WOOD, Circuit Judge.
This case presents a challenge to the practice known in the 401 (k) services industry as “revenue sharing” — an arrangement allowing mutual funds to share a portion of the fees that they collect from
investors with entities that provide services to the mutual funds, the investors, or both. Although the practice has been commonplace for years, until quite recently it was opaque to both individual investors and many 401(k) plan sponsors. As the existence and extent of revenue sharing has become more widely known, some have expressed concern that the practice unduly benefits mutual funds and 401(k) service providers to the detriment of plan participants. This concern has fueled a number of lawsuits alleging that the practice violates the Employee Retirement Income Security Act of 1974 (ERISA). This is one such suit.
The district court awarded summary judgment to the American United Life Insurance Company (AUL), which is an Indiana-based insurance company that offers investment, record-keeping, and other administrative services to 401(k) plans. The court ruled that AUL was not a fiduciary of the Leimkuehler, Inc. Profit Sharing Plan (the Plan) with respect to AUL’s revenue-sharing practices. The Plan and Robert Leimkuehler, its Trustee, have appealed. Although very little about the mutual fund industry or the management of 401(k) plans can plausibly be described as transparent, we agree with the district court that AUL is not acting as a fiduciary for purposes of 29 U.S.C. § 1002(21)(A) when it makes decisions about, or engages in, revenue sharing. We find it unnecessary to express any view on the question whether revenue sharing yields net benefits to individual 401(k) investors, and we thus affirm the district court.
I
Leimkuehler, Inc., a small company that manufactures prosthetic limbs and braces, operates a 401(k) plan for its employees. (So-called 401 (k) plans are, more formally, private, employer-based defined-contribution retirement plans that meet the requirements of Internal Revenue Code Section 401(k). 26 U.S.C. § 401(k). We commented on the importance to millions of people of this type of retirement plan in
Spano v. The Boeing Co.,
633 F.3d 574, 576 (7th Cir.2011).) Robert Leimkuehler, president of Leimkuehler, Inc., and Tras-tee of the Plan, brought this suit against AUL, which has provided services to the Plan since 2000.
One of the services AUL provides to the Plan is the use of á group variable annuity contract, which enables individual Plan participants to invest their 401(k) contributions “in” mutual funds. We use quotation marks because, as the contract is structured, no Plan participant invests in a mutual fund directly. Rather, participants’ contributions are deposited into a “separate account” — distinct because state insurance law and ERISA require AUL to keep retirement contributions separate from other assets — that AUL owns and controls. AUL uses the funds in the separate account to invest in whatever mutual funds the Plan participants have selected; it credits the proceeds of these investments back to the participants. Because the performance of the separate account mirrors that of the mutual funds, investing in the separate account is the equivalent from the perspective of a Plan participant of investing in the funds directly.
While the separate account means little to a Plan participant, it makes quite a difference to the mutual fund companies. If every individual participant in the Plan were to invest directly in the mutual funds that AUL services, the funds would have to keep track of and service thousands of individual accounts, many of which would contain little money. By pooling individual contributions into the separate account, AUL radically simplifies matters for the participating funds. From the funds’ per
spective, AUL is a single investor. The use of the separate account thus substantially reduces the mutual funds’ administrative, marketing, and service costs.
These costs do not, however, disappear altogether. Instead, AUL must perform many of the services that the mutual funds would otherwise handle themselves. Among other things, AUL keeps track of individual accounts, takes responsibility for calculating the daily value of assets in the separate account, distributes information to the Plan sponsor and participants, and provides a customer-service hotline.
In principle, AUL could cover the costs of providing these services in one of two ways. One way would be to bill the Plan sponsor or Plan participants directly. The other way would be to engage in a practice known as “revenue sharing,” whereby the mutual fund companies pay a portion of the fees they charge investors — fees that are referred to as a fund’s “expense ratio” and that are expressed as a percentage of a fund’s assets — to AUL. Because a portion of a mutual fund’s expense ratio is typically intended to cover the costs of providing the participant-level services that the mutual fund would be furnishing if it were not for AUL, the mutual funds are willing to pay some of these fees to AUL as compensation for AUL’s provision of these services.
