KING, Chief Judge:

Plaintiffs-Appellants Robert A. Bussian and James J. Keating appeal from the district court's grant of summary judgment to Defendant-Appellee RJR Nabisco, Inc. and its denial of class certification. We reverse in part, vacate in part, and remand for further consideration by the district court.

I. FACTUAL AND PROCEDURAL BACKGROUND This case is yet another litigating who must bear the cost of the collapse of Executive Life Insurance Company of California ("Executive Life") in the late 1980s and early 1990s. The issue before us is whether Defendant-Appellee RJR Nabisco, Inc. ("RJR") acted consistently with its fiduciary obligations under § 1104 of the Employee Retirement Income Security Act of 1974, 29 U.S.C. § 1001 et seq. (1994) ("ERISA"), when it chose to purchase a single-premium annuity from Executive Life in August, 1987.

Because this case comes to us from a grant of RJR's motion for summary judgment, our presentation of the facts reflects in part the requirement that we view the evidence in the light most favorable to Plaintiffs-Appellants Robert A. Bussian and James J. Keating ("Appellants"). Many of the underlying facts are uncontested. RJR's involvement in this case comes about through its purchase, in 1976, of Aminoil USA, Inc. ("Aminoil"), a Houston-based oil company. Aminoil administered a pension plan for its employees that was governed by ERISA. RJR sold Aminoil in 1984, and the purchaser assumed the pension obligations for all then-current employees. At the time of the sale, other employees had ceased employment with the oil company and were either already receiving pension benefits or were vested in the Aminoil pension plan but were not yet eligible to receive benefits. RJR retained the obligation of administering pension benefits for these former employees, including Appellants, under an ERISA-defined benefit pension plan ("the Plan").

On October 16, 1986, RJR's Board of Directors approved resolutions authorizing the termination of the Plan and several other plans of former RJR subsidiaries. The Board also approved the purchase of an annuity to cover all pension obligations to the participants and beneficiaries of all the plans. The Plan's documents provided that upon termination any excess funds would revert to RJR. At the time the decision to terminate was made, the Plan was over-funded, and the Board was informed that a reversion could be expected. By December 1986, RJR was assuming that an annuity would cost about $62.5 million, and allowing for a $10 million cushion, was anticipating a reversion of about $55 million.

Members of RJR's Pension Asset Management Department were given the responsibility of selecting an annuity provider. Paul Tyner was involved from the beginning; Robert Shultz, hired in March, 1987 as RJR's Vice President of Pension Asset Management, had responsibility for making the final decision. In October, 1986, RJR hired Buck Consultants, Inc. ("Buck") to assist in the endeavor. William Overgard, an investment consultant with Buck Pension Funds Services, was asked to participate in the process in January, 1987.

Overgard was told that his role in the transaction was to identify insurance companies and to provide those companies with appropriate information in order to solicit the best bid from each one that was interested in the business so that RJR could select the carrier that was appropriate to its needs. Overgard compiled an initial list of insurance companies that could provide the annuity. That list included providers with which Buck was familiar, that had a reputation for providing good service to their clients, and that would have the capacity for a placement covering approximately 10,000 individuals. In January, 1987, a letter was sent to thirteen companies inviting comments on several issues related to the purchase of the annuity. In the letter, RJR was not identified as the buyer of that annuity.

Executive Life was not among those receiving the January letter. This was because it was involved in a nontraditional investment strategy: its portfolio had a higher percentage of low-quality bonds and a lower percentage of other investments than other insurance companies. Low-quality bonds, which are also referred to as "high-yield" or "junk" bonds, are rated below investment grade, i.e., ratings agencies have determined that the issuing entity is a greater than average credit risk. In order to compensate for the increased risk of default, such bonds must offer a higher interest rate. See, e.g., Levan v. Capital Cities/ABC, Inc., 190 F.3d 1230, 1235 (11th Cir. 1999). After Overgard discussed Executive Life's strategy with one of his colleagues, the two decided that the company should not be included on the initial list.

Overgard understood that by 1987, over 50% of Executive Life's portfolio was in low-quality bonds. In this Executive Life was indeed unusual compared to its competitors in the insurance industry. Information in the record suggests that the average percentage of low-quality bond holdings was on the order of 6% to 7%. Executive Life allegedly held the largest original issue low-quality bond portfolio ever assembled, with most of its acquisitions coming through Drexel Burnham Lambert ("Drexel"). Overgard understood Executive Life's low-quality bond holdings to be broadly diversified.

Based on his experience with Executive Life in the course of bidding he conducted for guaranteed investment contracts, and his desire to increase the competitiveness of the final bidding for the annuity contract, on or about April 3, 1987, Tyner requested that Executive Life be added to the list of carriers. In Tyner's opinion, Executive Life's inclusion would facilitate bringing other bidders down in price because it would come in with a lower quote. According to William J. Wolliver, a former Manager of Annuity Pricing for Prudential Insurance Company, Executive Life's low-quality bond portfolio enabled the company to underbid his firm. At the time he requested that Executive Life be added, Tyner did not think that the provider would be seriously considered in the final bidding process. Instead, he believed that RJR would go with a more well-known company.

To check up on Executive Life's solvency and financial health, Overgard reviewed the reports and ratings of four rating agencies (Standard Poor's Corp. ("SP"), Moody's Investor's Services ("Moody's"), A.M. Best ("Best"), Conning Company ("Conning")). He reviewed the pros and cons of including Executive Life on the list of carriers to be contacted with Henry Anderson, an actuarial expert with Buck who, as the account executive, had brought Overgard in on the RJR purchase. They discussed the high-quality ratings that Executive Life had received, the company's interest in doing business, its reputation for providing good service and for being knowledgeable in the business, and its nontraditional investment portfolio. Overgard believed that a broadly diversified portfolio of low-quality bonds was a viable investment strategy. Based on his investigation, Overgard determined that Executive Life should be included on the bid list because the ratings the company received from SP, Best, and Conning were high; its low-quality bond portfolio was broadly diversified and its investment strategy sound; and its administrative capabilities and reputation in the annuity business were strong.

On April 8, 1987, Buck solicited bids from fourteen potential annuity providers, including Executive Life. Buck had previously explained to RJR that companies would make initial bids and that Buck would select possibly three companies from which final bids would be solicited. In May, five potential providers submitted preliminary bids: AIG Life Insurance Company ("AIG"), Aetna Life Insurance ("Aetna"), Executive Life, Mutual Life Insurance Company of New York, and Prudential Asset Management Company ("Prudential"). The other companies declined to participate, primarily because of the complexity associated with the numerous plans.

Between May and August, Overgard provided additional information to the companies interested in bidding. The bulk of his time was spent working with the companies to make sure they had correct data and enough data to enable them to submit a qualified bid, testing whether alternative strategies might be available for placing the bid on the final bid day, and assessing how hard he could push the companies in final negotiations.

