Many defined-contribution pension plans offer participants an opportunity to select investments from a portfolio, which often includes mutual funds. In recent years participants in pension plans have contended that the sponsor offers too few funds (not enough choice), too many funds (producing confusion), or too expensive funds (meaning that the funds’ ratios of expenses to assets are needlessly high). See, e.g., Hecker v. Deere & Co., 556 F.3d 575, rehearing denied, 569 F.3d 708 (7th Cir.2009); Howell v. Motorola, Inc., 633 F.3d 552 (7th Cir.2011); Spano v. Boeing Co., 633 F.3d 574 (7th Cir.2011); George v. Kraft Foods Global, Inc., 641 F.3d 786 (7th Cir.2011). The district court decided that the current suit is a replay of Hecker and dismissed it on the pleadings. 2009 WL 4667092, 2009 U.S. Dist. Lexis 114626 (N.D.Ill. Dec. 9, 2009).

Exelon’s defined-contribution pension plan allows participants to choose how their retirement assets will be invested. It offers 32 options, including 24 mutual funds that are open to the public. These funds are no-load vehicles. In other words, they do not charge investors a fee to buy or sell shares. Purchases and sales occur at net asset value, calculated daily. A no-load fund covers its expenses by deducting them from the assets under management. So if these assets appreciate 10% in a given year, and the expenses come to 1%, investors receive a net gain of 9%; if the assets decline 5% in the market, investors’ net return is -6% that year. The funds available to participants in the Exelon Plan have expense ratios ranging from 0.03% to 0.96%. The low-expense funds tend to be passively managed (index funds, for example, which do not make any[*670] independent investment choices but simply track a designated portfolio such as the Standard & Poor’s 500 Index) and have features that discourage turnover (an index fund typically disallows new investments for a month or more following any withdrawal). The high-expense funds tend to be actively managed (that is, the fund’s investment advisers try to find and buy underpriced securities while selling ones that the advisers think are overvalued) and to allow rapid turnover both in the funds’ holdings and the participants’ investments. Higher turnover means higher brokerage fees and higher administrative expenses.

Plaintiffs, participants in Exelon’s Plan, contend that its administrators have violated their fiduciary duties under the Employee Retirement Income Security Act, see 29 U.S.C. § 1104(a), in two ways: by offering “retail” mutual funds, in which participants get the same terms (and thus bear the same expenses) as the general public; and by requiring participants to bear the economic incidence of those expenses themselves, rather than having the Plan cover these costs. Plaintiffs contend that Exelon should have arranged for access to “wholesale” or “institutional” investment vehicles. Some mutual funds offer a separate “institutional” class of shares, and Exelon’s Plan also could have participated in trusts and investment pools to which the general public does not have access.

Similar arguments were made in Hecker but did not prevail. Deere offered 25 retail mutual funds with expense ratios from 0.07% to just over 1% annually. We held that as a matter of law that was an acceptable array of investment options, observing that “all of these funds were also offered to investors in the general public, and so the expense ratios necessarily were set against the backdrop of market competition. The fact that it is possible that some other funds might have had even lower ratios is beside the point; nothing in ERISA requires every fiduciary to scour the market to find and offer the cheapest possible fund (which might, of course, be plagued by other problems).” 556 F.3d at 586. By offering a wide range of options, Hecker held, Deere’s plan complied with ERISA’s fiduciary duties.

Plaintiffs contend that the panel in Hecker retreated from this holding when denying a petition for rehearing. It did not. Two principal issues were disputed in Hecker: first, whether ERISA plans must offer “wholesale” or “institutional” funds; second, whether Deere’s portfolio of funds was covered by a safe harbor, 29 U.S.C. § 1104(c), that made the answer to the first question irrelevant. The opinion denying rehearing principally concerned the second issue. (Exelon does not rely on § 1104(c).) The panel reaffirmed its negative answer to the first question, stating that plaintiffs

argued- — and especially in their Petition for Rehearing they continue to argue— that the Plans were flawed because Deere decided to accept ‘retail’ fees and did not negotiate presumptively lower ‘wholesale’ fees. The opinion discusses a number of reasons why that particular assertion is not enough, in the context of these Plans, to state a claim, and we adhere to that discussion.

569 F.3d at 711. Unless Hecker is to be overruled, our plaintiffs cannot prevail. Two other circuits have agreed with Heck-er. See Renfro v. Unisys Corp., — F.3d - (3d Cir.2011); Braden v. Wal-Mart Stores, Inc., 588 F.3d 585 (8th Cir.2009). Plaintiffs do not persuade us to overrule Hecker and create a conflict.

Nothing in Jones v. Harris Associates, L.P., — U.S. -, 130 S.Ct. 1418, 176 L.Ed.2d 265 (2010), undermines Hecker1s analysis. The petition for rehearing in Hecker was denied three months after[*671] Jones came down. That case dealt with the fiduciary duties of investment advisers, which as the Court observed have a conflict of interest when seeking management fees from mutual funds under their effective control. Plaintiffs do not contend that the funds that Exelon selected had any control over it, or it over them; there is no reason to think that Exelon chose these funds to enrich itself at participants’ expense. To the contrary, Exelon had (and has) every reason to use competition in the market for fund management to drive down the expenses charged to participants, because the larger participants’ net gains, the better Exelon’s pension plan is. That enables Exelon to recruit better workers, or reduce wages and pension contributions without making the total package of compensation (wages plus fringe benefits) less attractive. Competition thus assists both employers and employees, as Hecker observed. (By contrast, the plaintiffs in Bra-den alleged that the plan sponsor limited participants’ options to ten funds as a result of kickbacks; while adopting the approach of Hecker, the eighth circuit held this allegation sufficient to state a fiduciary claim under ERISA. Nothing of the sort is alleged in this case.)

