United States Court of Appeals For the Eighth Circuit ___________________________ No. 15-2792 ___________________________ Ronald C. Tussey; Charles E. Fisher; Timothy Pinnell lllllllllllllllllllllPlaintiffs - Appellants v. ABB, Inc.; John W. Cutler, Jr.; Pension Review Committee of ABB, Inc.; Pension & Thrift Management Group of ABB, Inc.; Employee Benefits Committee of ABB, Inc. lllllllllllllllllllllDefendants - Appellees Fidelity Management Trust Company; Fidelity Management & Research Company lllllllllllllllllllllDefendants -----------------------------American Association of Retired Persons lllllllllllllllllllllAmicus on Behalf of Appellants Securities Industry and Financial Markets Association lllllllllllllllllllllAmicus on Behalf of Appellees
___________________________ No. 16-1127 ___________________________ Ronald C. Tussey; Charles E. Fisher; Timothy Pinnell lllllllllllllllllllllPlaintiffs - Appellees v. ABB, Inc.; John W. Cutler, Jr.; Pension Review Committee of ABB, Inc.; Pension & Thrift Management Group of ABB, Inc.; Employee Benefits Committee of ABB, Inc. lllllllllllllllllllllDefendants - Appellants Fidelity Management Trust Company; Fidelity Management & Research Company lllllllllllllllllllllDefendants ____________ Appeals from United States District Court for the Western District of Missouri - Jefferson City ____________ Submitted: September 21, 2016 Filed: March 9, 2017 ____________ Before RILEY, Chief Judge, MURPHY and SMITH, Circuit Judges. ____________ RILEY, Chief Judge. A class of employees who participated in ABB, Inc.’s retirement plans—“401(k) defined contribution savings plans,” to be precise, see generally -2-
26 U.S.C. § 401(k)—accuse ABB and its agents (collectively, the ABB fiduciaries) of managing the plans for their own benefit, rather than the participants’. In an earlier appeal, we directed the district court to “reevaluate” how the participants might have been injured if the ABB fiduciaries breached their fiduciary duties under the Employee Retirement Income Security Act of 1974 (ERISA), 29 U.S.C. §§ 1001, et seq., when they changed the investment options for the plans. See Tussey v. ABB, Inc., 746 F.3d 327, 338 (8th Cir. 2014). Because the district court apparently mistook that direction for a definitive ruling on how to measure plan losses, and as a result entered judgment in favor of the ABB fiduciaries despite finding they did breach their duties, we vacate the judgment on that claim and remand for further consideration regarding whether the participants can prove losses to the plans. Because we thus reopen one of the participants’ substantive claims, we also vacate and remand the district court’s award of attorney fees. BACKGROUND1 The plans offered participants a menu of options for investing the money in their accounts.2 In 2000, ABB’s Pension Review Committee adopted a written policy statement describing “the underlying philosophy and process for the selection, monitoring and evaluation and, if necessary, removal of investment options.” The policy statement said the plans would offer investments in three “tiers,” organized by how much active involvement they demanded from investors. The first tier, meant “[f]or participants unwilling or unable to make a personal asset allocation decision,” was to “offer several ‘managed allocation’ funds designed to offer the participant a I.