One additional complication plays an important role in our case. Within a single mutual fund, there are often several different expense-ratio/revenue-sharing levels available, because most mutual funds offer multiple “share classes” to investors. Although each share class within a given fund is invested in an identical portfolio of securities, the classes have differing price structures. The share classes typically made available to 401(k) investors vary primarily (and possibly exclusively) in terms of expense ratio and revenue sharing (if any).
As a general matter, expense ratios and revenue-sharing payments move in tandem: the higher a given share class’s expense ratio, the more the fund pays AUL in revenue sharing. It is also generally the case that the more AUL receives in revenue sharing, the less it charges plan sponsors or participants directly for its services. AUL employees stated in deposition testimony (and Leimkuehler does not contest) that AUL offers a range of 401(k) investment products, some of which offer mutual funds with relatively high expense ratios and relatively low billed fees, and others with relatively low expense ratios and relatively high billed fees.
None of this is meant to suggest that there is necessarily a one-to-one correspondence between the cost to AUL of providing participant-level services and the amount that AUL receives in revenue-sharing payments. AUL may be making a profit, perhaps even a sizeable profit, from revenue sharing (just as it may be making a profit when it bills a plan directly for its services). The foregoing discussion simply places AUL’s revenue sharing in context. (We note as well that to the extent Leim-kuehler’s concerns about revenue sharing arise from AUL’s historical failure to disclose its revenue-sharing practices, that issue has been addressed recently. The Department of Labor (DOL) promulgated a final rule, effective July 1, 2012, that requires entities like AUL to disclose their revenue-sharing arrangements. 29 C.F.R. § 2550.408b-2 (2012).)
AUL’s contract with the Leimkuehler Plan did not enable Plan participants to invest in all of the roughly 7,500 mutual funds currently available on the market. Instead, Plan participants had a significantly narrower range of investment options for their 401 (k) contributions. The
winnowing of investment options occurred in two stages. At stage one, AUL selected a “menu” of mutual funds and presented this menu to Leimkuehler, in his capacity as Plan Trustee. As of 2010, this investment menu contained 383 funds. For each fund on the menu, AUL also selected a particular share class, and thus a particular expense ratio and level of revenue sharing. As counsel for Leimkuehler conceded at oral argument, share classes were selected at the time AUL developed the menu; they did not change thereafter. Although AUL did not disclose to Leimkueh-ler or to Plan participants which share class was associated with each fund on the menu, all parties agree that AUL did disclose each fund’s expense ratio. Leim-kuehler therefore knew how much each mutual fund cost, though he did not know how those costs were allocated between the fund companies and AUL.
At stage two, Leimkuehler selected from the menu the specific funds that he wished to make available to Plan participants. Plan participants could then direct their contributions to one or more of the investment options that Leimkuehler had selected. Under the contract, Leimkuehler retained the right to change his selections, and he in fact did make changes to the mix of available funds at least twice between 2000 and 2010. AUL also reserved the right to make substitutions to or deletions from Leimkuehler’s selections; it exercised this right twice — once in 2000 to substitute one S & P 500 index fund for another, and again in 2011 to substitute one Vanguard fund for another.
Leimkuehler filed this suit against AUL on behalf of the Leimkuehler Plan individually and as a class action, alleging that AUL’s revenue-sharing practices breached a fiduciary duty to the Plan under ERISA. The district court granted AUL’s motion for summary judgment, concluding that AUL did not owe any fiduciary responsibility to the Plan with respect to its revenue-sharing practices and that it therefore was not a “functional fiduciary” within the meaning of 29 U.S.C. § 1002(21)(A). In a separate ruling, the district court declined to grant AUL’s motion for either attorney’s fees and costs under ERISA, or costs under Federal Rule of Civil Procedure 54(d). Both parties now appeal: the Plan challenges the grant of summary judgment for AUL; and AUL challenges the rulings on fees and costs.
II
AUL is not named as a fiduciary to the Leimkuehler Plan. Accordingly, any fiduciary responsibility that AUL owes to the Plan must stem from its status as a “functional fiduciary.” The general term “fiduciary” is defined as follows:
[A] person is a fiduciary with respect to a plan to the extent (i) he exercises any discretionary authority or discretionary control respecting management of such plan or exercises any authority or control respecting management or disposition of its assets, (ii) he renders investment advice for a fee or other compensation, direct or indirect, with respect to any moneys or other property of such plan, or has any authority or responsibility to do so, or (iii) he has any discretionary authority or discretionary responsibility in the administration of such plan.