Sometime prior to August, 1987, Overgard learned that Moody's had given Executive Life a rating of A3, which was two grades below that of SP's AAA rating for the company. He also read media reports speculating that problems in the market for low-quality bonds might affect Executive Life. Overgard determined from a discussion with an individual at Moody's that the rating agency had not talked with Executive Life management prior to issuing its rating, and he pursued "industry intelligence" regarding the company. Overgard concluded that the lower Moody's rating was an attempt on the part of the agency to gain publicity, but did not recall a specific discussion with the individual at Moody's regarding why the agency rated Executive Life as it did, or how the agency viewed the provider's nontraditional portfolio. He found that the opinions of other insurance companies were mixed: "some were not concerned about Executive Life and some were willing to put the fear of God into us," the latter describing low-quality bonds as a bad investment strategy. Concerned about what would happen to the market for low-quality bonds should Drexel collapse, Overgard talked to investment bankers. In Overgard's opinion, those bankers were quite eager to move into the market for low-quality bonds. Overgard also viewed Executive Life as working the case harder and asking more questions during the bid solicitation process than the other companies. Overgard concluded that Executive Life should remain on the bid list.

Final bid day was set for August 12, 1987. On that day, Overgard met with representatives of RJR (Tyner, and representatives from RJR's Employee Benefits and Legal Departments) to review the preliminary bids. The sole documentation RJR had comparing providers was a listing of the final companies' ratings and their initial bids. Buck did not recommend any particular company; instead, it saw each of the four remaining companies as qualified and competent to provide the annuity. As a result, Overgard saw his role on final bid day as obtaining from each company its best (lowest) bid. Overgard negotiated with the four companies in one room; RJR representatives were in another room. Overgard determined midday that Aetna and AIG had given their best bid, and so concentrated for the remaining period on obtaining lower bids from Prudential and Executive Life. The following provides the final bids along with other information Buck supplied RJR:

INSURER SP BEST MOODY'S CONNING BID

Aetna AAA A+ AAA 102/104 $61.9 M

AIG AAA A AAA N/A $60.2 M Executive Life AAA A+ A3 100/106 $54 M

Prudential AAA A+ AAA 98/91 $56.7 M

Aetna's bid was the highest at $61.9 million, and Executive Life's was the lowest at $54 million. According to Overgard, the numeric Conning ratings reflected historical information on liquidity over two years. Thus, Aetna's rating of 102/104 reflected an improvement, while Prudential's ratings reflected a decline.

RJR had established three requirements that "at a minimum" the company providing the annuity would have to meet: (1) receipt of an AAA rating from S P; (2) capacity to administer the plans; and (3) approval from Buck. On the final bid day, Shultz had a number of other things to do. Because he had full confidence in the RJR representatives present, and "because the dollar value of the assets involved in the transaction was insubstantial in comparison to RJR's total pension portfolio," he attended the meeting for about an hour and fifteen minutes at its outset. After the final bids came in, RJR representatives present identified Executive Life as the insurer from which to purchase the annuity, as it was the lowest bidder, had at least one AAA rating, and was capable of administering the annuity. Tyner telephoned Shultz to inform him of the recommendation. After a fifteen or twenty-minute conversation, Shultz gave the go-ahead to select Executive Life.

Unlike Tyner, Shultz was aware of a number of facts regarding Executive Life, its chairman, Fred Carr, and the market for low-quality bonds. For example, Shultz was aware (1) of the percentage of Executive Life's portfolio that was devoted to low-quality bonds, (2) of allegations regarding a connection between Executive Life and Drexel's Michael Milken, (3) that Executive Life was one of Milken's largest customers, (4) that Drexel and Milken were the targets of SEC and Attorney General investigations of the 1986 insider trading scandal, (5) that Executive Life and Carr came within the scope of those investigations, and (6) that Executive Life of New York, a subsidiary of Executive Life, had been fined by New York insurance regulators due to the insurer's reinsurance practices, had $150 million of reinsurance disallowed, and had received from Executive Life $150 million to make up the difference. Shultz had not seen as much negative press regarding Aetna's or Prudential's holdings of low-quality bonds as he had seen with regard to the holdings of Executive Life, and he had not seen the diversity of reviews of the other companies that he had seen with respect to Executive Life. Shultz stated that he relied primarily on Tyner's input, and that his decision to concur in the purchase of Executive Life's annuity was made taking into account the fact that "Executive Life had the same SP rating as did Prudential, had a reputation equal to or better than Prudential's for being able to service complex annuity contracts and was recommended by Buck."

On August 17, 1987, RJR caused $54 million to be wired to Executive Life. A letter agreement was signed November 23 of the same year. RJR formally terminated the Plan on June 30, 1988. The total pre-tax reversion associated with the termination of all plans covered under the annuity was $82,080,000; this resulted in RJR receiving on May 27, 1989 a net reversion of $43,051,510.

Tyner was aware that by 1989, Executive Life was suffering financially. To his knowledge, however, no one at RJR considered extracting himself from the deal to buy Executive Life's annuity. RJR accepted the provider's annuity contract on December 13, 1989.

By late 1989, the low-quality bond market was suffering significant losses. Because well over half of Executive Life's portfolio consisted of low-quality bonds, it felt the brunt of those losses. In January, 1990, First Executive Corporation, the parent of Executive Life, announced that its low-quality bond portfolio had lost $1 billion in market value and that it would take a $515 million write down. In April, 1991, California insurance regulators placed Executive Life in conservator ship, and for a period of time, certain Executive Life policyholders received reduced benefits. Eventually, the market for low-quality bonds rebounded, and Executive Life was taken over by a consortium of French companies, which formed Aurora National Life Assurance Company. Unfortunately, Appellants and some other Plan participants have not received their full benefits.

Appellants filed suit, on their own behalf and on behalf of a class, against RJR in Texas state court in 1991, alleging violations of RJR's fiduciary duties. RJR removed the case to federal court and moved for summary judgment in 1992. In 1998, the district court granted summary judgment and, consequently, denied Appellants' motion to certify a class. See Bussian v. RJR Nabisco, Inc., 21 F. Supp.2d 680 (S.D.Tex. 1998). Appellants timely appeal.

II. SUMMARY JUDGMENT A. Standard of Review

We review the granting of summary judgment de novo, applying the same criteria used by the district court in the first instance. See Norman v. Apache Corp., 19 F.3d 1017, 1021 (5th Cir. 1994); Conkling v. Turner, 18 F.3d 1285, 1295 (5th Cir. 1994). Summary judgment is proper "if the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there is no genuine issue as to any material fact and that the moving party is entitled to a judgment as a matter of law." Fed.R.Civ.P. 56(c); see Celotex Corp. v. Catrett, 477 U.S. 317, 327 (1986). "[T]here is no issue for trial unless there is sufficient evidence favoring the non moving party for a jury to return a verdict for that party. If the evidence is merely colorable, or is not significantly probative, summary judgment may be granted." Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 249 (1986) (citations omitted). We must view all evidence in the light most favorable to the party opposing the motion and draw all reasonable inferences in that party's favor. See id. at 255.

