Plaintiffs John Brotherston and Joan Glancy are two former employees of Putnam Investments, LLC who participated in Putnam's defined-contribution 401(k) retirement plan (the "Plan"). They brought this lawsuit on behalf of a now-certified class of other participants in the Plan, and on behalf of the Plan itself pursuant to the civil enforcement provision of the Employee *23 Retirement Income Security Act ("ERISA"). See 29 U.S.C. § 1132 (a)(2). They claim that Putnam (as well as other Plan fiduciaries) breached fiduciary duties owed to Plan participants by offering participants a range of mutual fund investments that included all of (and, for most of the class period, only) Putnam's own mutual funds without regard to whether such funds were prudent investment options. They also claim that Putnam structured fees and rebates in a manner that was both unreasonable and treated Plan participants worse than other investors in those Putnam mutual funds. In a series of rulings before and after plaintiffs presented their evidence at trial, the district court found that plaintiffs failed to prove that any lack of care in selecting the Plan's investment options resulted in a loss to the Plan, and that the manner in which Putnam transacted with the Plan was neither unreasonable nor less advantageous than the manner in which Putnam dealt with other investors in its mutual funds. Finding several errors of law in the district court's rulings, we vacate the district court's judgment in part and remand for further proceedings.

I.

We begin with a basic outline of the undisputed facts and the procedural history of this case, reserving further details for our analysis. 1 Putnam is an asset management company that creates, manages, and sells mutual funds. Under the Plan, eligible employees of Putnam and its subsidiaries make contributions to individual 401(k) accounts and personally direct those contributions among a menu of investment options. Putnam itself also contributes to the employees' Plan accounts. Pursuant to the Plan's governing documents, Putnam Benefits Investment Committee ("PBIC") is one of the Plan's named fiduciaries and is responsible for selecting, monitoring, and removing investments from the Plan's offerings.

The investment options offered under the Plan include many of Putnam's proprietary mutual funds. Between 2009 and 2015, over 85% of the Plan's assets were invested in these funds. Putnam offers two classes of shares in these funds: Y shares and R6 shares. 2 Most of Putnam's mutual funds offered under the Plan are "actively managed"; that is, they are operated by an investment advisor seeking to beat the market. From the beginning of the class period in November 2009 through January 31, 2016, the PBIC selected no mutual funds other than the propriety Putnam funds for inclusion in the portfolio of investment vehicles offered to Plan participants. During this period, Plan participants were given the option to invest in non-affiliated funds only through a self-directed brokerage account.

The Plan itself did instruct the PBIC to include as investment options "any publicly offered, open-end mutual fund (other than tax-exempt funds) that are generally made available to employer-sponsored retirement plans and underwritten or managed by Putnam Investments or one of its affiliates," as well as several other Putnam funds and a collective investment trust administered by Putnam's affiliate, PanAgora Asset Management, Inc. But the parties presume, at least for purposes of this case, *24 that this instruction does not immunize defendants from potential liability based on the duty of prudence in selecting investment offerings under the Plan. See Fifth Third Bancorp v. Dudenhoeffer , --- U.S. ----, 134 S.Ct. 2459 , 2468, 189 L.Ed.2d 457 (2014) ("[T]he duty of prudence trumps the instructions of a plan document ....").

The district court found that the PBIC did not independently investigate Putnam funds before including them as investment options under the Plan, did not independently monitor them once in the Plan, 3 and did not remove a single fund from the Plan lineup for underperformance, even when certain Putnam funds received a "fail" rating from Advised Asset Group, a Putnam affiliate. 4

In November 2015, Brotherston and Glancy filed this lawsuit against Putnam, the PBIC, and various other Putnam individuals and entities (collectively, "defendants"). On behalf of themselves, two certified subclasses of other Plan participants, and the Plan itself pursuant to 29 U.S.C. § 1132 (a)(2) (collectively, "plaintiffs"), they press two types of claims under ERISA. First, they claim that the fees charged by Putnam subsidiaries to the mutual funds offered in the Plan constituted prohibited transactions under ERISA. Second, they claim that Putnam, through its committees operating the Plan, breached its fiduciary duties by blindly stocking the Plan with Putnam-affiliated investment options merely because they were proprietary. 5 Three months after this lawsuit commenced, the PBIC added six BNY Mellon collective investment trusts to the Plan's investment options. It is undisputed that the process for choosing the BNY Mellon funds was prudent.

By agreement of the parties, the district court decided plaintiffs' prohibited transactions claims on a case-stated basis at summary judgment. After seven days of a bench trial, during which plaintiffs but not defendants presented their case, the district court entered judgment on partial findings under Federal Rule of Civil Procedure 52(c). On all counts, the court found against plaintiffs, who now appeal.

II.

We begin our analysis with the order that dismissed plaintiffs' prohibited transactions claims. The case-stated procedure allows a court in a nonjury case to "engage in a certain amount of factfinding, including the drawing of inferences," where "the basic dispute between the parties concerns only the factual inferences that one might draw from the more basic facts to which the parties have agreed." Pac. Indem. Co. v. Deming , 828 F.3d 19 , 22 (1st Cir. 2016) (quoting United Paperworkers Int'l Union Local 14 v. Int'l Paper Co. , 64 F.3d 28 , 31-32 (1st Cir. 1995) ). In reviewing the entry of summary judgment *25 on a case-stated record, we review legal questions de novo and factual determinations for clear error. See United Paperworkers Int'l , 64 F.3d at 31-32 .

A brief sketch of the statutory background frames our analysis. ERISA "supplements the fiduciary's general duty of loyalty to the plan's beneficiaries by categorically barring certain transactions deemed 'likely to injure the pension plan.' " Harris Tr. and Sav. Bank v. Salomon Smith Barney, Inc. , 530 U.S. 238 , 241-42, 120 S.Ct. 2180 , 147 L.Ed.2d 187 (2000) (citation omitted) (quoting Comm'r v. Keystone Consol. Indus., Inc. , 508 U.S. 152 , 160, 113 S.Ct. 2006 , 124 L.Ed.2d 71 (1993) ). Two particular prohibitions, and their related exemptions, are at issue here. 6 The first prohibition appears in section 1106(a)(1), which states:

Except as provided in section 1108 of this title:

(1) A fiduciary with respect to a plan shall not cause the plan to engage in a transaction, if he knows or should know that such transaction constitutes a direct or indirect-- ...

(C) furnishing of goods, services, or facilities between the plan and a party in interest ....

29 U.S.C. § 1106 (a)(1)(C). The second prohibition appears in section 1106(b), which provides:

A fiduciary with respect to a plan shall not-- ...

(3) receive any consideration for his own personal account from any party dealing with such plan in connection with a transaction involving the assets of the plan.

29 U.S.C. § 1106 (b).

The design and operation of the Plan implicates both of these prohibitions. The Plan contracts with parties-in-interest (Putnam subsidiaries) for services, thereby implicating section 1106(a)(1). 7 And Putnam, through the service fees it charges the Putnam funds in which the Plan invests, receives a benefit "in connection with a transaction involving the assets of the [P]lan" (that transaction being the Plan's purchase of shares in the Putnam funds), thereby implicating section 1106(b). Putnam therefore runs afoul of each prohibition unless it qualifies for an applicable exemption. Defendants argue that several such exemptions apply. We address each in turn, beginning with those potentially applicable to the otherwise broad reach of the prohibition imposed by section 1106(a)(1) for causing a plan to purchase services from a party-in-interest.

