were only slightly above the London Interbank Offered Rate (LIBOR), the "risk-free interest rate" that was the LDBF's benchmark. (Tr. 1210-11 (Carron); PX 487.)
The LDBF sought to achieve excess returns by investing in increasing amounts of home-equity Asset Backed Securities ("ABS"), otherwise known as subprime securities.*fn7 By January 2007, approximately 93% of the LDBF's notional holdings ($2.035 billion out of $2.190 billion) were invested in various forms of subprime securities. (See PX 601, at SSE000000816 (Excel Spreadsheet, tab labeled "jan07," table labeled "Sector Distribution (Assets)").)
State Street limited the amount of money any one State Street fund could invest in the LDBF. Before 2006, that limit was 20% of the market value of the fund that chose to invest in the LDBF. (See PX 436.) The limit later was increased to 25%. (See Tr. 854:5-9 (Reigel).) The limit was calculated by looking both at the fund's direct investments in the LDBF and also at the fund's exposure to the LDBF obtained by investing in other State Street funds that themselves invested in the LDBF. (Tr. 854:13-22 (Reigel).) In other words, State Street employed "a look-through analysis" in calculating the 25% limit. (Id.) The Bond Funds, however, also could invest directly in the same securities as the LDBF without contributing to the 25% limit, and the Bond Funds did so. (Compare PX 824 (showing the IBF's total exposure to subprime ABS, including the exposure obtained through LDBF), with PX 825 (showing the IBF's exposure to subprime excluding the exposure obtained through LDBF).)
The Bond Funds invested a substantial portion of their assets in the LDBF. As of June 29, 2007, the IBF had invested 23.45% of its assets (based on market value) in the LDBF. (Tr. 855.) At the time, the market value of the IBF's assets was $1.279 billion. Similarly, the GCBF, as of June 29, 2007, had exposure to the LDBF equal to 22.74% of the GCBF's net asset value. (DX 351, Ex. 4.)
The Bond Funds' notional exposure to the LDBF was much greater than 25% of the market value of its assets. As of June 29, 2007, the IBF's notional exposure to the LDBF totaled $1.5 billion, or approximately 117% of the total market value of the IBF. (See Tr. 856 (Reigel).) Virtually all of this exposure was to securities backed by subprime mortgages. (Compare PX 824, with PX 825.)
The Bond Funds also sought to achieve excess-returns by investing in derivative securities. State Street defined derivative securities as "securities that derived their value from another or other securities." (PX 284.)
The Bond Funds took substantial positions in two derivative instruments that derived their value from the performance of underlying home-equity ABS (or subprime securities): (1) Total Return Swaps ("TRS") on the Lehman Brothers Floating-Rate Asset Backed Securities Home Equity AAA and AA indices, and (2) the ABX Index.*fn8
The TRS on the Lehman Brothers Floating-Rate Asset Backed Securities Home Equity AAA and AA indices worked as follows. State Street and a counterparty (as relevant here, usually a subsidiary of Lehman Brothers (see Tr. 80)) would agree on a "notional" amount for the swap, for example, $1 million. The counterparty then would agree to pay State Street the return on the relevant Lehman Brothers Index over a specified period of time based on the notional amount. State Street, in turn, agreed to pay the counterparty some established interest rate, usually an amount above LIBOR, likewise based on the notional amount of the swap. Thus, if the Lehman Brothers Index generated a higher return than the rate State Street agreed to pay the counterparty, State Street would make money. By contrast, if the Lehman Brothers Index underperformed the counterparty's rate, State Street would lose money. (See Tr. 77-78 (Blume testimony).) The TRS allowed State Street to gain exposure to the underlying indices while putting little or no cash up front. (Id.)
The TRS were not standardized financial instruments and were traded over-the-counter in a thin market. (Kinney Dep. 7/7/2009 at 173.) The underlying assets in the Lehman Brothers indices were priced only weekly, which suggests that the market for the underlying assets was not strong and that the underlying assets were rather illiquid. (Tr. 78-79 (Blume).) In addition, since Lehman Brothers was the only counterparty from whom State Street could receive a bid or cancel the swap, (Kinney Dep. 7/7/2009 at 173-74), the swaps themselves were not very liquid. (Tr. 78:11-13 (Blume testimony).) State Street's head trader Andrew Tenczar, however, testified that he consistently monitored the liquidity of the TRS by conversing with dealers and other market participants about prices and spreads, and that he never had difficulty with liquidity until July and August of 2007. (Tenczar Dep. 51, 59, 149.)