B. The Standard

Section 1104(a) sets forth the general duties imposed upon ERISA fiduciaries:

(1) Subject to sections 1103(c) and (d), 1342, and 1344 of this title, a fiduciary shall discharge his duties with respect to a plan solely in the interest of the participants and beneficiaries and —

(A) for the exclusive purpose of:

(i) providing benefits to participants and their beneficiaries; and

(ii) defraying reasonable expenses of administering the plan;

(B) with the care, skill, prudence, and diligence under the circumstances then prevailing that a prudent man acting in a like capacity and familiar with such matters would use in the conduct of an enterprise of a like character and with like aims;

(C) by diversifying the investments of the plan so as to minimize the risk of large losses, unless under the circumstances it is clearly prudent not to do so; and

(D) in accordance with the documents and instruments governing the plan insofar as such documents and instruments are consistent with the provisions of this subchapter and subchapter III of this chapter.

29 U.S.C. § 1104(a)(1). We have recognized that this provision imposes several overlapping duties. See, e.g., Metzler v. Graham, 112 F.3d 207, 209 (5th Cir. 1997) (involving the duty to diversify and the duty of loyalty); Donovan v. Cunningham, 716 F.2d 1455, 1464 (5th Cir. 1983) ("Section [1104] imposes upon fiduciaries a duty of loyalty and a duty of care."). Appellants assert that the district court erred in holding that, as a matter of law, RJR satisfied its obligations under ERISA. They argue that RJR was required to attempt to select the safest available annuity to satisfy its duty of loyalty. They also contend that RJR failed to conduct an investigation that satisfied its duty of care, and that it acted inconsistently with its duty to diversify in selecting an insurance carrier that held 50% to 60% of its portfolio in low-quality bonds.

1. The Duty to Diversify

We first narrow the focus of our inquiry by disposing of one of Appellants' arguments. They assert that § 1104(a)(1)(C) imposes on a fiduciary selecting an annuity the duty to select an insurance provider whose portfolio is sufficiently diversified. We disagree. Section 1104(a)(1)(C) deals specifically with "investments of the plan." As RJR points out, the purchase of an annuity to facilitate plan termination is not an investment of the plan. It is, as 29 U.S.C. § 1341(b)(3) provides, a "final distribution of assets." Section 1104(a)(1)(C) therefore does not impose upon a plan fiduciary the obligation to investigate or ensure the adequate diversification of an annuity provider's portfolio. This is not to say that a plan fiduciary has no obligation to consider the diversification of an annuity provider's portfolio; such an obligation may exist under § 1104(a)(1)(B), a possibility we address infra. Cf. 29 U.S.C. § 1104(a)(2) (stating that the "diversification requirement of paragraph (1)(C) and the prudence requirement (only to the extent that it requires diversification) of paragraph (1)(B)" do not apply to certain transactions). We are therefore left to determine the proper standard to guide our inquiry into whether summary judgment is appropriate to dispose of Appellants' claims that RJR breached its duties of loyalty and care in purchasing Executive Life's annuity.

2. The Duty of Loyalty

ERISA's duty of loyalty is "the highest known to the law." Donovan v. Bierwirth, 680 F.2d 263, 272 n. 8 (2d Cir.), cert. denied, 459 U.S. 1069 (1982); cf. Meinhard v. Salmon, 164 N.E. 545, 546 (1928) (Cardozo, J.) ("Many forms of conduct permissible in a workaday world for those acting at arm's length, are forbidden to those bound by fiduciary ties. A trustee is held to something stricter than the morals of the market place. Not honesty alone, but the punctilio of an honor the most sensitive, is then the standard of behavior."). The Supreme Court recently had occasion to describe ERISA's duty of loyalty, in so doing again recognizing the duty's source in the common law of trusts. See Pegram v. Herdrich, — S.Ct. —, 2000 WL 743301, at *7 (U.S. June 12, 2000) ("`The most fundamental duty owed by the trustee to the beneficiaries of the trust is the duty of loyalty. . . . It is the duty of a trustee to administer the trust solely in the interest of the beneficiaries.'" (quoting 2A A. Scott W. Fratcher, Trusts § 170, at 311 (4th ed. 1987))).

Although ERISA's duties gain definition from the law of trusts, the usefulness of trust law to decide cases brought under ERISA is constrained by the statute's provisions. See Varity Corp. v. Howe, 516 U.S. 489, 497 (1996) ("We also recognize . . . that trust law does not tell the entire story."); Cunningham, 716 F.2d at 1464. Under ERISA, for example, a fiduciary may have financial interests adverse to beneficiaries, but under trust law a "trustee `is not permitted to place himself in a position where it would be for his own benefit to violate his duty to the beneficiaries.'" See Pegram, 2000 WL 743301, at *8 (quoting 2A Scott Fratcher, § 170, at 311)). Despite the ability of an ERISA fiduciary to wear two hats, "ERISA does require . . . that the fiduciary with two hats wear only one at a time, and wear the fiduciary hat when making fiduciary decisions." Id. (citing Hughes Aircraft Co. v. Jacobson, 525 U.S. 432, 443-44 (1999)); see also Varity, 516 U.S. at 497.

That ERISA contemplates that a plan fiduciary may have multiple roles is reflected in the language of § 1104(a). That section begins with the phrase "[s]ubject to sections 1103(c) and (d), 1342, and 1344 of this title," which explicitly refers to ERISA provisions that allow plan assets to be returned to the employer under some circumstances. See Borst v. Chevron Corp., 36 F.3d 1308, 1320 (5th Cir. 1994), cert. denied, 514 U.S. 1066 (1995); District 65, U.A.W. v. Harper Row, Publishers, Inc., 576 F. Supp. 1468, 1477-78 (S.D.N.Y. 1983); Daniel Fischel J.H. Langbein, ERISA's Fundamental Contradiction: The Exclusive Benefit Rule, 55 U. Chi. L. Rev. 1105, 1154 (1988). As a result, although the balance of § 1104(a)(1) would appear to make a return of assets to an employer a violation of the duty to act "solely in the interest of participants and beneficiaries and for the exclusive purpose of providing benefits to participants," § 1104(a)(1)(A)(i), the provision's initial phrase precludes such an interpretation.

Under ERISA, neither the decision to terminate an overfunded plan, nor a reversion of plan assets that is consistent with § 1344(d), is a per se violation of § 1104(a)(1). See § 1108(a)(9) (exempting from prohibited transactions "[t]he making by a fiduciary of a distribution of the assets of the plan in accordance with the terms of the plan if such assets are distributed in the same manner as provided under § [1344] . . . ."); Lockheed Corp. v. Spink, 517 U.S. 882, 890-91 (1996) (extending to pension benefit plans the notion that when employers terminate employee welfare plans, they do not act as fiduciaries and instead are analogous to settlors of a trust); Izzarelli v. Rexene Prods. Co., 24 F.3d 1506, 1524 (5th Cir. 1994). Prior to termination, a defined benefit plan, such as the one involved in the case before us, "consists of a general pool of assets," Hughes Aircraft, 525 U.S. at 439, and "no plan member has a claim to any particular asset that composes a part of the plan's general asset pool." Id. at 440. Instead, plan members have a right only to their accrued benefit — a plan's surplus need not be made available for distribution to plan members. See id. at 440-41; Borst, 36 F.3d at 1315. Because an employer may, consistent with ERISA's provisions, receive a plan's surplus upon termination, the fact that the employer terminates a plan specifically to gain access to that surplus is not a violation. See District 65, 576 F. Supp. at 1478 (dismissing plaintiffs' breach of fiduciary-duty claim challenging a sponsor's termination of a plan in order to use the surplus to prevent a third party from taking control of the company).