A.

By its very terms, the prohibition of section 1106(a)(1) on transactions with parties-in-interest applies "[e]xcept as provided in section 1108." Section 1108 in turn provides two exemptions upon which defendants rely. The first exemption allows for:

Contracting or making reasonable arrangements with a party in interest for office space, or legal, accounting, or other services necessary for the establishment or operation of the plan, if no more *26 than reasonable compensation is paid therefor.

29 U.S.C. § 1108 (b)(2) (emphasis added). The second exemption provides that a fiduciary shall not be barred from:

receiving any reasonable compensation for services rendered, or for the reimbursement of expenses properly and actually incurred, in the performance of his duties with the plan; except that no person so serving who already receives full time pay from an employer or an association of employers, whose employees are participants in the plan, or from an employee organization whose members are participants in such plan shall receive compensation from such plan, except for reimbursement of expenses properly and actually incurred.

Id. § 1108(c)(2) (emphasis added).

Relevant here, Putnam mutual funds pay a monthly management fee to Putnam Investment Management, LLC ("Putnam Management") for investment management services and a monthly investor servicing fee to Putnam Investor Services, Inc. ("Putnam Services") for transfer agent services. Both Putnam Management and Putnam Services operate as part of Putnam and their profits flow directly to the parent company. So in the context of this case, the applicability of the two exemptions set forth in sections 1108(b)(2) and 1108(c)(2) hinges in the first instance on the answer to a common question: Were the payments received by these Putnam subsidiaries for their services to Putnam mutual funds reasonable?

The district court made several findings on this question based on the case-stated record. First, it found that the net expense ratios for the funds in which the Plan invested ranged from 0% to 1.65% as of December 2011, and that there was no evidence that the range was materially different for the relevant class period. Brotherston v. Putnam Invs., LLC , 15-cv-13825-WGY, 2017 WL 1196648 , at *6 (D. Mass. Mar. 30, 2017) Relatedly, the district court noted that other courts have upheld similar ranges. Id. Second, the court observed that, "[i]mportantly, all of the Putnam mutual funds the Plan invested in were also offered to investors in the general public, therefore, their expense ratios were 'set against the backdrop of market competition.' " Id. (quoting Hecker v. Deere & Co. , 556 F.3d 575 , 586 (7th Cir. 2009) ). Finally, the court rejected the analysis of plaintiffs' expert, Dr. Steve Pomerantz, who purported to show that Putnam's fees were materially higher on average than the fees paid by other funds, on the grounds that his comparators were flawed. Id. at *7.

In context, we read the district court's second finding as saying that the Putnam funds were both offered to and acquired by at least some other individuals and entities who had the freedom to invest in other funds in the marketplace. Such was precisely what defendants' expert, Dr. Erik Sirri, said in one of his reports. 8 Sirri's supplemental report stated that, in contrast to the conclusion drawn by plaintiffs' expert, the data "do not indicate that Putnam's funds have generally been rejected by impartial, unaffiliated fiduciaries of non-Putnam retirement plans." Rather, the report noted, "all but nine of the funds were offered by at least one other plan and several funds were offered by over one hundred different plans. Two-thirds of the funds were offered by at least nine other plans, and half were offered by at least 23 *27 other plans." 9 In addition, Sirri concluded in his original report that the Plan paid about $500,000 less in expenses from 2009 to 2014 than it would have paid had it invested at the average expense ratio for peer group funds identified by independent analyst Lipper, Inc.

Plaintiffs' position, supported by Pomerantz's report, was that very few plans as large as the Plan invested in any of the Putnam funds. And, as we noted, Pomerantz put forward an analysis to the effect that Putnam charged more for its funds than did other funds the expert deemed comparable. Based on this testimony, perhaps the district court could have found the fees unreasonable even though other investors paid them. But our review of the district court's finding to the contrary is for clear error. See United Paperworkers Int'l , 64 F.3d at 31-32 ; see also Chao v. Hotel Oasis, Inc. , 493 F.3d 26 , 35 (1st Cir. 2007) (noting in another context that we review "factual findings related to good faith and reasonableness for clear error"). And on this record we see no clear error in that finding. Moreover, the fact that the district court did not explicitly frame its conclusion that Putnam charges reasonable management fees as what it plainly was -- a finding of fact -- does not preclude us from treating it as such.

Plaintiffs also complain that Putnam did not offer the Plan the same revenue sharing rebates it offered other plans. And they contend that Sirri's analysis failed to account for this fact. But plaintiffs do not develop this argument in connection with their section 1106(a)(1) claim, the exemption that calls for an analysis of precisely why a fee is not "reasonable." So, we will review the revenue sharing rebates only as part of the inquiry into "other dealings" relevant to the exemption from the prohibition of section 1106(b).

We therefore affirm the district court's determination that defendants are not liable under the prohibited transaction provision of section 1106(a)(1)(C).

B.

Next, we ask whether defendants are liable under section 1106(b) because Putnam received fees from the funds in which the Plan invested. To avoid liability under that provision, defendants seek to rely on a prohibited transaction exemption adopted by the Department of Labor. Known as PTE 77-3, the exemption renders the prohibition of section 1106 inapplicable to employee benefit plans that invest in in-house mutual funds, provided that four conditions are met. See 42 F.R. 18734 ; see also Krueger v. Ameriprise Fin., Inc. , No. 11-cv-02781-SRN/JSM, 2012 WL 5873825 , at *14 (D. Minn. Nov. 20, 2012). Plaintiffs challenge only the satisfaction of one of these conditions. We therefore limit our analysis to that condition, which reads as follows:

[a]ll other dealings between the plan and the investment company, the investment adviser or principal underwriter for the investment company, or any affiliated person of such investment adviser or principal underwriter, are on a basis no less favorable to the plan than such dealings are with other shareholders of the investment company.

42 F.R. 18734, 18735. So the question for the district court was: Are "[a]ll other dealings" between the Plan and Putnam any less favorable to the Plan than dealings *28 between Putnam and other shareholders investing in the same Putnam funds?

The dealings upon which the parties focus are payments of service fees and revenue sharing that Putnam provides for the benefit of plans that invest in its funds. When a third-party plan ( i.e. , a plan other than the Putnam Plan) invests in Y shares of a typical Putnam mutual fund, the third-party plan pays fees to a company that provides certain services to the plan, such as recordkeeping. In many instances, the manager of the Putnam mutual fund in which the plan invests pays the recordkeeper a share of the fund's revenue to reimburse the recordkeeper for services the manager would otherwise have to provide or pay for. The recordkeeper in turn may credit this payment to the plan. And sometimes the investment manager provides the revenue sharing directly to the plan.