29 U.S.C. § 1002(21)(A).
Focusing on the second clause of sub-part (i), Leimkuehler offers two theories of how AUL satisfies its requirements. First, he asserts that AUL exercises authority or control over the management or disposition of the Plan’s assets by selecting which mutual fund share classes to include on its investment menu. Second, he asserts that AUL exercises authority or con
trol through the various activities associated with maintaining the separate account. DOL, which appeared in this appeal as
amicus curiae
on behalf of Leimkuehler, offers a third theory: it argues that AUL’s contractual reservation of the right to substitute or delete funds made available to the Leimkuehler Plan’s participants is itself an exercise of authority or control over the Plan’s assets, even if AUL never affirmatively exercises its contractual right in a way that gives rise to a claim. We address each argument in turn.
A
Leimkuehler’s first theory of AUL’s fiduciary status might broadly be termed a “product design” theory, as it centers on actions that AUL takes when designing the products it offers to its 401(k) plan customers. In crafting its menu of investment options, AUL decides which mutual funds to include and which share classes of those funds to select. In making both these decisions, AUL is also setting the stage for any revenue sharing in which it wishes to engage. These product-design decisions shape the disposition of Plan assets: they limit the universe of funds, as well as the share classes within those funds, in which Plan assets are invested. Leimkuehler urges that this suffices to make AUL a fiduciary under the terms of Section 1002(21)(A)(i).
The problem with this theory is that it is functionally indistinguishable from the one this court rejected in
Hecker v. Deere & Co.,
556 F.3d 575 (7th Cir.2009). In
Hecker,
participants in Deere & Company’s 401(k) plan sued two Fidelity entities that provided investment services to the plan.
Id.
at 578. Fidelity offered a menu of mutual funds and other investment options to Deere, which then selected which of those investment options it wished to make available to plan participants.
Id.
The menu of funds presented to Deere did not include every mutual fund on the market; rather, it was a select, though still expansive, list compiled by Fidelity in advance.
Id.
at 581, 583. The plaintiffs alleged that Fidelity’s control over which funds made it onto the list gave it authority or control over the plan’s assets, because the menu limited the universe of funds in which a plan participant could invest.
Id.
at 583. This court concluded that the act of selecting which funds will be included in a particular 401(k) investment product, without more, does not give rise to a fiduciary responsibility, both because there is “no authority that holds that limiting funds ... automatically creates discretionary control sufficient for fiduciary status,” and because, in any event, “the Trust Agreement gives Deere, not Fidelity Trust, the final say on which investment options will be included.”
Id.
Although
Hecker
did not address share classes specifically, its facts are otherwise strikingly similar to those in this case. Indeed, we perceive only two factual distinctions that might even conceivably be of any practical significance. First, AUL, unlike Fidelity, is an insurance company and therefore operates through the separate account. As discussed below in part II.B, however, this difference does not alter the result. Second, AUL reserves the right to make substitutions to the funds that Leimkuehler chooses to offer to Plan participants, and thus there is at least some basis for questioning whether Leim-kuehler has “the final say on which investment options will be included.” But Le-imkuehler concedes that AUL has never exercised this contractual right in a way that could give rise to a claim, and so on the present record this distinction falls away as well. (As mentioned above, AUL has exercised this right on two occasions; one instance fell outside the limitations
period, and the other involved a substitution of Vanguard funds, neither of which made revenue-sharing payments to AUL.)
Nor does adding the concept of share classes to the mix meaningfully differentiate this case from
Hecker.
We grant that the failure to offer every share class of every fund that AUL includes on its menu results in the limitation of the universe of investment options available to Plan participants. But we fail to see how this is significantly different from Fidelity’s limiting the universe of investment options by offering certain mutual funds and not others. True, some share classes are more expensive than others, but the cheapest option may not inevitably be the best option. There is also no particular reason to think that AUL would not seek to make up the revenue it missed by offering cheaper share classes by charging higher direct fees to plans like Leimkuehler’s. Furthermore, given that AUL does disclose the bottomline cost of every fund that it offers, Leimkuehler was free to seek a better deal with a different 401(k) service provider if he felt that AUL’s investment options were too expensive. In short, we see no basis for distinguishing AUL’s actions here from those in
Hecker.