The Lehman Brothers indices on which the TRS were based were created in May 2005, with data backfilled to 2005. (Tr. 78-79 (Blume).) The limited history of performance data meant that attempts to model the risk characteristics of the TRS through different market cycles would likely prove unreliable. (Id.) Indeed, in July 2007, one State Street employee remarked to another, "Well the problem is you and I both know that there is no history, no experience upon which to draw here. You can't watch this through other cycles, because the product did not exist." (Kelly Dep. 82-83.) This statement reflects the concept of "model risk" which occurs "when a model is calibrated incorrectly relative to the environment it is trying to capture." (Tr. 1011:23-25 (Armstrong testimony).) Model risk increases the risk of large losses relative to a fund's benchmark. (See Tr. 90-91 (Blume testimony).)
In late August 2007, in preparation for a presentation to State Street's board of directors, State Street employees prepared a list of talking points about SSgA's investment strategies. (See PX 585.) One of the bullet points in that document states, "How does SSgA identify the risks associated with new and relatively untested instruments?" (Id.) Under a sub-bullet point entitled "To manage," the author indicates that SSgA employs the concept of "naive allocation (cap allocation to manageable size -- from perspective of notional at risk)." (Id.)
Despite the stated approach to managing the risks of untested instruments, each Bond Fund held, by the end of March 2007, over $1 billion in notional exposure to the AAA and AA TRS. (See Pickett Dep. 8/12/2009 at 1019:15-19; PX 319.) One billion dollars in notional exposure to an untested instrument (when the net asset value of the entire fund is less than $1.3 billion) is inconsistent with State Street's stated methodology of "cap[ping] allocation to manageable size -- from perspective of notional at risk." (PX 585.)
In addition, the AAA and AA ratings in the names of the TRS do not refer to the credit quality of the swap, but to the credit quality of the assets underlying the indices. (Tr. 79-80 (Blume).) A subprime security with a AAA rating is intended by the rating agency to reflect the same level of credit risk as a AAA rating on a corporate bond. (Tr. 668 (Flannery testimony).) As early as September of 2006, however, State Street considered any subprime security rated below AAA to be the equivalent of a "distressed" security. (See Tenczar Dep. 146-48 (State Street's head trader describes an email exchange with Michael Wands, State Street's Head of North American Fixed Income).)
The ABX Index actually was a series of six subindices, each of which was offered in four vintages, two in 2006 and two in 2007. The subindices ranged in credit quality from AAA to BBB-. The value of each subindex was determined with reference to a basket of twenty other derivatives known as credit default swaps. The credit default swaps, in turn, derived their value from the value of an underlying individual home equity Asset Backed Security. (See Tr. 81.) The ratings of the subindices were derived from the ratings of the underlying individual home equity Asset Backed Securities. (Statchel Dep. 127-28.)
There were two ways in which an investor could participate in a particular ABX Index. One way was to "buy protection;" the other was to "sell protection." (Tr. 81:23-24.) The buyer of protection would be required to pay the seller of protection a certain periodic premium (based on the notional value of the trade), but also would be entitled to receive a specified payment (also based on the notional value of the trade) from the counterparty if certain events negatively affected the value of the underlying reference assets. The seller of protection would be entitled to receive the buyer's periodic premium payments, but also would be obligated to pay the buyer if the same certain events negatively affected the value of the underlying reference assets. (See Tr. 81.)
State Street sold protection in the 2006-2 vintage of the ABX BBB Index, and also in the ABX AA Index. (See Tr. 81-82; PX 318; PX 319.) Thus, State Street was "long" the ABX Index, in the sense that it would profit if the Index performed well. (Tr. 82:14.)
Similar to the TRS on the Lehman Brothers indices, the ABX Index lacked historical data on which to model its risk characteristics. State Street originally modeled the risk of the instrument using the underlying individual home equity ABS as a "proxy." (Pham Dep. 57-58.) By February 2007, however, State Street realized that this approach understated the instrument's risk. At that time, Patrick Armstrong, State Street's Head of Risk Management, wrote,
The problem with the modelling [sic] of the ABX BBB trade was that given the short history of the synthetic, we did proxy its behaviour [sic] to the attendant cash instrument in order to obtain a robust time series. However, what has become apparent is that the synthetic is ...