However, simply because ERISA allows an employer to recoup surplus assets does not mean that a fiduciary's acts undertaken to implement a plan's termination may deviate from ERISA's command that a "fiduciary shall discharge his duties with respect to a plan solely in the interest of the participants and beneficiaries." § 1104(a)(1). The question whether an employer has access to a reversion because of a plan's termination is separate from the issue of the size of that reversion. See District 65, 576 F. Supp. at 1478. Undertaking steps to maximize the size of the reversion with the direct result of reducing benefits would be a violation of ERISA's commands. See Cooke v. Lynn Sand Stone Co., 673 F. Supp. 14, 27 (D.Mass. 1986) (denying summary judgment where a material fact question existed regarding whether sponsor had used higher interest rate to maximize its reversion); cf. Reich v. Compton, 57 F.3d 270, 291 (3d Cir. 1995) ("[T]rustees violate their duty of loyalty when they act in the interests of the plan sponsor rather than `with an eye single to the interests of the participants and beneficiaries of the plan'" (quoting Donovan v. Bierwirth, 680 F.2d 263, 271 (2d Cir.), cert. denied, 459 U.S. 1069 (1982))).

The Secretary of the Department of Labor (the "Secretary"), as amicus curiae, urges us to hold that the duty of loyalty requires that a fiduciary disposing of plan assets as part of a termination purchase "the safest annuity available." Interpretive Bulletin Relating to the Fiduciary Standard Under ERISA When Selecting an Annuity Provider, 29 C.F.R. § 2509.95-1(c) (1999) (hereafter "IB 95-1" or the "Bulletin"). Although the Bulletin was first published in March 1995 in response to the failure of Executive Life, the Federal Register notes an effective date for IB-95 of January 1, 1975. See Interpretive Bulletins Relating to the Employee Retirement Income Security Act of 1974 (hereafter "IB-ERISA"), 60 Fed. Reg. 12328, 12328 (1995). According to the Secretary, we owe deference to the interpretation of ERISA's fiduciary duties expressed in IB-95, see Chevron U.S.A., Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837 (1984), and should apply it to RJR's selection of Executive Life's annuity.

In Christensen v. Harris County, 120 S.Ct. 1655 (2000), the Supreme Court rejected an argument that it should give "Chevron deference" to a Department of Labor opinion letter. Noting that such interpretations are not "arrived at after, for example, a formal adjudication or notice-and-comment rule making" and "lack the force of law," id. at 1662, it concluded that interpretations in opinion letters and similar documents are instead "`entitled to respect' under [its] decision in Skidmore v. Swift Co., 323 U.S. 134, 140 (1944), but only to the extent that those interpretations have the `power to persuade'." Id. at 1663; see also Martin v. Occupational Safety Health Review Comm'n, 499 U.S. 144, 157 (1991) (noting that interpretive rules and enforcement guidelines are "not entitled to the same deference as norms that derive from the exercise of the Secretary's delegated lawmaking powers").

IB-95 is a Department of Labor interpretative bulletin that is not the product of notice-and-comment procedures established by the Administrative Procedure Act. See 5 U.S.C. § 553 (1994). Although the Department gave advance notice of proposed rule making, see Annuitization of Participants and Beneficiaries Covered Under Employee Pension Plans (hereafter "Annuitization"), 56 Fed. Reg. 28638 (1991), the focus of that notice was not the proper application of § 1104 to a fiduciary's selection of an annuity provider as part of plan terminations. Instead, the notice described the possibility of amending existing regulations defining the circumstances under which an individual is a participant covered under a plan. See id. at 28639. After receiving some responses, see IB-ERISA, 60 Fed. Reg. at 12329, the Department determined that "no regulatory action should be taken at this time to amend the minimum standards under the regulation at 29 CFR 2510.3-3(d)(2)(ii)." Id.

Rather than undertaking regulatory action, the Department, seeing a need for "further guidance regarding the selection of . . . annuity providers by plan fiduciaries," published the Bulletin. IB-ERISA, 60 Fed. Reg. at 12328. The Department noted that the "bulletin concerns solely the fiduciary standard and is published in addition to and independent of the regulatory minimum standard at 29 C.F.R. § 2510.3-3(d)(2)(ii)." Id. at 12329. The Secretary's position is that the Bulletin "announce[s] to the public the Department's legal view of ERISA." Secretary's Brief at 17-18. Because the Bulletin is not the product of notice-and-comment rule making, and does not have the force of law, we apply the standard referred to in Christensen, and determine the extent to which the Bulletin is "entitled to respect." Skidmore, 323 U.S. at 140.

We begin our inquiry with a discussion of the Bulletin's provisions. Subsection (c) provides that in discharging its duty of loyalty in purchasing an annuity, a fiduciary "must take steps calculated to obtain the safest annuity available, unless under the circumstances it would be in the interests of participants and beneficiaries to do otherwise." 29 C.F.R. § 2509.95-1(c) (1999). Although this would appear to impose on fiduciaries an obligation to attempt to obtain the safest annuity, the Bulletin also states that "there are situations where it may be in the interest of the participants and beneficiaries to purchase other than the safest available annuity." Id. § 2509.95-1(d). In cases involving overfunded plans, the Bulletin provides that a fiduciary "must make diligent efforts to assure that the safest available annuity is purchased." Id. This language strongly suggests that the Secretary interprets ERISA's duty of loyalty as requiring that a fiduciary selecting an annuity for purposes of plan termination actually purchase the safest annuity, unless circumstances of the type indicated exist. These circumstances include where the safest annuity is only marginally safer yet disproportionately more expensive and where the insurer offering the safest annuity is unable to administer the plan. See id.

The Secretary's brief also argues that a fiduciary under the circumstances of this case is obligated to purchase the safest annuity available. The Secretary contends that the relevant issue before us is not whether Executive Life was a viable or sound candidate, as RJR argues, but instead "whether Executive Life's annuity was the safest available annuity." According to the Secretary, Shultz and Tyner acted consistently with their fiduciary duties only if they could answer this question in the affirmative.

We agree with the Bulletin and the Secretary that once the decision to terminate a plan has been made, the primary interest of plan beneficiaries and participants is in the full and timely payment of their promised benefits. We agree that beneficiaries and participants whose plan is being terminated gain nothing from an annuity offered at a comparative discount by a provider that brings to the table a heightened risk of default. We would even add that the purchase of such an annuity can be considered an example of the imposition on annuitants of uncompensated risk — the risk of default is borne by the annuitants and, in those states that have guaranty associations, by those associations, while the benefit is granted to the sponsor in the form of a lower price and larger reversion.