We therefore confirm that, standing alone, the act of selecting both funds and their share classes for inclusion on a menu of investment options offered to 401(k) plan customers does not transform a provider of annuities into a functional fiduciary under Section 1002(21) (A) (i).
B
Leimkuehler argues, however, that AUL does more than merely select funds and share classes for its investment menu: AUL, he says, exercises control over the management and disposition of the Plan’s assets by maintaining the separate account, which AUL alone controls. In order to manage that account, AUL must keep track of individual Plan participants’ contributions and investment directions. It then must invest participants’ funds in the mutual funds they select and credit returns from the funds to the participants’ accounts. Although these tasks are essentially ministerial, Leimkuehler argues that they are nevertheless sufficient to make AUL a fiduciary, because Section 1002(21)(A)(i) requires only that AUL exercise
“any
authority or control respecting management or disposition of its assets.” (Emphasis added.)
The district court concluded that Leim-kuehler was reading too much into the word “any,” and that fiduciary status arises only if the authority or control permits the exercise of discretion. Although Section 1002(21)(A)(i) does not spell out such a limitation, the district court read several of this court’s prior decisions as holding that discretion is an essential prerequisite for finding a fiduciary duty under ERISA, rather than a characteristic that is often present but is not an ironclad requirement. We recognize that some imprecise language in our prior decisions in this area has generated confusion. See,
e.g., Hecker,
556 F.3d at 583 (“In order to find that they were ‘functional fiduciaries,’ we must look at whether either Fidelity Trust or Fidelity Research exercised discretionary authority or control over the management of the Plans, the disposition of the Plans’ assets, or the administration of the Plans.”);
Pohl v. National Benefits Consultants,
956 F.2d 126, 129 (7th Cir.1992) (“At all events, ERISA makes the existence of discretion a sine qua non of fiduciary duty.”). Our prior decisions tended to discuss Section 1002(21)(A) (which does make frequent mention of “discretion”) as a whole. That section, however, not only contains three subparts, but subpart (i) identifies two different situ
ations: “[first] [the person] exercises any
discretionary
authority or
discretionary
control respecting management of such plan or [second] [the person] exercises
any
authority or control respecting management or disposition of its assets.” 29 U.S.C. § 1002(21)(A)(i) (emphasis added). The concept of discretion is thus integral to Plan management, but it is conspicuously missing when it comes to asset management or disposition.
A number of our sister circuits have taken note of this distinction and concluded that discretionary control is not required with regard to the management or disposition of plan assets. See
Briscoe v. Fine,
444 F.3d 478, 492-94 (6th Cir.2006);
Chao v. Day,
436 F.3d 234, 237-38 (D.C.Cir.2006);
Coldesina v. Estate of Simper,
407 F.3d 1126, 1132 (10th Cir.2005);
Board of Trustees of Bricklayers & Allied Craftsmen Local 6 of N.J. Welfare Fund v. Wettlin Assoc., Inc.,
237 F.3d 270, 273 (3d Cir.2001);
Herman v. Nations-Bank Trust Co., (Georgia),
126 F.3d 1354, 1365 (11th Cir.1997);
LoPresti v. Terwilli-ger,
126 F.3d 34, 40 (2d Cir.1997);
IT Corp. v. General Am. Life Ins. Co.,
107 F.3d 1415, 1421 (9th Cir.1997);
FirsTier Bank, N.A. v. Zeller,
16 F.3d 907, 911 (8th Cir.1994). We agree with them that this reading is most faithful to the language of the statute, and we now make explicit that insofar as “management or disposition of assets” is concerned, there is no separate requirement of discretionary authority or control.
Unfortunately for Leimkuehler, however, this does not help him as much as he might think. Critically, Section 1002(21)(A) additionally states that an entity is a fiduciary only “to the extent” it exercises its authority or control. The Supreme Court has interpreted this phrase as requiring that an entity exercise authority or control with respect to the action at issue in the suit in order to be subject to liability as a fiduciary under this section. In
Pegram v. Herdrich,
530 U.S. 211, 120 S.Ct. 2143, 147 L.Ed.2d 164 (2000), the Court explained:
In every case charging breach of ERISA fiduciary duty, [ ] the threshold question is not whether the actions of some person employed to provide services under a plan adversely affected a plan beneficiary’s interest, but whether that person was acting as a fiduciary (that is, was performing a fiduciary function)
when taking the action subject to complaint.