Our prior opinion recounts the facts of the case more completely. See Tussey, 746 F.3d at 330-33. Here, we focus on what is relevant to this appeal. 2
We accept the parties’ representations that the differences between the two plans at issue—one for ABB’s unionized employees, one for the others—are irrelevant here. -3-
professionally managed, well diversified fund or funds appropriate for the participants’ [sic] investment goals.” John Cutler, Jr., the director of the committee’s staff, thought those “‘managed allocation’” funds should be “target-date” or “life-cycle” funds, which dynamically change their mix of investments to become more conservative as a specified date (such as an employee’s expected retirement) approaches. Cutler favored the Fidelity Freedom Funds, a family of funds with target dates at ten-year intervals from 2000 to 2040. Cutler also suggested removing the Vanguard Wellington Fund, an established fund that invested in stocks and bonds in a generally static ratio. The committee agreed on both points. Removing the Wellington Fund raised the question of what to do with the money participants had invested in it—roughly $123 million, representing about 8.4% of the total assets in the plans. The ABB fiduciaries decided to move the money into the Freedom Funds, a process called “mapping” assets from one investment to another. Participants whose money was mapped to a Freedom Fund remained free to choose a different investment option (or options) at any time, but the mapping decision made the Freedom Funds the default for anyone who had been invested in the Wellington Fund and did not take affirmative steps to do something else with their money. In 2006, the participants sued the ABB fiduciaries and two Fidelity companies—the recordkeeper for the plans and the investment advisor for the Fidelity mutual funds included in the plans—under ERISA. See 29 U.S.C. §§ 1109, 1132(a)(2) (fiduciary liability and cause of action). After a bench trial, the district court found both sets of defendants “breached some fiduciary duties that they owed to the . . . Plans.” See id. § 1104(a)(1) (fiduciary duties). In particular, the district court found the ABB fiduciaries breached their fiduciary duties by (1) deciding effectively to replace the Wellington Fund with the Freedom Funds based on selfinterest rather than what was best for the plans, (2) failing to properly monitor and control recordkeeping costs, and (3) agreeing to make the plans overpay for Fidelity’s -4-
services in return for Fidelity charging ABB less for corporate services it bought for itself. The Fidelity defendants were liable as well, according to the district court, because interest earned when money in the process of being added to or taken out of plan investments was invested overnight—called “float income”—should have been credited to the plans, not back to the investments. The district court awarded the participants $21.8 million against the ABB fiduciaries for swapping the Wellington and Freedom Funds, $13.4 million for the ABB fiduciaries’ other breaches, and $1.7 million against the Fidelity defendants on the float claim, plus attorney fees of $12.9 million from all the defendants jointly and severally, see id. § 1132(g)(1) (attorney fees). The defendants appealed. We vacated the finding of breach for changing the investment options, explaining the district court should have afforded more deference to the discretion the plans explicitly granted the ABB fiduciaries. See Tussey, 746 F.3d at 338. Because the issue could be relevant again on remand—if the district court still found a breach after a properly deferential review—we added that, “[a]s calculated,” the original award for switching the funds was “speculative and exceed[ed] the ‘losses to the plan[s] resulting from’ any fiduciary breach.” Id. at 338 n.7, 339 (quoting 29 U.S.C. § 1109(a)). The district court had calculated the plans’ losses by comparing the returns on the Freedom Funds to what the participants would have earned if they had invested in the Wellington Fund instead. We suggested “it seems the participants’ mapping damages, if any, would be more accurately measured by comparing the difference between the performance of the Freedom Funds and the minimum return of the subset of managed allocation funds the ABB fiduciaries could have chosen without breaching their fiduciary obligations.” Id. at 339. We affirmed the ABB fiduciaries’ liability on the other claims, reversed the judgment against the Fidelity defendants, and vacated the fee award—now only against the ABB fiduciaries—so the district court could account for the resolution of the remanded issue. See id. at 336-37, 340-41.
On remand, the district court again held the ABB fiduciaries breached their fiduciary duties. Yet the district court concluded the participants had failed to prove any losses under the theory we “tacitly approved” in the first appeal—comparing the Freedom Funds’ returns to the worst-performing of the funds the ABB fiduciaries could have properly chosen—so the ABB fiduciaries nonetheless prevailed on that claim.3 In light of that result, the district court reduced the participants’ attorney fee award for work through trial by almost $2.2 million, to $10,768,474. The district court also awarded the participants $900,000 for work on the appeal—just over twothirds of what they requested—for a total of $11,668,474. The participants appeal the district court’s ruling on measuring losses and liability for the breach. In a consolidated cross-appeal, the ABB fiduciaries argue both parts of the fee award are still too high. See 28 U.S.C. § 1291 (appellate jurisdiction). II.