However, we are not persuaded that § 1104(a) imposes on fiduciaries the obligation to purchase the "safest available annuity" in order to fulfill their fiduciary duties. We hold that the proper standard to be applied to this case is the standard applicable in other situations that involve the potential for conflicting interests: fiduciaries act consistently with ERISA's obligations if "their decisions [are] made with an eye single to the interests of the participants and beneficiaries." Bierwirth, 680 F.2d at 271; see, e.g., Metzler, 112 F.3d at 213; Pilkington PLC v. Perelman, 72 F.3d 1396 (9th Cir. 1995); Compton, 57 F.3d at 291; Deak v. Masters, Mates Pilots Pension Plan, 821 F.2d 572, 580 (11th Cir. 1987), cert. denied, 484 U.S. 1005 (1988); Leigh v. Engle, 727 F.2d 113, 125 (7th Cir. 1984) ("Leigh I"). That standard does not require that a fiduciary under the circumstances of this case purchase the "safest available annuity." Cf. Riley v. Murdock, No. 95-2414, 1996 WL 209613, at *1 (4th Cir. Apr. 30, 1996) (unpublished) (rejecting the standard advocated by the Department of Labor).

The Bulletin's standard focuses on the quality of the selected annuity. The standard we apply focuses instead on the fiduciary's conduct. It requires that fiduciaries keep the interests of beneficiaries foremost in their minds, taking all steps necessary to prevent conflicting interests from entering into the decision-making process. See Metzler, 112 F.3d at 213 (noting that steps necessary to reduce the effects of potential conflicts are dependent upon the circumstances); Bierwirth, 680 F.2d at 276 (stating that the conflicted trustees "were bound to take every feasible precaution to see that they had carefully considered the other side . . . ."). Although a fiduciary's ultimate choice may be evidence that the duty of loyalty has been breached, the proper inquiry has as its central concern the extent to which the fiduciary's conduct reflects a subordination of beneficiaries' and participants' interests to those of a third party. Cf. Leigh v. Engle, 858 F.2d 361 (7th Cir. 1988) ("Leigh II") ("[W]hether the investments were speculative is irrelevant. The administrators' breach did not consist of investment in speculative assets. Rather, the administrators breached their duties when they made investment decisions out of personal motivations, without making adequate provision that the trust's best interests would be served.").

3. The Duty of Care

We recently addressed an ERISA fiduciary's duty of care in Laborers National Pension Fund v. Northern Trust Quantitative Advisors, Inc., 173 F.3d 313 (5th Cir.), cert. denied sub nom, Laborers Nat'l Pension Fund v. American Nat'l Bank Trust Co., 120 S.Ct. 406 (1999). The issue in Laborers was whether a pension fund's investment manager violated its duty of care when it purchased interest-only mortgage-backed securities. Although the case before us arises in a different context, we find the discussion in Laborers instructive:

In determining compliance with ERISA's prudent man standard, courts objectively assess whether the fiduciary, at the time of the transaction, utilized proper methods to investigate, evaluate and structure the investment; acted in a manner as would others familiar with such matters; and exercised independent judgment when making investment decisions. [ERISA's] test of prudence . . . is one of conduct, and not a test of the result of performance of the investment. The focus of the inquiry is how the fiduciary acted in his selection of the investment, and not whether his investments succeeded or failed. Thus, the appropriate inquiry is whether the individual trustees, at the time they engaged in the challenged transactions, employed the appropriate methods to investigate the merits of the investment and to structure the investment.

Id. at 317 (alterations in original) (internal citations and quotation marks omitted); see also In re Unisys Sav. Plan Litig., 173 F.3d 145, 153 (3d Cir.) ("Unisys II") (noting that the prudence requirement focuses on whether "a fiduciary employed the appropriate methods to investigate and determine the merits of a particular investment"), cert. denied sub nom, Meinhardt v. Unisys Corp., 120 S.Ct. 372 (1999); DeBruyne v. Equitable Life Assurance Soc'y, 920 F.2d 457, 465 (7th Cir. 1990) (agreeing with the lower court that ERISA's duty of care requires "prudence, not prescience"). What the appropriate methods are in a given situation depends on the "character" and "aim" of the particular plan and decision at issue and the "circumstances prevailing" at the time a particular course of action must be investigated and undertaken. 29 U.S.C. § 1104(a)(1)(B); see also Cunningham, 716 F.2d at 1467.

A fiduciary's duty of care overlaps the duty of loyalty. See Bierwirth, 680 F.2d at 271. The presence of conflicting interests imposes on fiduciaries the obligation to take precautions to ensure that their duty of loyalty is not compromised. As we have noted, "[t]he level of precaution necessary to relieve a fiduciary of the taint of a potential conflict should depend on the circumstances of the case and the magnitude of the potential conflict." Metzler, 112 F.3d at 213. To ensure that actions are in the best interests of plan participants and beneficiaries, fiduciaries under certain circumstances may have to "at a minimum" undertake an "intensive and scrupulous independent investigation of [the fiduciary's] options." Leigh I, 727 F.2d at 125-26 (citing Bierwirth, 680 F.2d at 272). In some instances, the only open course of action may be to appoint an independent fiduciary. See Leigh I, 727 F.2d at 125; Bierwirth, 680 F.2d at 271-72.

With regard to the duty of care in the circumstances of this case, IB 95-1 states that ERISA "requires, at a minimum, that plan fiduciaries conduct an objective, thorough and analytical search for the purpose of identifying and selecting providers from which to purchase annuities." Id. § 2509.95-1(c). The Bulletin notes several factors that should be considered in the search, including the "quality and diversification" of an insurer's portfolio, the size of the insurer, the insurer's exposure to liability, and the safety provided by the structure of the annuity contract. See id. § 2509.95-1(c)(1)-(5). "Reliance solely on ratings provided by insurance rating services would not be sufficient . . . ." Id. § 2509.95-1(c). The Bulletin suggests that fiduciaries with a conflict of interest take special precautions in a reversion situation. It exhorts such fiduciaries "to obtain and follow independent expert advice calculated to identify those insurers with the highest claims-paying ability willing to write the business." Id. § 2509.95-1(e).

We view the Bulletin's description of the nature of the investigation to be undertaken in the circumstances of this case as fully consistent with ERISA's provisions and with courts' holdings, including our own. See, e.g., Laborers, 173 F.3d at 317. When selecting an annuity provider to facilitate the termination of a vastly over-funded defined benefit pension plan, the plan's fiduciary must structure and conduct a "careful and impartial investigation" aimed at identifying providers whose annuity the fiduciary may "reasonably conclude best to promote the interests of participants and beneficiaries" of the plan. Bierwirth, 680 F.2d at 271. Of course, many factors must be weighed in determining which provider or providers are best-suited to promote those interests.