Id.
at 226, 120 S.Ct. 2143 (emphasis added); see also
Chicago Dist. Council of Carpenters Welfare Fund v. Caremark, Inc.,
474 F.3d 463, 471-72 (7th Cir.2007) (ERISA plaintiff must show that entity “was acting in its capacity as a fiduciary at the time it took the actions that are the subject of the complaint”). Thus, AUL’s control over the separate account can support a finding of fiduciary status only if Leimkuehler’s claims for breach of fiduciary duty arise from AUL’s handling of the separate account.
They do not. Leimkuehler does not allege that AUL in any way mismanaged the separate account — say, by losing track of participants’ contributions or withdrawing funds in the separate account to pay for a company-wide vacation to Las Vegas.
Cf. Chao,
436 F.3d at 235 (defendant that received plan assets for purposes of purchasing insurance policies was a fiduciary under Section 1002(21)(A)(i) when he kept the money and provided fake insurance policies). Rather, Leimkuehler’s claims focus on share-class selection and revenue sharing, and AUL’s maintenance of the separate account involves neither. As we noted earlier and as Leimkuehler concedes, AUL selects share classes and decides how much it will receive in revenue sharing when it designs its investment-
options menu. Those steps occur well before a Plan participant deposits her contributions in the separate account and directs AUL where to invest those contributions. Because the actions Leimkuehler complains of do not implicate AUL’s control over the separate account, the separate account does not render AUL a fiduciary-under the circumstances of this case.
C
Finally, DOL proposes that AUL is a fiduciary because, in section 3.3 of its contract with the Plan, it retains the right to delete or substitute the funds Leim-kuehler has selected for the Plan. DOL acknowledges that AUL can be liable as a fiduciary only “to the extent” it exercises this contractual authority. DOL also acknowledges that neither of the two occasions on which AUL exercised its right under section 3.3 gives rise to an ERISA claim. In DOL’s view, however, AUL need never affirmatively exercise its section 3.3 authority in order to incur fiduciary responsibilities to the Plan. Instead, it “exercises” this authority in a negative sense every time it invests a participant’s contributions in one of the chosen mutual fund share classes, as opposed to a less expensive share class of that same mutual fund. This is effectively a “non-exercise” theory of exercise: because AUL
could,
unilaterally substitute less expensive share classes, its failure to do so amounts to an exercise of its authority.
This theory is unworkable. It conflicts with a common-sense understanding of the meaning of “exercise,” is unsupported by precedent, and would expand fiduciary responsibilities under Section 1002(21)(A) to entities that took no action at all with respect to a plan. In contrast to a named fiduciary, a functional fiduciary under Section 1002(21)(A) owes a duty to a plan through its actions, regardless of whether it chose to assume fiduciary responsibilities or even anticipated that such responsibilities might arise. Section 1002(21)(A)’s “reach is limited to circumstances where the individual actually exercises some authority,”
Trustees of the Graphic Commc’ns Int’l Union Upper Midwest Local 1M Health & Welfare Plan v. Bjorkedal,
516 F.3d 719, 733 (8th Cir.2008), and “people may be fiduciaries when they do certain things but be entitled to act in their own interests when they do others,”
Johnson v. Georgia-Pacific Corp.,
19 F.3d 1184, 1188 (7th Cir.1994). We agree with the Eighth Circuit that “[a]n act of omission fails to satisfy the requirement that the individual
exercise
discretionary authority over plan assets.”
Bjorkedal,
516 F.3d at 733 (emphasis in original). This means that AUL’s decision
not
to exercise its contractual right to substitute different (less expensive) funds for the Leimkuehler Plan does not make it a fiduciary.