DISCUSSION A. Liability We start with the ABB fiduciaries’ assertion, presented as an alternative ground for affirming the district court’s judgment on the merits, that they did not actually breach their fiduciary duties, so there is no need to reach the issue of how much the plans lost. Whether the ABB fiduciaries’ actions constituted a breach is a legal question we must answer de novo. See, e.g., Tussey, 746 F.3d at 333. But what they did, and why, are factual matters on which we accept the district court’s findings unless they are clearly wrong. See, e.g., id.; Herman v. Mercantile Bank, N.A., 137 F.3d 584, 586 (8th Cir. 1998); see also Fed. R. Civ. P. 52(a)(6). The heart of the ruling on remand was the district court’s conclusion that “the removal of the Wellington Fund from the . . . platform and the mapping of its assets
At the end of its order, the district court reassured the participants that “[i]f the Court has misread the Eighth Circuit, its decision . . . can be corrected on appeal.” Thus, here we are. -6-
to the Freedom Funds . . . was motivated in large part to benefit Fidelity Trust and ABB, not the Plan participants.” The ABB fiduciaries insist that determination was the product of speculation, unfounded inferences, and the fallacy of “equat[ing] the effect of a decision with its purpose.” To the contrary, strong evidence supported the district court’s finding, notably the fact that Cutler—the director of the Pension Review Committee’s staff—and ABB’s director of employee benefits openly communicated with Fidelity about the “pricing implications” of changes to the plans’ investment lineup and the specific dollar amounts by which Fidelity would cut its fees “if the Wellington money map[ped]/default[ed] to the Freedom Funds.” The ABB fiduciaries stress that there was a great deal of other evidence too, and some of it showed them acting against Fidelity’s interests in various ways. For instance, the ABB fiduciaries dropped other lucrative Fidelity funds from the plans and mapped their assets to non-Fidelity options. However, their examples all relate to other investment decisions, not the Wellington-Freedom swap. The fact the ABB fiduciaries apparently did not always favor Fidelity as much as they could, or seize every opportunity to send Fidelity more of the participants’ money, does little to undermine the district court’s finding about why they made the particular decisions at issue in this case. That is all the more true given that we know conclusively the ABB fiduciaries, in still other contexts, did “fail to monitor and control [Fidelity’s] recordkeeping fees” and did “pay [Fidelity] excessive revenue sharing from Plan assets.” Tussey, 746 F.3d at 336. In addition to the direct evidence of meetings about “pricing implications,” the district court’s finding was based on what it saw as “circumstantial evidence” of the ABB fiduciaries’ motives. The ABB fiduciaries take issue with the district court’s approach, which they say once again failed to afford proper deference to their choices “in implementing the redesign and evaluating and selecting Plan investment options
in accordance with the Plan.”4 Id. at 338. According to the ABB fiduciaries, the district court was wrong to focus on perceived flaws in the ABB fiduciaries’ decisionmaking process, such as not considering other possible replacements for the Wellington Fund and giving an explanation for mapping the assets from the Wellington Fund to the Freedom Funds that seemingly conflicted with the reason they had given for dropping the Wellington Fund, because the ABB fiduciaries had discretion over such things. The ABB fiduciaries overstate the deference they are owed. “[A] Plan administrator deserves discretion to the extent its ex ante investment choices were reasonable given what it knew at the time.” Id. When the district court considered the circumstances of how the ABB fiduciaries decided to swap the funds, it was not second-guessing whether their decision was reasonable. Nor was it naysaying their underlying determination of the appropriate procedure to use to make such a decision. Rather, the district court was observing that the ABB fiduciaries’ actions—discretionary or not—when taken together and viewed in context, shed light on their motivations. In particular, those actions tended to suggest the ABB fiduciaries did what they did not because they thought it was best for the plans, but
The ABB fiduciaries, along with the Securities Industry and Financial Markets Association (SIFMA) as amicus curiae, also point out that some parts of the district court’s reasoning suggest it took a dim view of target-date funds in general, relative to more traditional funds like Wellington, perhaps based on hindsight bias and a failure to appreciate the differences between the two kinds of products. We agree that simply comparing upfront costs or looking back at the funds’ earnings after the fact would not have been a valid way to determine whether choosing the Freedom Funds over Wellington was prudent. But the district court explicitly recognized that “[t]he relative performance of the Freedom Funds and the Wellington Fund is only relevant to determine whether a loss was sustained by the Plan,” not whether there was a breach in the first place, and that “a non-conflicted fiduciary could have chosen the Freedom Funds,” so we are satisfied any misunderstanding did not affect the district court’s findings. -8-