In this regard, we find the factors enumerated in IB 95-1 instructive. The relevant inquiry in any case is whether the fiduciary, in structuring and conducting a thorough and impartial investigation of annuity providers, carefully considered such factors and any others relevant under the particular circumstances it faced at the time of decision. If so, a fiduciary satisfies ERISA's obligations if, based upon what it learns in its investigation, it selects an annuity provider it "reasonably concludes best to promote the interests of [the plan's] participants and beneficiaries." Bierwirth, 680 F.2d at 271. If not, ERISA's obligations are nonetheless satisfied if the provider selected would have been chosen had the fiduciary conducted a proper investigation. See Unisys II, 173 F.3d at 153-54 (affirming district court's holding, after a bench trial, that a hypothetical prudent person would have invested in Executive Life guaranteed investment contracts for an ongoing plan); Roth v. Sawyer-Cleator Lumber Co., 16 F.3d 915, 919 (8th Cir. 1994) ("Even if a trustee failed to conduct an investigation before making a decision, he is insulated from liability if a hypothetical prudent fiduciary would have made the same decision anyway.").

A fiduciary must consider any potential conflict of interest, such as a potential reversion of plan assets, and structure its investigation accordingly. Engaging the services of an independent, outside advisor may serve the dual purposes of increasing the thoroughness and impartiality of the relevant investigation, and of relieving the fiduciary of any taint of a potential conflict. In the circumstances of this case, such purposes are served when the outside advisor's task is directed to identifying the provider or providers that best promote the beneficiaries' interests.

Fiduciaries investigating annuity providers to facilitate the termination of an over-funded defined benefit plan, like fiduciaries in other circumstances, are entitled to rely on the advice they obtain from independent experts. See Cunningham, 716 F.2d at 1474 ("ERISA fiduciaries need not become experts in the valuation of closely-held stock — they are entitled to rely on the expertise of others."). Those fiduciaries may not, however, rely blindly on that advice. See id. ("An independent appraisal is not a magic wand that fiduciaries may simply waive over a transaction to ensure that their responsibilities are fulfilled. It is a tool and, like other tools, is useful only if used properly."); Howard v. Shay, 100 F.3d 1484, 1490 (9th Cir. 1996) ("Conflicted fiduciaries do not fulfill ERISA's investigative requirements by merely hiring an expert."), cert. denied, 520 U.S. 1237 (1997); Donovan v. Mazzola, 716 F.2d 1226, 1234 (9th Cir. 1983) ("[R]eliance on counsel's advice, without more, cannot be a complete defense to an imprudence charge."), cert. denied, 464 U.S. 1040 (1984); Bierwirth, 680 F.2d at 272. In order to rely on an expert's advice, a "fiduciary must (1) investigate the expert's qualifications, (2) provide the expert with complete and accurate information, and (3) make certain that reliance on the expert's advice is reasonably justified under the circumstances." Howard, 100 F.3d at 1489 (citing Cunningham, 716 F.2d at 1467, 1474) (other citation omitted); see also Hightshue v. AIG Life Insurance Co., 135 F.3d 1144, 1148 (7th Cir. 1998); In re Unisys Sav. Plan Litig., 74 F.3d 420, 435-36 (3d Cir. 1996) ("Unisys I") ("[W]e believe that ERISA's duty to investigate requires fiduciaries to review the data a consultant gathers, to assess its significance and to supplement it where necessary.").

A determination whether a fiduciary's reliance on an expert advisor is justified is informed by many factors, including the expert's reputation and experience, the extensiveness and thoroughness of the expert's investigation, whether the expert's opinion is supported by relevant material, and whether the expert's methods and assumptions are appropriate to the decision at hand. See, e.g., Hightshue, 135 F.3d at 1148; cf. Howard, 100 F.3d at 1490 ("To justifiably rely on an independent appraisal, a conflicted fiduciary need not become an expert in the valuation of closely held corporations. But the fiduciary is required to make an honest, objective effort to read the valuation, understand it, and question the methods and assumptions that do not make sense."). The goal is not to duplicate the expert's analysis, but to review that analysis to determine the extent to which any emerging recommendation can be relied upon. Cf. Cunningham, 716 F.2d at 1474 (holding that fiduciaries, who had information available to them indicating that assumptions underlying an expert's appraisal were no longer valid, breached their duties under ERISA by not analyzing the effect of changes on those assumptions).

Just as with experts' advice, blind reliance on credit or other ratings is inconsistent with fiduciary standards. See Pilkington, 72 F.3d at 1400 ("Legal authority does not support [the fiduciaries'] contention that a mere ratings scan satisfied their fiduciary duty of loyalty to the plan."); Unisys I, 74 F.3d at 436-37 (citing Cunningham in support of its determination that whether a "rating was a reliable measure of Executive Life's financial status under the circumstances and whether Unisys was capable of using the rating effectively" were matters to be decided at trial). Reviewing the ratings assigned by different rating agencies may be a good place to begin the inquiry, but it certainly is not a proper place to end it.

As with an expert's advice, fiduciaries must determine the extent to which reliance on ratings is reasonably justified under the circumstances. Some ratings agencies are more highly regarded than others. Ratings in general reflect an agency's evaluation of a company, not its evaluation of a company's particular product line. Different agencies' ratings reflect the application of different methodologies. At any given time, agencies' ratings will vary as to their recency. As evidence in the record before us suggests, an agency's rating of a particular company may be perceived by investors and industry insiders as incorrect. Reports accompanying ratings provide fiduciaries with a means of assessing the basis for the particular rating and of identifying what additional information may need to be considered. As with the use of experts, a fiduciary need not duplicate the analysis conducted by the ratings agencies. However, the duty of care imposes on the fiduciary an obligation to ascertain the extent to which the ratings can be relied upon in making the decision at hand.

Assuming a proper investigation has been conducted, a fiduciary does not violate its duties under ERISA simply because an action it determines best promotes participants' and beneficiaries' interests "incidentally benefits the corporation." Bierwirth, 680 F.2d at 271. Appellants charge that RJR selected Executive Life because it submitted the lowest bid and in so doing, violated its duty of loyalty. RJR does not deny that cost was a basis for its decision, and instead contends that it could choose the lowest-cost provider under the circumstances. Under the standard we apply, an annuity's price cannot be the motivating factor until the fiduciary reasonably determines, through prudent investigation, that the providers under consideration are comparable in their ability to promote the interests of participants and beneficiaries. Without such a prior determination, consideration of an annuity's price, because it directly benefits the employer, can be taken as evidence that a fiduciary has placed an interest in a reversion above the interests of plan beneficiaries.

Of course, the comparison of annuity providers must be made considering factors relevant to plan beneficiaries' and participants' interests. As a general matter, we expect that a proper investigation of potential annuity providers will reveal that each has its own "warts." We do not view the presence of such blemishes, by itself, to be sufficient to cause a fiduciary to eliminate those providers from further consideration. The issue is whether a provider's warts, viewed qualitatively and quantitatively, are such that a prudent person in like circumstances would determine that the purchase of that provider's annuity was not in the best interests of plan beneficiaries and participants. Having concluded that all remaining providers are comparable in their ability to serve the best interests of plan beneficiaries and participants, a fiduciary does not violate ERISA's commands by subsequently considering which provider offers its annuity at a lower price.