With the Putnam Plan, the arrangement differs. Putnam itself directly pays the recordkeeper for the Plan, the recordkeeper does not charge any fees to the Plan, and Putnam's investment managers pay no revenue sharing to or for the benefit of the Plan, even in relation to Y shares of Putnam mutual funds.

Plaintiffs claim that this alternative arrangement operated to the Plan's disadvantage because it resulted in Plan participants paying higher expenses compared to third-party plan participants who benefitted from revenue-sharing rebates. This theory only works if the value of the revenue sharing that third-party plans receive exceeds the value of the service fees borne by those plans. Otherwise, third-party plans are simply being compensated for costs that the Plan never bears in the first place, which puts the Plan no worse off on net.

The district court did not find whether or to what extent the revenue sharing paid to or for the benefit of some third-party plans would have exceeded the fees borne by third-party plans but not by the Plan. Instead, at defendants' behest, the district court pointed to the fact that Putnam paid into the Plan (for the benefit of most participants) discretionary 401(k) employer contributions that totaled much more than the rebates would have. Pointing to the fact that PTE 77-3 calls for an assessment of "[a]ll other dealings between the plan and the investment company," the district court reasoned that, on a net basis, Putnam treated its Plan even more favorably than it treated those that received the benefit of revenue-sharing payments.

We do not agree with this analysis because we do not regard Putnam's payment of discretionary contributions to be a relevant "dealing" between Putnam and the Plan. As noted, PTE 77-3, which directs our focus to "all other dealings," is an exemption to section 1106(b), which otherwise prohibits "[a] fiduciary" from receiving payment or other consideration in connection with its own plan. As the Supreme Court has recognized, "the trustee under ERISA may wear many different hats." Pegram v. Herdrich , 530 U.S. 211 , 225, 120 S.Ct. 2143 , 147 L.Ed.2d 164 (2000). Putnam wore at least two hats: that of an employer dealing with its employees and that of a fiduciary dealing with the Plan. In making discretionary contributions, it acted as employer providing compensation to its employees, not as fiduciary. See ERISA Practice & Litigation § 3:32 ("In the single employer plan context, decisions relating to the timing and amount of contributions are generally not thought of as being fiduciary in nature."); cf. Akers v. Palmer , 71 F.3d 226 , 230 (6th Cir. 1995) (noting that "courts have no authority to decide which benefits employers must confer upon their employees") (quoting *29 Moore v. Reynolds Metals Co. Ret. Prog. , 740 F.2d 454 , 456 (6th Cir. 1984) ).

Putnam's own documents confirm that it understood this to be the law. Putnam's Fiduciary Planning Guide explains the basic contours of fiduciary responsibility. Under a heading labeled "A Fiduciary -- But Only for 'Fiduciary Functions,' " Putnam explains that various decisions, including determining "the level of benefits" for a retirement plan, are made in a party's "capacity as employer" and "are not subject to, and cannot be challenged, under ERISA's fiduciary rules."

In other words, the term "employer contribution" commonly used to describe the discretionary payments at issue here is no misnomer. Because Putnam's discretionary contributions were made in Putnam's capacity as employer for the benefit of its employees qua employees, they are irrelevant to the analysis under PTE 77-3, which, as we have noted, provides an exception to a prohibition on actions by fiduciaries. Putnam cannot point to those contributions to offset funds Putnam charges (or withholds from) the Plan in its capacity as a plan fiduciary. To hold otherwise would be to allow employers to claw back with their fiduciary hands compensation granted with their employer hands.

Taking an alternative tack, defendants contend that revenue sharing payments are not relevant to PTE 77-3(d) because they are paid to third-party service providers, rather than to the plans that own shares in the funds (the "shareholders" under PTE 77-3). The record supports defendants' assertion that revenue sharing payments are often paid directly to third-party service providers. However, defendants do not contest that these payments may well benefit the associated plans by offsetting payments the plans would otherwise make to those providers. Given this beneficial link, these payments fall within PTE 77-3's instruction to consider dealings between the "investment company" (Putnam) and "other shareholders" (third-party plans). Cf. NLRB v. Cabot Carbon Co. , 360 U.S. 203 , 211, 79 S.Ct. 1015 , 3 L.Ed.2d 1175 (1959) (construing "dealing with" as a "broad term").

Defendants' final rejoinder is that, for the Plan alone, Putnam pays recordkeeping fees upfront, rather than passing those costs along to the Plan. But, as we have already noted, this assertion does not definitively answer whether the Plan is treated less favorably than other shareholders. It is undisputed that the Plan's recordkeeping expenses that Putnam pays upfront are 3 basis points (0.03% of plan assets). It is also undisputed that Putnam pays revenue sharing of up to 25 basis points (0.25% of fund assets) in connection with Y shares of Putnam mutual funds held by other plans. Given the gap between these two figures, the Plan may in fact be missing out on net revenue sharing benefits being recouped by other plans. Pomerantz asserted in his report that this is precisely what has happened. According to his calculations, which he adjusted to present value, the Plan lost out on over $5 million from 2010 to 2016 as a result of the Plan's inability to capture revenue sharing payments. This analysis took into account the fact that Putnam paid recordkeeping fees and so-called "trustee fees."

Defendants assert that in addition to paying "[a]ll recordkeeping expenses," Putnam also pays "the cost of a service that provides individualized investment advice to participants," as well as the annual fee associated with the brokerage window that allows Plan participants to access non-Putnam investments. But defendants do not quantify this payment in their briefing. Nor do they address whether the figures for "total administrative fees" to which they stipulated below include the additional *30 cost of the window or the fees identified in other documents in the record.

Without guidance from the parties on how to analyze these various documents and without the benefit of the district court's assessment on the matter, we think it best not to sift through the record to reach our own unaided conclusions. We therefore vacate the judgment against plaintiffs on their claim under section 1106(b) and remand for the district court to reconsider whether the requirement of PTE 77-3(d) is satisfied in light of revenue sharing payments Putnam makes to some other plans. 10 In considering whether, by not receiving the benefit of such payments, the Plan was treated any less favorably on net than other comparably situated plans, the district court should consider, among other things, the administrative fees paid by Putnam, as well as any fees paid by the Plan itself. The district court should not consider the discretionary contributions made by Putnam to Plan participants.

III.

We turn now to the district court's ruling mid-trial dismissing plaintiffs' claims that Putnam acted imprudently in selecting the Plan's investment options and that it breached the duty of loyalty by engaging in self-dealing. "If a party has been fully heard on an issue during a nonjury trial and the court finds against the party on that issue," Rule 52(c) allows the court to "enter judgment against the party on a claim or defense that, under the controlling law, can be maintained or defeated only with a favorable finding on that issue." Fed. R. Civ. P. 52(c) ; see also Morales Feliciano v. Rullán , 378 F.3d 42 , 59 (1st Cir. 2004). In resolving a Rule 52(c) motion, "the court's task is to weigh the evidence, resolve any conflicts in it, and decide for itself in which party's favor the preponderance of the evidence lies." 9C Charles Alan Wright & Arthur R. Miller, Federal Practice & Procedure § 2573.1 (3d ed.) (footnotes omitted). As with a case-stated summary judgment ruling, we review Rule 52(c) judgments under a mixed standard of review, "evaluat[ing] the district court's conclusions of law de novo and typically examin[ing] the district court's underlying findings of fact for clear error." Mullin v. Town of Fairhaven , 284 F.3d 31 , 36-37 (1st Cir. 2002) (internal quotation marks and citations omitted).