because they wanted to get a better deal for themselves. As the district court explained, “there [were] too many coincidences to make the beneficial outcome for ABB serendipitous, particularly considering the powerful draw of self-interest when transactions are occurring out of sight and are unlikely to ever be discovered.” A fiduciary can abuse its discretion and breach its duties by acting on improper motives, even if one acting for the right reasons might have ended up in the same place.5 Cf. Metro. Life Ins. Co. v. Glenn, 554 U.S. 105, 117 (2008); id. at 123 (Roberts, C.J., concurring in part and concurring in the judgment); id. at 131 (Scalia, J., dissenting). Because the choice of whose interests to favor, the plans’ or their own, was not one over which the ABB fiduciaries could claim any discretion, cf. 29 U.S.C. § 1104(a) (“[A] fiduciary shall discharge his duties with respect to a plan solely in the interest of the participants and beneficiaries.” (emphasis added)); Tussey, 746 F.3d at 333-35, there was no place for deference in the district court’s factfinding on their motives. There was no clear error in that factfinding either, so we will not disturb the district court’s determination that “but for its conflict of interest, ABB would not have made the same decisions.” And we see no reason to reject the legal conclusion the district court drew from that fact—the ABB fiduciaries abused their discretion and breached their fiduciary duties. That brings us back to the question which prompted this appeal: how to determine what that breach cost the plans? We answer de novo, because the method for measuring losses, as distinct from the calculation itself, is a legal issue, see id. at 338-39, as is the explication of our previous decisions.
Such a breach might not be actionable, if it truly did not change the result and thus did not cause any harm, but that is a separate question. See Roth v. SawyerCleator Lumber Co., 16 F.3d 915, 919 (8th Cir. 1994). -9-
What we said on this point was: On remand, the district court should reevaluate its method of calculating the damage award, if any, for the participants’ investment selection and mapping claims. First, the district court awarded the amount that participants who had invested in the Wellington Fund presumably would have had if (1) ABB had not replaced the Wellington Fund with the Freedom Funds, and (2) the participants remained invested in the Wellington Fund for the entire period at issue. In light of the [policy statement’s] requirement to add a managed allocation fund, it seems the participants’ mapping damages, if any, would be more accurately measured by comparing the difference between the performance of the Freedom Funds and the minimum return of the subset of managed allocation funds the ABB fiduciaries could have chosen without breaching their fiduciary obligations. Second, the district court determined “it [was] a reasonable inference that participants who invested in the Freedom Funds would have invested in the Wellington Fund had it not been removed from the Plan’s investment platform.” Such an inference appears to ignore the investment provisions of the [policy statement], participant choice under the Plan, and the popularity of managed allocation funds. And the participants fail to cite any evidentiary support for inferring the participants’ voluntary, post-mapping investments in the Freedom Funds would have instead been made in the Wellington Fund, even if that fund remained as a Plan option for all of the years at issue. A reasonable inference is one which may be drawn from the evidence without resort to speculation. As calculated, the $21.8 million damage award for the participants’ mapping claim is speculative and exceeds the losses to the plan resulting from any fiduciary breach. Id. (second alteration in original) (emphasis added) (footnote, quotation marks, and citations omitted). According to the ABB fiduciaries, the italicized language was a “binding alternative holding,” and thus became the “law of the case,” because it was
necessary to give the district court a standard to apply on remand.6 We think that gives the highlighted language too much weight. To be sure, we did clearly rule that the original award was wrong “[a]s calculated.” Id. at 339. At the same time, we left open exactly how it should be fixed. We suggested two points in the district court’s explanation that “seem[ed]” or “appear[ed]” to be mistaken, but we did so in tentative, qualifying terms, rather than the firm, definite language we used for our holdings (the award “is speculative and exceeds” the plans’ losses, and the district court “should reevaluate” its methodology).7 Id. at 338-39 (emphasis added). Properly read, the passage at issue proposed an alternative we thought warranted consideration (if measuring the plans’ losses became necessary again on remand), it did not require that the district court adopt our proffered approach. That is why our overarching instruction to the district court was to “reevaluate its method of calculating the damage award.” Id. at 338. With that phrasing, we meant to make clear both that there was work—“reevaluat[ion]”—left for the district court to do and the work involved resolving the “method of calculating” losses, not just their ultimate amount. Such a directive would have made little sense if, as the ABB fiduciaries assert, all we meant 6