C. RJR's Compliance with its Fiduciary Obligation

Keeping in mind the standards set forth above, we must determine whether reasonable and fair-minded persons could conclude from the summary judgment evidence that RJR breached its fiduciary duties in selecting Executive Life's annuity. Based upon a careful review of the record in this case, we conclude that it was inappropriate for the district court to grant summary judgment in favor of RJR. Viewing the evidence in the light most favorable to Appellants, a reasonable fact finder could conclude that RJR failed to structure, let alone conduct, a thorough, impartial investigation of which provider or providers best served the interests of the participants and beneficiaries. Even if the fact finder were to conclude that RJR's investigation was appropriate, it could conclude, based on the evidence, that RJR could not reasonably determine that Executive Life best promoted the interests of plan participants and beneficiaries. Finally, moving on to the hypothetical prudent person standard, a reasonable fact finder could also conclude that Executive Life was not an objectively reasonable choice based upon the information RJR should have gathered.

1. The Investigation A reasonable fact finder could conclude that RJR did not structure or conduct an independent and impartial investigation directed to identifying a carrier that it could "reasonably conclude [was] best to promote the interests of participants and beneficiaries" of the plan. Bierwirth, 680 F.2d at 271. Given the decision to terminate a defined benefit plan, the primary interest of participants and beneficiaries was in the full and timely payment of their promised benefits. The record shows that RJR employed Buck to assist it in selecting an annuity provider, and looked to Buck to assess the solvency and safety of the bidding companies. Overgard, a Buck consultant, stated in his deposition that his analysis of the insurers' financial health was limited to a review of the rating agencies' ratings and reports. He also stated that he had spent less time on evaluating companies than, as Overgard put it, "on stuff that [Buck] had been hired to do, and that is to work with the insurance companies to get the best bids."

Overgard, who was responsible for compiling an initial list of insurance companies that could provide the annuity, determined, after a discussion with a colleague, that Executive Life ought not be included on that list because it used a nontraditional investment strategy that featured a high percentage of low-quality bonds. When the list compiled by its expert did not include Executive Life, Tyner specifically requested that the company be added because its expected lower bid could be used to drive down the bids of other providers. Tyner, at the time he requested Executive Life be included, did not think that "Executive Life should be seriously considered in the final bidding process." He anticipated that another, "more well-known" company would ultimately be selected.

The record contains evidence that Overgard undertook some investigation of Executive Life beyond his examination of the ratings (e.g., determining that Moody's had not talked with Executive Life management prior to assigning its rating, talking with investment bankers, pursuing industry intelligence). He found opinions regarding Executive Life to be mixed, with some industry insiders viewing the company's investment strategy as bad. Again, Overgard did not review any of Executive Life's financial statements, reports, or disclosures, or conduct a special financial analysis of Executive Life or any other provider. The record indicates that Overgard was not aware that California regulators were looking into Executive Life's reinsurance practices, and did not recall whether he knew, prior to the final bid day, that states' regulators had capped, or were considering capping, insurance companies' investment in low-quality bonds. In Overgard's opinion, positive attributes, such as Executive Life working the case harder, being more professional, and asking more questions, kept the company on the list. A reasonable fact finder could conclude that Buck included Executive Life on the final list of bidders in spite of its nontraditional investment strategy specifically because of the request of RJR, its client. Executive Life's low bid could not be used to drive down the bids of other providers unless it was included on the final list.

The record also includes indications that RJR did not ascertain, prior to selecting Executive Life, what Overgard had done to assess the safety of the companies interested in RJR's business other than look at the ratings, which Overgard had provided to RJR. It could be concluded based on evidence in the record that despite RJR's request that Executive Life be placed on the list to drive down other providers' bids, RJR did not ascertain the basis for Buck's statement that the company was "qualified." A reasonable fact finder could also conclude that RJR failed to assess the basis for Buck's statement that all four providers were "qualified" to provide the annuity, cf. Unisys I, 74 F.3d at 435-36 (concluding, when confronted with similar evidence, that summary judgment in favor of the defendant was inappropriate), and failed to ascertain whether Buck's statement meant that RJR could view each of the companies as comparable.

Focusing on whether RJR undertook activities to investigate the safety of the carriers interested in bidding, a reasonable fact finder could conclude that the company relied entirely on ratings that Buck provided it. The record indicates that RJR looked to those ratings to examine the safety of Executive Life. Both Shultz and Tyner stated that they had not read the accompanying reports. Tyner assumed that negative information that existed would be reflected in agency ratings. In his deposition, Tyner stated that to his knowledge, no one checked why Moody's had given Executive Life a lower rating. Tyner also stated that he did not look at Executive Life's annual reports or SEC filings. As with Buck's recommendation, a fact finder could conclude that RJR failed to assess the extent to which it was justified in relying upon the ratings assembled by Buck, and that the bulk of RJR's investigation was a review of those ratings.

A fact finder could conclude that the absence of an independent investigation by RJR is made more egregious by the fact that Shultz (who bore the responsibility for making the final decision on behalf of RJR) apparently possessed a good deal of information about Executive Life and the emerging problems in the market for low-quality bonds. See Part I supra. Yet he did nothing to ascertain whether Tyner was in possession of that information, let alone whether he had conducted further investigation (either personally or through Buck) to determine that Executive Life was a provider qualified to be on the final list.

A fact finder could conclude that as far as RJR knew on August 12, 1987, its investigation of the providers involved (1) hiring Buck, which scanned the ratings, and (2) scanning the ratings itself. RJR asserts that this represents the normal investigation undertaken at the time by fiduciaries purchasing annuities from insurance companies that are heavily regulated by the states, and points to a statement of one of its experts, who had not acted as a fiduciary, for support for this contention. However, the record also contains statements from Appellants' experts, three of whom had acted as a fiduciary, that RJR's practices breached its fiduciary duties. Given this case is before us on summary judgment, we leave to the fact finder the task of making credibility assessments. See Anderson, 477 U.S. at 255 ("Credibility determinations, the weighing of the evidence, and the drawing of legitimate inferences from the facts are jury functions, not those of a judge, whether he is ruling on a motion for summary judgment or for a directed verdict."). We note that a reasonable fact finder could conclude from the evidence that application of the "normal" investigation was not sufficient under the circumstances. Executive Life's investment strategy deviated significantly from the norm, was comparatively untested, and was the subject of debate among industry insiders. Moreover, evidence in the record suggests that some investors viewed Executive Life's SP rating as incorrect.

In short, a reasonable fact finder could conclude from evidence in the record that RJR made an insufficient attempt to identify which provider or providers was best positioned to promote the interests of the participants and beneficiaries. Based upon its lack of understanding of the basis for Buck's statement that all four bidders on the final list were "qualified," its failure to assess the extent to which ratings could be reasonably relied upon, and its failure to consider factors beyond ratings provided by Buck, a reasonable fact finder could conclude that RJR failed to structure and conduct a prudent investigation. Even if it had long been the practice of those purchasing annuities to rely solely on a ratings scan, a fact finder could conclude that such an investigation was inappropriate in light of lack of experience that the industry, and its regulators, had with Executive Life's investment strategy. Were a fact finder to conclude that RJR's investigation was inadequate under the circumstances, RJR would no longer be entitled to rely on the reasonableness of its final selection based upon the information its investigation produced.