A.

We begin with the duty of prudence. Pursuant to ERISA, a fiduciary must act "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." 29 U.S.C. § 1104 (a)(1)(B). A fiduciary who breaches that duty must "make good" to the plan "any losses to the plan resulting from such breach." Id. § 1109(a). Although the parties in this case dispute the precise requirements for making out a duty of prudence claim, both sides agree that the claim has three elements: breach, loss, and causation. We address each in turn.

1.

The district court fairly summarized the plaintiffs' theory of breach: "[T]he Defendants violated their fiduciary duty of prudence by failing to implement or follow a prudent objective process for investigating *31 and monitoring the individual merits of each of the Plan's investments in terms of costs, redundancy, or performance." Brotherston v. Putnam Invs., LLC , No. 15-cv-13825-WGY, 2017 WL 2634361 , at *8 (D. Mass. June 19, 2017). Because the district court terminated the trial before Putnam could present its defense, the district court did not make a definitive ruling on whether such a violation occurred. Rather, it concluded that the evidence presented would be sufficient to support a finding that the PBIC "failed to monitor the Plan investments independently" and that it therefore failed to discharge its fiduciary duty. Id. at *9. Presumably because of the tentative nature of the district court's conclusion, Putnam lodges no cross-appeal from that determination, so we accept it at face value and move on to the question of loss.

2.

The question of loss in this case might at first blush seem quite simple. If one invests $1,000 in shares of a mutual fund, and two years later the shares are worth $1,000, many people would say that there has been no loss. Certainly the IRS agrees. And if the investment increases in absolute dollar value, rather than remaining constant, many would similarly claim no loss.

Any reasonably sophisticated investor, though, would think about loss -- and gain -- very differently. To the extent that the investor had a choice of investments, the decision to pick one investment over another might result in a measurable loss of opportunity. It follows that a trustee who decides to stuff cash in a mattress cannot assure that there is no loss merely by holding onto the mattress. This more sophisticated view of loss aligns with most people's expectations regarding their financial fiduciaries who have broad investment discretion. It also aligns with what has become known as the "total return" measure of loss and damages for breach of trust. See Restatement (Third) of Trusts, § 100 cmt. a(3); see also id. § 100 cmt. b(1).

The Restatement calls "for determining whether and in what amount the breach has caused a 'loss[ ]' ... by reference to what the results 'would have been if the portion of the trust affected by the breach had been properly administered .' " Id. ch. 19, intro. note (emphasis in original) (quoting Id. § 100 ). The Restatement expands on this principle as follows: The recovery from a trustee for imprudent or otherwise improper investments is ordinarily "the difference between (1) the value of those investments and their income and other product at the time of surcharge and (2) the amount of funds expended in making the improper investments, increased (or decreased) by a projected amount of total return (or negative total return) that would have accrued to the trust and its beneficiaries if the funds had been properly invested." Id. § 100 cmt. b(1). Finally, the Restatement specifically identifies as an appropriate comparator for loss calculation purposes "return rates of one or more ... suitable index mutual funds or market indexes (with such adjustments as may be appropriate)." Id.

ERISA itself is not so specific. Rather, it states that a breaching fiduciary shall be liable to the plan for "any losses to the plan resulting from each such breach." 29 U.S.C. § 1109 (a). Certainly this text is broad enough to accommodate the total return principle recognized in the Restatement. Behind the text, too, stands Congress's clear intent "to provide the courts with broad remedies for redressing the interests of participants and beneficiaries when they have been adversely affected by breaches of fiduciary duty."

*32 Eaves v. Penn , 587 F.2d 453 , 462 (10th Cir. 1978) (relying on S. Rep. No. 93-127 ). And as the Supreme Court has instructed, when we confront a lack of explicit direction in the text of ERISA, we often find answers in the common law of trusts. See Varity Corp. v. Howe , 516 U.S. 489 , 496-97, 116 S.Ct. 1065 , 134 L.Ed.2d 130 (1996) ; see also id. at 502, 506-07 , 116 S.Ct. 1065 (relying on "ordinary trust law principles" to fill gaps created by ERISA's lack of definition regarding the scope of fiduciary conduct and duties).

In this instance, the trust law that we have described provides an answer that both requires no stretch of ERISA's text and accords with common sense. Otherwise, hoarding plan assets in cash would become a fail-safe option for ERISA fiduciaries. We therefore hold that an ERISA trustee that imprudently performs its discretionary investment decisions, including the design of a portfolio of funds to offer as investment options in a defined-contribution plan, "is chargeable with ... the amount required to restore the values of the trust estate and trust distributions to what they would have been if the portion of the trust affected by the breach had been properly administered." Restatement (Third) of Trusts, § 100.

Applying this definition of chargeable loss to the case at hand, we begin with the district court's tentative finding that PBIC breached its fiduciary duty in automatically including Putnam funds as investment options for the Plan and then failing to independently monitor the performance of those funds. The district court correctly observed that such a breach does not mean that the Plan necessarily suffered any loss. So the question was, did any loss occur?

Plaintiffs attempted to answer this question with the testimony of their expert, Pomerantz. As we have noted, most of the Putnam funds were actively managed and therefore carried higher fees than passively-managed funds. For each Putnam fund held by the Plan, Pomerantz asked whether the Plan got something for those higher fees. Pomerantz began by comparing one at a time the total return for each Putnam fund to the total return for two passive comparators, a Vanguard index fund that belonged to the same Morningstar category 11 as the Putnam fund and a BNY Mellon collective investment trust, for every quarter from the beginning of the class period through mid-2016, and then adding together each quarterly differential. Pomerantz also did a second analysis with the same comparators, focusing on the fees charged by the Putnam fund compared to the comparator fund, to be able to pinpoint what portion of the difference in total returns stemmed from the fee differential. Where an automatically-included Putnam fund generated returns equal to or greater than its benchmark, Pomerantz calculated no loss for that fund, and credited any differential gain to Putnam. But where an automatically-included Putnam fund generated lower returns than its benchmark, he deemed the differential to be a loss. Pomerantz testified that overall, the portfolio of actively managed Putnam funds, when compared to a portfolio of passively managed Vanguard funds, suffered total damages (converted to present value) of about $45.6 million. Most of this figure, about $31.7 million, was attributable to the difference in fees between the two sets of funds. When compared to a portfolio of BNY Mellon funds, the Putnam portfolio suffered total damages of about $44.3 million, of which about $35.1 was fee damage.

*33 In short, according to Pomerantz's testimony, the Plan and its beneficiaries paid a premium of $30 to $35 million to obtain overall net returns that fell below the returns generated by the passive investment options that the PBIC could have offered.