The ABB fiduciaries also invoke the rules that one panel of this court cannot overrule another, see, e.g., Maxfield v. Cintas Corp., No. 2, 487 F.3d 1132, 1135 (8th Cir. 2007), and that the district court must strictly obey our mandate, see, e.g., Bethea v. Levi Strauss & Co., 916 F.2d 453, 456 (8th Cir. 1990). Those theories boil down to the same thing as the “law of the case” argument, namely an insistence that our statement amounted to a definitive ruling on how to measure losses from the breach. 7
The ABB fiduciaries urge us to disregard this contrast in phrasing because in 1899 another court of appeals considered itself bound by a statement of law in a Supreme Court decision even though it contained the words “it would seem.” See Anderson v. Reid, 14 App. D.C. 54, 81-82 (D.C. Cir. 1899). We do not mean to suggest such qualifications automatically render any statement to which they are attached nonbinding. Yet they are not always superfluities either. Like any other words in a judicial proclamation, they must be read and understood in context. -11-
for the district court to do was carry out the calculations under an approach we had dictated on appeal. The district court therefore should have considered other ways of measuring the plans’ losses from the ABB fiduciaries’ breach, as well as the participants’ contentions about why, in their view, our proposal was misguided and contradicted persuasive authority and the trust-law principles that generally inform ERISA decisions, cf., e.g., Donovan v. Bierwirth, 754 F.2d 1049, 1056 (2d Cir. 1985) (“Where several alternative investment strategies were equally plausible, the court should presume that the funds would have been used in the most profitable of these. The burden of proving that the funds would have earned less than that amount is on the fiduciaries found to be in breach of their duty.”). We leave to the district court’s discretion whether and how to expand the record and hear additional argument from the parties on this issue. Although this case arguably illustrates the dangers of saying more than we need to, we add two observations to inform the analysis on remand. First, the district court treated our reference to measuring the Freedom Funds against “the subset of managed allocation funds the ABB fiduciaries could have chosen,” Tussey, 746 F.3d at 339, as limiting the comparison to “alternative target[date] fund[s].” That reading is echoed in the ABB fiduciaries’ claim on appeal that “this Court determined that the performance of the Freedom family had to be measured against a dynamic ‘managed allocation’ alternative.” It is unclear where the ABB fiduciaries find the requirement of a “dynamic” comparator—we certainly did not use that word. See id. And we see no basis for reading such a restriction into the phrase “managed allocation funds,” either in our opinion or in the investment policy statement we took it from. According to the minutes of the meetings where the changes to the investment options were considered, both times Cutler briefed the Pension Review Committee about managed allocation funds he “not[ed] that there were two types: static and dynamic.” So while dynamic target-date funds are one kind of managed allocation fund, they are not the only one. Indeed, absent other -12-
evidence of a more limited technical definition, as a matter of plain meaning the Wellington Fund itself would appear to be a static managed allocation fund—a portfolio manager actively monitors how its assets are allocated, keeping the overall mix of investments roughly constant but selecting the stocks, bonds, and other securities it holds at any given time. Second, it is a mistake to argue, as the ABB fiduciaries and SIFMA do, that measuring any portion of the losses by comparing the returns from the Freedom Funds with what the plans would have earned from the Wellington Fund is necessarily inappropriate because it involves “an apples-to-oranges comparison.” True, the funds are designed for different purposes and thus choose their investments differently, so there is no reason to expect them to make similar returns over any given span of time. But the point of the comparison here would just be to determine, as a factual matter, the effect of owning one fund rather than the other. The reason for any difference in returns would be immaterial.8 This is not to say we are sure a comparison with the returns on the Wellington Fund must be part of measuring what the ABB fiduciaries’ breach cost the plans, just that we are unpersuaded by these two arguments against it and do not want them to unduly occupy the parties and the district court on remand. The measure and amount of the plans’ losses remain for the district court to resolve. See generally Martin v. Feilen, 965 F.2d 660, 671-72 (8th Cir. 1992) (insisting the district court perform its “function ‘to fashion the remedy best suited to the harm’” even though the plaintiff had “presented only an unsound . . . damage theory,” and explaining “the measure of 8