Even if RJR's investigation were to be found proper, a reasonable fact finder could conclude that RJR, based on the information it had, was unreasonable in considering the four providers comparable in their ability to serve the interests of plan beneficiaries and participants. The record indicates that the four companies were identical in only one dimension — the ratings given by SP. Beyond this, there was variation in the ratings given to the four companies, with Executive Life receiving a Moody's rating two grades lower than AAA. A fact finder could conclude that Shultz was aware of a number of facts regarding Executive Life, including that over 50% of its portfolio was in low-quality bonds, that in this way Executive Life was unusual among insurance companies, and that there was mixed opinion regarding both Executive Life's strategy (involving, as it did, investing over 50% of its portfolio in low-quality bonds) and the soundness of investing in low-quality bonds generally. Shultz understood the connection between Drexel and Executive Life, and that Executive Life came within the scope of then-on-going government investigations. Shultz had not seen the same variation in views of the other companies as he had seen with Executive Life. There is evidence in the record that of the final four companies, RJR first used price to reduce the field to two, and then simply went with the lowest bidder. For example, Aetna was dropped from consideration midday because of price; Prudential and Executive Life were considered further because they were the low bidders. Executive Life was chosen over Prudential because of price. From this, and other evidence in the record, a reasonable fact finder could conclude that RJR placed its interests in the reversion ahead of the beneficiaries' interests in full and timely payment of their benefits.

2. The Hypothetical Prudent Person Standard

Similarly, a reasonable fact finder could conclude that Executive Life was not an appropriate choice based upon the investigation that RJR should have conducted. There is evidence that many voices in the industry had concerns about Executive Life's investment strategy — a strategy that was substantially different from that used by the industry and that had not stood the test of time. As such, there was more uncertainty (and more associated risk) with Executive Life than with the other candidates. A fact finder could conclude on this basis alone that a prudent person would not select Executive Life's annuity over the annuities offered by those candidates.

The record supplies the fact finder with considerable additional evidence that leads to the same conclusion. A fact finder could conclude from evidence in the record that the vast majority of insurance companies at the time rejected the type of investment strategy that Executive Life had adopted, despite Executive Life's ability to underbid other firms and their resulting economic incentive to adopt a similar strategy. Evidence in the record also suggests that some were critical of SP giving a high rating to Executive Life, that Duff Phelps gave the company its ninth rating, and that Moody's had assigned its lower rating to Executive Life in part because of the quality of its bonds. Moreover, Executive Life was, during the relevant period, under investigation by both New York and California regulators. New York regulators had levied a hefty fine against Executive Life's New York subsidiary, and had placed a cap (of 20%) on the low-quality bond holdings of insurance companies that state regulated. Documentation filed by First Executive indicated that the company saw adoption of caps by New York regulators as threat to its future growth, competitiveness, and profitability. Other states' regulators, including those in California, were considering capping investment in such bonds. Although evidence was presented that investment banking firms (in addition to Drexel) were eager to make a market in low-quality bonds, there is also evidence that the low-quality bond market as a whole would suffer as a result of investigations of Drexel that were ongoing at the time RJR chose Executive Life. There is evidence that one reputable consultant had removed Executive Life from its Approved List in 1985. A reasonable fact finder could conclude that an appropriate investigation would have revealed this information and that such information, when weighed against the information that should have been gathered on other providers, would cause a fiduciary to eliminate Executive Life as a final candidate well before price could be legitimately considered. Cf. Pilkington, 72 F.3d at 1401-02 (holding that summary judgment in favor of defendant was inappropriate where evidence in the record indicated the investigation of Executive Life relied on a "mere ratings scan," that "voices in the insurance industry had expressed misgivings about the soundness of those ratings," and that "reversion maximization figured prominently in [the sponsor's] spin-off/plan termination decision"); Unisys I, 74 F.3d at 435-37 (holding that summary judgment in favor of the defendant was inappropriate given, inter alia, evidence that allowed a fact finder to infer that Unisys "failed to analyze the bases underlying [its expert's] opinion of Executive Life's financial condition and to determine for itself whether credible data supported [the expert's] recommendation," a subsequent investigation "consisted of nothing more than confirming that Executive Life's credit ratings had not changed," and evidence in the record that raised issues as to whether reliance on ratings was justified).

Given the factual differences between the two cases and the fact-specific nature of our inquiry, we do not view Unisys II, 173 F.3d 145 (3d Cir. 1999), as dictating a different conclusion. In that case, the court affirmed the lower court's determination, after a bench trial and additional findings of fact, that the fiduciaries' purchase of Executive Life guaranteed investment contracts did not violate ERISA. We note that although those fiduciaries were buying products sold by Executive Life, they were not buying an annuity to facilitate the termination of a defined benefit pension plan. The investments at issue constituted only 15-20% of a fund that was just part of the retirement plan at issue in that case. See id. at 152 n. 10. As a result, we do not find Unisys II's ultimate conclusion dispositive.

For similar reasons, we also do not regard Riley v. Murdock, 890 F. Supp. 444 (E.D.N.C. 1995), aff'd, No. 95-2414, 1996 WL 209613 (4th Cir. Apr. 30, 1996) (unpublished), as dispositive. In that case, as in this, an Executive Life group annuity was purchased to facilitate the termination of an over-funded defined benefit pension plan. The Riley court explained that to assess prudence it first inquired "whether the fiduciary employed the appropriate methods to diligently investigate the transaction." 890 F. Supp. at 458. Next, it determined whether "the decision ultimately made was reasonable based upon the information resulting from the investigation." Id. The court detailed the extensive actions taken by the fiduciaries in that case, explained that the plaintiffs had presented no evidence that the fiduciaries should have known about problems with Executive Life in 1986, and concluded, "[a]ll of these efforts establish that [the fiduciaries] thoroughly investigated the purchase of the annuity from Executive Life and that the decision to purchase was reasonable based on the results of that investigation." Id. (emphasis added).

The Riley court's conclusion can not be translated into a pronouncement that the purchase of an Executive Life group annuity to facilitate plan termination was objectively reasonable in 1987 regardless of the investigation conducted. Not only did RJR have an additional year of information available to it, but the Riley court never addressed the objective prudence of a decision to invest in an Executive Life group annuity. Finding that the fiduciaries in that case conducted a prudent investigation and that their decision was reasonable based upon that investigation, the Riley court did not have cause to apply the hypothetical prudent person standard.

III. CLASS CERTIFICATION

The district court denied the motion to certify a class for the reason that "neither of the named plaintiffs will recover anything by this suit." Bussian, 21 F. Supp.2d at 684. We have concluded that summary judgment was inappropriate. Under the circumstances, it seems appropriate to vacate the district court's order denying class certification and allow it to consider the issue more fully on remand.

IV. CONCLUSION

For the foregoing reasons, the grant of summary judgment in favor of RJR is REVERSED, and the order denying class certification is VACATED. The case is REMANDED to the district court. Costs shall be borne by RJR.