The district court ruled, as a matter of law, this evidence was insufficient to make out a prima facie case of loss. It is not clear why the district court so concluded. The court stated at one point that proof that Putnam lacked a prudent process to monitor Plan investment vehicles did not make "the entire Plan lineup imprudent." Brotherston , 2017 WL 2634361 , at *12. It further stated that "a person could lack an independent process to monitor his investment and still end up with prudent investments, even if it was the result of sheer luck." Id. In so stating, the district court appeared concerned that approving what it characterized as the "broad sweep of the Plaintiffs' 'procedural breach' theory," id. at *10, would implicitly decide, without proof on the matter, that every fund in the Plan's portfolio was "per se imprudent," id. at *12, in the sense of being substantively an unwise investment. But nothing in Pomerantz's methodology so presumed. Rather, he simply calculated which funds generated a loss relative to a benchmark.

Of course, the court's concern regarding holding defendants liable for losses stemming from funds that may in fact be good investment options even if selected without due care is legitimate; ERISA defendants are not liable for damages that the Plan would have suffered even with a prudent fiduciary at the helm. See 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."). But this is an issue of causation (and possibly damage calculation), not loss. See id. (framing the question of whether a fiduciary's decision was objectively reasonable as part of ERISA's causation requirement). And for the reasons we explain in the following section, the burden of showing that a loss would have occurred even had the fiduciary acted prudently falls on the imprudent fiduciary. By allowing its analysis on loss to be driven by its concern regarding the objective prudence of the Putnam funds, the district court in essence required plaintiffs to show causation as part of its case on loss -- even as it correctly sought to reserve that requirement to defendants. 12 Brotherston , 2017 WL 2634361 , at *9 n.15.

The district court's concern may also have been implicitly informed by a point it summarized in its statement of facts but did not revisit in its analysis: that Putnam included in the Plan's investment lineup so-called qualified default investment alternatives ("QDIAs"), also known as Putnam's Retirement Ready funds. As the district court pointed out, plaintiffs' "fiduciary process expert" at trial, Dr. Martin Schmidt, testified that the process for reviewing and monitoring these funds was prudent, although plaintiffs dispute on appeal the precise meaning of Schmidt's testimony. The presence of prudently managed Putnam funds in the Plan's investment menu suggests that a portion of Pomerantz's estimate of total portfolio-wide loss may be subject to challenge for that reason, among others. 13 It *34 does not, however, establish that Pomerantz's approach was across-the-board inadequate as a matter of law.

The point remains: With the exception of the QDIAs, the entire portfolio of investment options (through January 31, 2016) was selected by the use of imprudent means, or so the district court itself conditionally found. So to determine whether there was a loss, it is reasonable to compare the actual returns on that portfolio to the returns that would have been generated by a portfolio of benchmark funds or indexes comparable but for the fact that they do not claim to be able to pick winners and losers, or charge for doing so. Restatement (Third) of Trusts, § 100 cmt. b(1) (loss determinations can be based on returns of suitable index mutual funds or market indexes); cf. Evans v. Akers , 534 F.3d 65 , 74 (1st Cir. 2008) ("Losses to a plan from breaches of the duty of prudence may be ascertained, with the help of expert analysis, by comparing the performance of the imprudent investments with the performance of a prudently invested portfolio."). This is what Pomerantz purported to do.

This is not to say that Pomerantz necessarily picked suitable benchmarks, or calculated the returns correctly, or focused on the correct time period. Putnam raises some of these issues on appeal, arguing that Pomerantz's comparators were not plausible and that he improperly focused on damages at a particular point in time. But these are questions of fact. 14 And the district court never reached these questions precisely because it concluded that Pomerantz's approach to establishing that the investment funds selected by Putnam incurred losses was insufficient as a matter of law. Correctly recognizing that its resolution of that issue was not clear cut, the district court explicitly invited de novo review on the question of legal sufficiency, which we have now provided by determining that plaintiffs' evidence was sufficient to support a finding of loss.

3.

We now turn to the question of causation. Assuming the Plan suffered a loss, the district court was certainly correct that the lack of prudence in the procedures used to select investments may not have caused the loss. See Plasterers' Local Union No. 96 Pension Plan , 663 F.3d at 218 (4th Cir. 2011) ("[T]he mere fact that the [fiduciaries] failed to investigate alternative investment options does not mean that their actual investments were necessarily imprudent ones."). A prudent investor may have selected fee-burdened funds, perhaps even Putnam's specific funds, that over the relevant years performed worse than market index funds for reasons that would have been reasonably unforeseeable to or discounted by the prudent investor. Since ERISA only allows for the recovery of loss "resulting from" the fiduciary's breach, a beneficiary is not eligible to recover damages in that situation. 29 U.S.C. § 1109 (a). All of this means that a court need find causation before awarding damages. See Roth , 16 F.3d at 919 ; see also Brock v. Robbins , 830 F.2d 640 , 647 (7th Cir. 1987) (rejecting the idea that, in enacting ERISA, Congress intended to deter "imprudent but harmless conduct").

So far, so good, in that all parties agree that causation must be found to sustain a *35 recovery for plaintiffs. What the parties dispute is who bears the burden of proving (or disproving) causation. To answer this question, we begin with the extant precedent, followed by our own analysis.

Our sister courts are split on who bears the burden of proving or disproving causation once a plaintiff has proven a loss in the wake of an imprudent investment decision. Compare Tatum v. RJR Pension Inv. Comm. , 761 F.3d 346 , 363 (4th Cir. 2014) (adopting in the ERISA context the "long recognized trust law principle ... that once a fiduciary is shown to have breached his fiduciary duty and a loss is established, he bears the burden of proof on loss causation"); McDonald v. Provident Indem. Life Ins. Co., 60 F.3d 234 , 237 (5th Cir. 1995) (holding that once an ERISA plaintiff proves "a breach of a fiduciary duty and a prima facie case of loss to the plan[,] ... the burden of persuasion shifts to the fiduciary" to disprove causation (internal quotation marks omitted) ); Martin v. Feilen , 965 F.2d 660 , 671 (8th Cir. 1992) ("[O]nce the ERISA plaintiff has proved a breach of fiduciary duty and a prima facie case of loss to the plan or ill-gotten profit to the fiduciary, the burden of persuasion shifts to the fiduciary to prove that the loss was not caused by, or his profit was not attributable to, the breach of duty.") with Pioneer Centres Holding Co. Emp. Stock Ownership Plan & Tr. v. Alerus Fin., N.A. , 858 F.3d 1324 , 1337 (10th Cir. 2017), cert. dismissed per stipulation , No. 17-667, --- U.S. ----, --- S.Ct. ----, --- L.Ed.2d ----, 2018 WL 4496523 (U.S. Sept. 20, 2018) (adopting the ordinary default rule to hold that "the burden falls squarely on the plaintiff asserting a breach of fiduciary duty claim under § 1109(a) of ERISA to prove losses to the plan 'resulting from' the alleged breach of fiduciary duty"); Saumer v. Cliffs Natural Resources Inc. , 853 F.3d 855 , 863 (6th Cir. 2017) ("[A] plaintiff must show a causal link between the failure to investigate and the harm suffered by the plan." (internal quotation marks omitted) ); Wright v. Oregon Metallurgical Corp. , 360 F.3d 1090 , 1099 (9th Cir. 2004) (same); Willett v. Blue Cross & Blue Shield of Ala. , 953 F.2d 1335 , 1343 (11th Cir. 1992) (instructing that "[o]n remand, the burden of proof on the issue of causation will rest on the beneficiaries"). 15

We join those circuits that approve a burden-shifting approach. Our reasoning begins with the language of the statute. As we have already noted, that language -- establishing that a breaching fiduciary shall be liable for any losses to the plan "resulting from" its breach, 29 U.S.C. § 1109 (a) -- clearly requires a causal connection between a breach and a loss in order to justify compensation for the loss. Like many statutes that provide a cause of action, section 1109(a) does not explicitly state whether the plaintiff bears the burden of proving that causal link or whether the defendant must prove the absence of causation. Two interpretative approaches offer potential for resolving that question in the face of the text's silence.