By the same logic, making the comparison would not imply a (mistaken) view that whichever fund earned more over the relevant time frame “should” have been offered to the participants, or even that it performed “better” in a meaningful sense. Not only can good bets go bust and bad bets hit the jackpot, but some investments are simply meant to pay off less than others, in return for lower risks, different exposures, or countless other considerations. -13-
such damages need not be exact—‘it will be enough if the evidence show [sic] the extent of the damages as a matter of just and reasonable inference, although the result be only approximate’” (first quoting Garnatz v. Stifel, Nicolaus & Co., 559 F.2d 1357, 1360 (8th Cir. 1977); and then quoting Story Parchment Co. v. Paterson Parchment Paper Co., 282 U.S. 555, 563 (1931))). B. Attorney Fees In light of that disposition, we again vacate the award of attorney fees and remand for the district court to adjust the award as appropriate if the participants ultimately prevail on the liability issue.9 What happens on remand cannot change the outcome of the first appeal. Therefore, we need not wait to review the portion of the fee award corresponding to appellate work. The ABB fiduciaries urge us to conduct this analysis de novo. There is no reason for deference to the district court’s determinations, they explain, because the district court had no particular familiarity with what happened on appeal—if any court has special insight into the work involved, it is ours. We note the district court is still significantly more experienced than we are when it comes to analyzing and awarding attorney fees as a general matter, even if it did not see this part of the litigation firsthand. Questions of relative expertise notwithstanding, our case law is certain that when a district court’s fee award is for work on appeal, “we review its amount for abuse of discretion.” Little Rock Sch. Dist. v. Arkansas, 127 F.3d 693, 697 (8th Cir. 1997). The main justification the ABB fiduciaries give for reducing the fee award is that it was based on hourly rates the district court drew from another large ERISA case involving the same plaintiffs’ law firm, Abbot v. Lockheed Martin Corp., No. 06-cv-701, 2015 WL 4398475 (S.D. Ill. July 17, 2015), which the ABB fiduciaries
We dismiss as moot the participants’ motion to strike a portion of the ABB fiduciaries’ reply brief, because the challenged passages concern an issue we do not reach. -14-
say were too high. They emphasize that the fee request in the other case was unopposed—the case had settled, the award was to come out of the settlement pot, and there was no meaningful objection from the plaintiffs’ class—and the amount was initially set as a percentage of the total recovery, with the hourly rates serving only as a check on the product of that calculation. See id. at *1-3. Wherever the rates originally came from, a federal judge, relying on the findings of a special master, reviewed them and specifically concluded they were “reasonable and consistent with market rates.” Id. at *3-4. The ABB fiduciaries do not counter the participants’ showing with any contrary evidence affirmatively suggesting the rates are excessive, consequently, we are satisfied the district court’s use of them was not an abuse of discretion. The ABB fiduciaries also emphasize the participants’ lawyers won on only one issue (out of three) before this court and accuse them of spending too much time on the appeal. The district court acknowledged both points and reduced the amount it calculated based on the hourly rates by about a third to account for them. We cannot say it was an abuse of discretion not to reduce it further. Finally, we agree with the ABB fiduciaries that we can review and correct the incentive awards the district court ordered for the three named plaintiffs. The awards—$25,000 each—should be paid out of the class recovery, not by the ABB fiduciaries on top of the other amounts they owe. This might well be the result the district court intended—it is what the district court ordered in its original award, which its order on remand claimed to be simply “reaffirm[ing]” on this point. And it makes sense. Incentive awards compensate lead plaintiffs for their work and the benefit they have conveyed on the rest of the class. See, e.g., Koenig v. U.S. Bank Nat’l Ass’n, ND (In re US Bancorp Litig.), 291 F.3d 1035, 1038 (8th Cir. 2002). There is no sound reason to make defendants pay for these awards.
CONCLUSION The district court did not err in deciding that the ABB fiduciaries abused their discretion and breached their fiduciary duties by acting on improper motives when they replaced the Wellington Fund with the Freedom Funds as investment options for the plans. But the district court should have decided for itself how to measure what the plans lost as a result, rather than considering itself bound by our prior comments on the issue. That question is still for the district court to answer in the first instance, so the judgment in favor of the ABB fiduciaries is at best premature. We therefore vacate the judgment, vacate the award of attorney fees and costs, and remand the case for further proceedings. ______________________________