Ill
In its cross-appeal, AUL challenges the district court’s order denying its motion for attorney’s fees and costs under ERISA, 29 U.S.C. § 1132(g), or costs under Federal Rule of Civil Procedure 54(d). The decision to award fees and/or costs under either provision is committed to the discretion of the district court, and we will reverse only in the case of an abuse of discretion. See
Holmstrom v. Metropolitan Life Ins. Co.,
615 F.3d 758, 779 (7th Cir.2010) (fees and costs under ERISA);
Rivera v. City of Chicago,
469 F.3d 631, 636 (7th Cir.2006) (costs under Rule 54(d)).
Although it acknowledged that AUL was entitled to a “modest presumption” that it would recover its fees and costs under ERISA, see,
e.g., Herman v. Central States, Se. & Sw. Areas Pension Fund,
423 F.3d 684, 695-96 (7th Cir.2005), the
district court declined to do so because it concluded that Leimkuehler’s position was “substantially justified.” (Although the Supreme Court’s decision in
Hardt v. Reliance Standard Life Ins. Co.,
560 U.S. 242, 130 S.Ct. 2149, 176 L.Ed.2d 998 (2010), has generated some confusion about the test that governs fee and costs determinations under Section 1132(g), compare
Kolbe & Kolbe Health & Welfare Benefit Plan v. Medical Coll. of Wis., Inc.,
657 F.3d 496, 505-06 (7th Cir.2011), with
Loomis v. Exelon Corp.,
658 F.3d 667, 675 (7th Cir.2011), AUL stated to the district court that it preferred the substantial justification test, and AUL does not challenge the district court’s use of that test on appeal.) The district court found that Leimkuehler’s suit was substantially justified for four reasons: that there were legitimate arguments for distinguishing this case from
Hecker;
that case law in other circuits supported Leimkuehler’s position; that AUL officials had themselves expressed reservations about disclosing revenue sharing; and that DOL had threatened AUL with suit over its revenue-sharing practices. AUL notes that this last point is incorrect (DOL was, in fact, threatening suit over something unrelated), but this error is not enough to undermine the district court’s substantial justification finding. The district court’s remaining reasons fall well within the bounds of its discretion and are enough by themselves to justify its denial of fees and costs under Section 1132(g).
We are likewise disinclined to disturb the district court’s denial of costs under Rule 54(d), which provides that “[ujnless a federal statute, these rules, or a court order provides otherwise, costs — other than attorney’s fees — should be allowed to the prevailing party.” At times, we have taken the view that Section 1132(g) “provides otherwise,” and that costs are therefore unavailable under Rule 54(d) in ERISA actions. See
Nichol v. Pullman Standard, Inc.,
889 F.2d 115, 121 (7th Cir.1989). That approach must be reconsidered in light of the Supreme Court’s recent decision in
Marx v. General Revenue Corp.,
— U.S. —, 133 S.Ct. 1166, 185 L.Ed.2d 242 (2013). In
Marx,
the Court was faced with the question whether Section 1692k(a)(3) of the Fair Debt Collection Practices Act “provided otherwise” than Rule 54(d)(1), or if instead the two could be harmonized.
Id.
at 1170-71. The Court opted for the latter approach. It began by noting that despite the “venerable presumption” in favor of granting costs under Rule 54(d), “the decision whether to award costs ultimately lies within the sound discretion of the district court.”
Id.
at 1172. It then held that a statute “provides otherwise” for purposes of Rule 54(d) only if it is literally contrary to the rule, in the sense that it constricts discretion that the rule recognizes.
Id.
at 1173. Applying that approach to Section 1132(g), we see nothing contrary to Rule 54(d) in the statute.
The district court anticipated all of this, however, when it proceeded under the assumption that costs were available under Rule 54(d). The only question was thus whether, as a matter of discretion, costs
should
be awarded. The court thought not, because it found that Leimkuehler (who was a party to the suit only in his capacity as Plan trustee) was unable to satisfy a costs award. The district court noted in this regard that AUL had advised the court that neither the Plan nor any individual Plan participants were necessary parties, and that the court had refrained from joining additional parties in partial reliance on AUL’s assurances. Because the district court had no authority to order an award of costs against a non-party, see
In re Cardizem CD Antitrust Litig.,
481 F.3d 355, 359 (6th Cir.2007),
and because Leimkuehler holds no assets in his capacity as trustee, the district court did not abuse its discretion in concluding that the losing party was unable to pay and that a costs award under Rule 54(d) was therefore unwarranted. See
Rivera,
469 F.3d at 636 (inability to pay can overcome Rule 54(d)’s “heavy presumption” that the losing party should pay costs).
The judgment of the district court is AFFIRMED.