First, there is what the Supreme Court has called the "ordinary default rule." Schaffer ex rel. Schaffer v. Weast , 546 U.S. 49 , 56, 126 S.Ct. 528 , 163 L.Ed.2d 387 (2005). Under this rule, courts ordinarily presume that the burden rests on plaintiffs "regarding the essential aspects of their claims." Id. at 57 , 126 S.Ct. 528 . That normal rule, however, "admits of exceptions." Id. For example, "[t]he ordinary *36 rule, based on considerations of fairness, does not place the burden upon a litigant of establishing facts peculiarly within the knowledge of his adversary," id. at 60 , 126 S.Ct. 528 (alteration in original) (quoting United States v. New York, N.H. & H.R. Co. , 355 U.S. 253 , 256 n.5, 78 S.Ct. 212 , 2 L.Ed.2d 247 (1957) ), although there exist qualifications on the application of this exception. Id.

Second, ERISA brings to bear its own interpretative guidance. As we have already pointed out, supra , and will explain in greater detail, when the Supreme Court confronts a lack of explicit direction in the text of ERISA, it regularly seeks an answer in the common law of trusts. See generally Varity Corp. , 516 U.S. at 496-97 , 116 S.Ct. 1065 ; see also id. at 502, 506-07 , 116 S.Ct. 1065 . The common law of trusts -- like ERISA -- classifies causation as an element of a claim for breach of fiduciary duty. See Restatement (Third) of Trusts, § 100 cmt. e. It also places the burden of disproving causation on the fiduciary once the beneficiary has established that there is a loss associated with the fiduciary's breach. Id. cmt. f. This burden allocation has long been the rule in trust law. See Tatum , 761 F.3d at 363 (describing it as a "long-recognized trust law principle").

So how much weight should we place on ERISA's borrowing of trust law in the face of Schaffer 's default rule? In answering this question, we are guided by three observations: that ERISA's borrowing of trust law principles is robust; that trust law's burden allocation best fits the balance ERISA seeks to achieve between the interests of fiduciaries and beneficiaries; and that in this case, borrowing trust law's burden allocation actually poses no conflict with Schaffer 's approach to burden allocation. We explain.

The Supreme Court has time and again adopted ordinary trust law principles to construe ERISA in the absence of explicit textual direction. In LaRue v. DeWolff, Boberg & Associates, Inc. , the Court confronted a demand to recover lost profits under one of ERISA's civil enforcement provisions, which makes no mention of lost profits. 552 U.S. 248 , 128 S.Ct. 1020 , 169 L.Ed.2d 847 (2008). It reasoned: "Under the common law of trusts, which informs our interpretation of ERISA's fiduciary duties, trustees are 'chargeable with ... any profit which would have accrued to the trust estate if there had been no breach of trust ....' " Id. at 253 , 128 S.Ct. 1020 n.4 (first alteration in original) (internal citation omitted) (quoting Restatement (Second) of Trusts, § 205 (1957) ). Confronting silence in the text on whether certain nonfiduciary parties in interest may be held accountable on a claim for equitable relief under ERISA § 502(a)(3), the Court in Harris Trust looked in part to the common law of trusts, which it found "plainly countenances the sort of relief sought." 530 U.S. at 250, 120 S.Ct. 2180 . And the Court relied on the experience of the common law to reject an argument that untoward effects might flow from allowing claims against nonfiduciaries. Id. at 251 , 120 S.Ct. 2180 . Most notably, in Firestone Tire & Rubber Co. v. Bruch , the Court mirrored ordinary trust law principles in construing the rules under ERISA that control the standard of review to be employed in reviewing denials of ERISA benefits. 489 U.S. 101 , 111, 109 S.Ct. 948 , 103 L.Ed.2d 80 (1989) ("In determining the appropriate standard of review ..., we are guided by principles of trust law."). As the Court noted, "ERISA abounds with the language and terminology of trust law." Id. at 110 , 109 S.Ct. 948 .

This is not to say that we automatically adopt ordinary trust law principles to fill in gaps in ERISA. Trust law provides no assistance when "it is inconsistent *37 with the language of the statute, its structure, or its purposes." Hughes Aircraft Co. v. Jacobson , 525 U.S. 432 , 447, 119 S.Ct. 755 , 142 L.Ed.2d 881 (1999) (internal quotation marks omitted). Here, though, the statutory language is silent, and Putnam points to nothing in ERISA's structure that conflicts with the allocation of burdens under ordinary trust law.

This brings us to our next consideration: the purposes Congress clearly sought to achieve with ERISA. In that vein, one of Putnam's amici argues that placing on the fiduciary the burden of disproving causation would be inconsistent with Congress's purpose of reducing the cost of litigation so as not to dissuade employers from establishing plans. There is no serious claim, though, that ordinary trust law does not incorporate a similar aim. More importantly, the Supreme Court has made clear that whatever the overall balance the common law might have struck between the protection of beneficiaries and the protection of fiduciaries, ERISA's adoption reflected "Congress'[s] desire to offer employees enhanced protection for their benefits." Varity Corp. , 516 U.S. at 497 , 116 S.Ct. 1065 (emphasis added); see also Mass. Mut. Life Ins. Co. v. Russell , 473 U.S. 134 , 156 n.17, 105 S.Ct. 3085 , 87 L.Ed.2d 96 (1985) (Brennan, J., concurring) ("[I]n enacting ERISA, Congress made more exacting the requirements of the common law of trusts relating to employee benefit trust funds." (emphasis in original) (internal quotation marks omitted) ); cf. Firestone , 489 U.S. at 114 , 109 S.Ct. 948 (rejecting an alternative standard of review on the grounds that it would "afford less protection to employees and their beneficiaries than they enjoyed before ERISA was enacted"). In other words, Congress sought to offer beneficiaries, not fiduciaries, more protection than they had at common law, albeit while still paying heed to the counterproductive effects of complexity and litigation risk. See Varity Corp. , 516 U.S. at 497 , 116 S.Ct. 1065 (noting the "competing congressional purposes" of protecting employees without "unduly discourag[ing] employers from offering welfare benefit plans in the first place"). And it still provided substantial cost and risk reduction to employers by establishing a uniform, federally preemptive regime with the prospect of uniform federal guidance and regulation by the Department of Labor.

ERISA's enhancement of the protections for beneficiaries that existed at common law is reflected by the Supreme Court's decisions in Central States, Southeast & Southwest Areas Pension Fund v. Central Transport, Inc. , 472 U.S. 559 , 105 S.Ct. 2833 , 86 L.Ed.2d 447 (1985) and Fifth Third Bancorp v. Dudenhoeffer , --- U.S. ----, 134 S.Ct. 2459 , 189 L.Ed.2d 457 (2014). Those are the clearest examples of the Court opting not to follow an applicable common law rule in applying ERISA. In both instances, the Court rejected the ordinary trust law rules in a manner that enhanced rather than reduced the protection of beneficiary interests to the arguable detriment of employers. Central States , 472 U.S. at 572 , 105 S.Ct. 2833 (holding ERISA fiduciaries to the "more specific trustee duties itemized in the Act"); Fifth Third Bancorp , 134 S.Ct. at 2469 (relying on Central States 's "holding that, by contrast to the rule at common law, trust documents cannot excuse trustees from their duties under ERISA" (internal quotation marks omitted) ). In short, when interpreting the application of ERISA in the absence of statutory guidance, the Supreme Court has usually opted for the common law approach except when rejection was necessary to provide enhanced beneficiary protections. But cf. Conkright v. Frommert , 559 U.S. 506 , 516, 130 S.Ct. 1640 , 176 L.Ed.2d 469 (2010)

*38 (adopting Varity 's guidance that "trust law does not tell the entire story" and extending the deference given to plan administrators' interpretation of plans on the grounds that it protects the interests of employers, in line with Congressional intent); Mertens v. Hewitt Associates , 508 U.S. 248 , 253-54, 113 S.Ct. 2063 , 124 L.Ed.2d 161 (1993) (suggesting in dicta that the common law trust rule allowing "knowing participation" liability to be imposed on both co-fiduciaries and third parties does not apply in the ERISA context). On such a record, it would be strange to reject trust law's rules on burden allocation in favor of an attempt to reduce employer costs, especially where the benefit of such a reduction would flow exclusively to employers whose breaches were followed by losses to the plan.

Finally, we work our way back to Schaffer . We began by presenting the two interpretative paths embodied in Schaffer and Varity . We could read these cases as establishing alternative rules of construction, one generally applicable and the other more specifically applicable to ERISA. Under such a reading, we would opt for Varity 's specific over Schaffer 's general. Or we might read Varity 's guidance as simply one of the exceptions to Schaffer 's ordinary, but not universally-applicable, default rule. Under both readings, we end up in the same place: applying trust law principles. We nevertheless prefer the latter approach in this case because one important reason behind the ordinary trust rule for allocating the burden of proof aligns so well with the exception to Schaffer 's default rule recognized in Schaffer itself. That exception recognizes that the burden may be allocated to the defendant when he possesses more knowledge relevant to the element at issue. Schaffer , 546 U.S. at 60 , 126 S.Ct. 528 . Trust law has long embodied similar logic. See Restatement (Third) of Trusts, § 100 cmt. f (noting that the general rule placing on the plaintiff the burden of proving his claim "is moderated in order to take account of ... the trustee's superior (often, unique) access to information about the trust and its activities"); cf. 1 Joseph Story, Commentaries on Equity Jurisprudence: As Administered in England and America , § 322 (1836) (noting that the trust beneficiary may "not have it in his power distinctly and clearly to show" that the trustee made a bargain advantageous to himself). In short, even if there were no freestanding expectation that the interpretation of ERISA would be informed by trust law generally, on the specific matter of allocating the burden of proving causation the ordinary trust law rule could stand on its own feet as an exception to the default rule that Schaffer itself recognizes.

Common sense strongly supports this conclusion in the modern economy within which ERISA was enacted. An ERISA fiduciary often -- as in this case -- has available many options from which to build a portfolio of investments available to beneficiaries. In such circumstances, it makes little sense to have the plaintiff hazard a guess as to what the fiduciary would have done had it not breached its duty in selecting investment vehicles, only to be told "guess again." It makes much more sense for the fiduciary to say what it claims it would have done and for the plaintiff to then respond to that.

It is also true that this common sense concern could be addressed by a mere shift in the burden of production rather than the burden of persuasion, and Schaffer applies only to the latter. 546 U.S. at 56 , 126 S.Ct. 528 . And because ERISA cases rarely involve jury instructions, it is likely that very few cases will actually leave the question of causation "in evidentiary *39 equipoise." Id. at 58 , 126 S.Ct. 528 . 16 So it would not be farfetched to chart a third route by defaulting to Schaffer 's ordinary rule on the burden of proof while nevertheless requiring the fiduciary to first put forward its view of what likely would have happened but for the alleged fiduciary breach. Neither party, though, has briefed such a middle ground. More importantly, we have many decades of experience with the allocation of the burden of proof called for routinely by trust law, with no evidence of any particular difficulties, unfairness, or costs in applying that rule in the few cases in which it actually makes a difference. Cf. Metropolitan Life Ins. Co. v. Glenn , 554 U.S. 105 , 113, 128 S.Ct. 2343 , 171 L.Ed.2d 299 (2008) ("[W]e note that trust law functions well with a similar standard."). We therefore opt for a well-trodden path rather than risk introducing unforeseeable complexities with a more novel approach.

For the foregoing reasons, we align ourselves with the Fourth, Fifth, and Eighth Circuits and hold that once an ERISA plaintiff has shown a breach of fiduciary duty and loss to the plan, the burden shifts to the fiduciary to prove that such loss was not caused by its breach, that is, to prove that the resulting investment decision was objectively prudent. See Tatum , 761 F.3d at 363 ; McDonald , 60 F.3d at 237 ; Martin , 965 F.2d at 671 . 17 In so ruling, we stress that nothing in our opinion places on ERISA fiduciaries any burdens or risks not faced routinely by financial fiduciaries. While Putnam warns of putative ERISA plans forgone for fear of litigation risk, it points to no evidence that employers in, for example, the Fourth, Fifth, and Eighth Circuits, are less likely to adopt ERISA plans. Moreover, any fiduciary of a plan such as the Plan in this case can easily insulate itself by selecting well-established, low-fee and diversified market index funds. And any fiduciary that decides it can find funds that beat the market will be immune to liability unless a district court finds it imprudent in its method of selecting such funds, and finds that a loss occurred as a result. In short, these are not matters concerning which ERISA fiduciaries need cry "wolf."

This holding, together with our conclusion that the district court erred in finding that plaintiffs failed as a matter of law to make even a prima facie showing of loss, requires vacatur of the district court's entry of judgment against plaintiffs on their prudence claim. We remand for the district court to complete the bench trial in order to definitively decide whether Putnam breached the duty of prudence and, if so, to decide whether plaintiffs have shown a loss to the Plan and, if so, to decide whether Putnam can meet its burden of showing that the loss most likely would have occurred even if Putnam had been prudent in its selection and monitoring procedures.

B.

Plaintiffs also argue that the district court erred in dismissing their claim for breach of the duty of loyalty. Under ERISA, fiduciaries "shall discharge their duties with respect to a plan 'solely in the *40 interest of the participants and beneficiaries,' that is, 'for the exclusive purpose of (i) providing benefits to participants and their beneficiaries; and (ii) defraying reasonable expenses of administering the plan.' " Pegram v. Herdrich , 530 U.S. 211 , 223-24, 120 S.Ct. 2143 , 147 L.Ed.2d 164 (2000) (citations omitted) (quoting 29 U.S.C. § 1104 (a)(1) ).

Plaintiffs' position is that Putnam failed to act in the best interests of Plan participants because it included Putnam funds by fiat, retained those funds even though they were underperforming, buried evidence that many of the funds were receiving failing grades, and failed to consider any alternative investment options from other companies. The district court reasoned that merely "identifying a potential conflict of interest alone is not sufficient to establish a breach of the duty of loyalty." Brotherston , 2017 WL 2634361 , at *3 ; see also id. at *8. Even pointing to self-dealing is not enough, reasoned the court, at least where the self-dealing (selecting proprietary funds for plan investments) is a common industry practice within the scope of an express exception. Id. at *3, *8. Rather, the district court found, to establish a claim for breach of the duty of loyalty plaintiffs were required to prove that defendant's motivation in taking these actions was to put its own interests ahead of those of Plan participants. Id. "Evaluating the totality of the circumstances," the district court also found that plaintiffs had failed to establish improper motivation. Id. at *8. It therefore dismissed plaintiffs' breach of loyalty claim. Id.

We review the district court's weighing of the evidence for clear error. See Mullin , 284 F.3d at 36-37 . Plaintiffs in turn offer four reasons for finding such error.

First, they argue that the district court incorrectly employed a balancing test to dismiss their loyalty claim by crediting Putnam for contributions it made as settlor. This argument misreads the district court's order, which plainly hinged its loyalty analysis on plaintiffs' failure to point to specific instances of disloyalty, rather than on Putnam's contributions as employer.

Second, plaintiffs argue that the district court erred in holding that a duty of loyalty claim requires a showing of improper motivation. Plaintiffs contend that "purported good intentions do not excuse disloyal actions." But to be loyal is to possess a certain state of mind, one "unswerving in allegiance." Merriam-Webster's Collegiate Dictionary 738 (11th ed. 2012) (definition of loyal); see also id. ("faithful in allegiance ..."). This is why, in reviewing ERISA duty of loyalty claims, we have asked whether the fiduciary's "operative motive was to further its own interests." Ellis v. Fid. Mgmt. Tr. Co., 883 F.3d 1 , 6 (1st Cir. 2018).

Third, plaintiffs claim that the district court treated the exceptions for prohibited transactions as "a safe harbor from breach of fiduciary duty claims." The district court did no such thing. Rather, the district court simply stated that plaintiffs did not carry their burden of persuasion merely by pointing to transactions that were expressly exempt from the prohibitions of sections 1106(a) and (b), particularly where such exempt transactions were common in the industry. And to the extent that Putnam engaged in a non-exempt prohibited transaction, it would be liable under section 1106 itself, which "supplements" the general duty of loyalty. Harris Trust , 530 U.S. at 241, 120 S.Ct. 2180 .

Finally, plaintiffs argue that, even if a breach of the duty of loyalty does require improper motivation, there is sufficient evidence that Putman's decisions were motivated by an intent to benefit itself. Even *41 assuming that to be so, the sufficiency of the evidence to prove plaintiffs' claim is not at issue on this appeal. Rather, the question before us is whether the evidence is so one-sided that we must deem the district court's fact finding as clear error. And since plaintiffs point to no action of Putnam that can be explained only by a disloyal motivation, the district court possessed ample discretion to find as it did.

C.

We discuss, finally, plaintiffs' claim for disgorgement. We have recognized, supra , that Putnam can be said to have received fees "in connection with a transaction involving the assets of the [P]lan," 29 U.S.C. § 1106 (b)(3). Such a receipt placed on Putnam the obligation to satisfy the requirements of PTE 77-3. And to the extent that Putnam fails on remand to qualify under that exemption, nothing in this opinion forecloses disgorgement as an available remedy. Plaintiffs, though, also seek to press a broader claim for disgorgement as part of their breach of fiduciary duty claim under 29 U.S.C. § 1109 (a), which requires a breaching fiduciary to "restore to [the] plan" any profits "made through use of assets of the plan." 18 The district court dismissed that claim as "legally insufficient" in view of its finding that plaintiffs had failed as a matter of law to show loss. Our ruling that plaintiffs' evidence may in fact be sufficient to establish a loss eliminates the district court's basis for dismissing plaintiffs' broader disgorgement claim, but we nevertheless affirm the dismissal of that claim on alternative grounds.

The object of plaintiffs' desired disgorgement is $27.9 million in fees (allegedly $37.3 million in present-day value) obtained by Putnam as a result of offering its proprietary funds as investment options to the Plan. The district court had independently ruled, as part of its earlier summary judgment decision, that those fees were not derived from Plan assets, and thus did not implicate the bar of 29 U.S.C. § 1106 (a)(1)(D) against any "use by or for the benefit of a party in interest, of any assets of the plan" or the bar of 29 U.S.C. § 1106 (b)(1) against a fiduciary "deal[ing] with the assets of the plan in his own interest or for his own account." Plaintiffs have expressly waived any challenge to that ruling. So defendants pointedly argue that plaintiffs are precluded from now claiming on appeal as part of their disgorgement claim that Putnam's fees were derived "through use of assets of the plan." 29 U.S.C. § 1109 (a).

Plaintiffs offer no argument at all for how the fees at issue could not have qualified as "use by or for the benefit of [Putnam] of any assets of the plan" under section 1106(a)(1)(D), or a "deal with the assets of the plan" under section 1106(b)(1), yet nevertheless be deemed to have been obtained by Putnam "through use of" Plan assets under § 1109(a). Plaintiffs have therefore waived any argument that the fees are subject to disgorgement under § 1109(a).

IV.

Regarding the district court's summary judgment ruling, we affirm the district court's dismissal of plaintiffs' prohibited *42 transaction claim under section 1106(a)(1)(C) ; we vacate the district court's dismissal of plaintiffs' prohibited transaction claim under section 1106(b)(3) ; and we remand for further proceedings. With respect to the district court's order entering judgment on partial findings, we affirm the dismissal of plaintiffs' breach of loyalty claim and the dismissal of their disgorgement claim, except to the extent that disgorgement may be a remedy for a prohibited transaction claim; we vacate the finding that plaintiffs have failed as a matter of law to show loss; and we remand for further consideration of plaintiffs' prudence claim in light of our holding on the burden-shifting issue. Costs are awarded to the plaintiffs.

None of this means, we add, that defendants have violated any duties or obligations owed to the Plan or its beneficiaries. Rather, it simply means that we have rejected two reasons for concluding that such a violation necessarily did not occur, and we have otherwise clarified for the district court several principles that should guide its subsequent rulings in this case.