True, the participants in Exelon’s Plan press an argument that was not presented to the panel in Hecker: that the Plan should have paid the expenses directly, allowing participants to reap the gross rather than the net return. But whether to cover these expenses is a question of plan design, not of administration. The participants want Exelon to contribute more to the Plan than it does. ERISA does not create any fiduciary duty requiring employers to make pension plans more valuable to participants. When deciding how much to contribute to a plan, employers may act in their own interests. See, e.g., Hughes Aircraft Co. v. Jacobson, 525 U.S. 432, 119 S.Ct. 755, 142 L.Ed.2d 881 (1999); Lockheed Corp. v. Spink, 517 U.S. 882, 116 S.Ct. 1783, 135 L.Ed.2d 153 (1996). Fiduciary duties under ERISA are limited to a requirement of honest and prudent management of the assets that are under an administrator’s control. So the participants’ argument that Exelon should have ponied up additional money, to cover the operating expenses of their retirement vehicles, is a non-starter. What remains is the argument that flopped in Hecker: that Exelon should have offered only “wholesale” or “institutional” funds. Exelon’s Plan has at least 8 options other than “retail” mutual funds, and plaintiffs do not complain about these; instead they insist that the number of “retail” funds must be zero.

Note that this is not an argument about the absolute level of fees. Any participant who wants a fund with expenses under 0.1% can get it through Exelon’s Plan. Nor is it an argument that Exelon has left participants adrift and apt to blunder into the high-expense funds when they would be better off with the low-expense funds. Cf. Warren Bailey, Alok Kumar & David Ng, Behavioral biases of mutual fund investors, 102 J. Fin. Econ. 1 (2011). Both Exelon and the funds distribute literature and hold seminars for the participants, educating them about how the funds differ and how to identify the low-expense vehicles.- Plaintiffs do not contest the adequacy of the Plan’s and the funds’ disclosures. What plaintiffs contend instead is that, if a pension plan offers only “institutional” vehicles, fees will be lower on average, and that participants tempted by a high-expense fund might save.

One reason Hecker rejected this argument that the administrator’s fiduciary duties require limiting choices to “institutional” funds is that “retail” funds, being open to the public, give participants the benefits of competition. A pension plan that directs participants into privately held[*672] trusts or commingled pools (the sort of vehicles that insurance companies use for assets under their management) lacks the mark-to-market benchmark provided by a retail mutual fund. It can be hard to tell whether a closed fund is doing well or poorly, or whether its expenses are excessive in relation to the benefits they provide. It can be hard to value the vehicle’s assets (often real estate rather than stock or bonds) when someone wants to withdraw money, and any error in valuation can hurt other investors.

A helpful amicus brief filed by the Investment Company Institute tells us that the average expense ratio of institutional-share classes in equity funds in 2009 was 1.09%, which is higher than that of any of the retail funds offered to the participants in Exelon’s Plan. (The ICI calculates the average expense ratio of retail equity funds at 0.76%.) Likely the professional investors who negotiate for these investments are getting something extra for the money, but this expense ratio is not compatible with plaintiffs’ belief that institutional shares always have lower expenses. Meanwhile, institutional investment vehicles come with a drawback: lower liquidity. The retail funds that Exelon offers allow daily transfers. Participants can move their money from one vehicle to another whenever they wish, without paying a fee. In retirement, they can withdraw money daily. Institutional trusts and pools do not offer that choice. It is not clear that participants would gain from lower expense ratios at the cost of lower liquidity.

Plaintiffs treat the situation as one in which Exelon, whose retirees have more than $1 billion in the Plan, could exercise “buying power” by negotiating lower fees in exchange for a promise to place more money with a given investment manager, while demanding the same retail services (such as daily transfers) for which mutual funds charge their normal expenses. Alternatively, plaintiffs contend, Exelon could use its “buying power” to insist that mutual funds charge a capitation fee (an annual flat price per investor) in lieu of expenses as a percentage of capital under management.

Now it isn’t clear to us why mutual funds would offer lower prices just because participants in this Plan have pension wealth that in the aggregate exceeds $1 billion. Exelon can’t commit that sum, or any portion of it, to any one fund without abandoning the arrangement under which the participants themselves choose where their money will be invested. The expenses of retail funds derive in large measure from the need to deal with investors one at a time: to receive and mail small checks, to print and mail individual prospectuses and account statements, frequently to exchange modest sums from one fund to another, and so on. Expenses per dollar under management necessarily are higher if the average account is $100,000 than if it is $100,000,000. Hertz gets a fleet discount from General Motors when it orders 10,000 cars at a time, but Hertz does not secure fleet discounts for members of its # 1 Club to buy their own GM cars; retail transactions occur at retail prices. So too with retail transactions in mutual funds.

Likewise it isn’t clear to us why participants would view a capitation fee as a gain. A flat-fee structure might be beneficial for participants with the largest balances, but, for younger employees and others with small investment balances, a capitation fee could work out to more, per dollar under management, than a fee between 0.03% and 0.96% of the account balance. (The same holds true if plaintiffs’ argument is limited to fees of the Plan’s own record-keeper; flat payments per participant may help some participants but hurt others,[*673] depending on the size of each participant’s account.)

Even if a restructured means of covering a fund’s costs would benefit participants, it is not something that Exelon could achieve. Mutual funds are regulated under the Securities Act of 1933, the Securities Exchange Act of 1934, and the Investment Company Act of 1940. These statutes, and their implementing regulations, require mutual funds to treat alike all investors holding the same. class of shares. See 17 C.F.R. §. 270.18Í-3. So the sponsor of a mutual fund could not agree with Exelon to offer a special deal (lower expense ratios, capitation fees rather than expenses paid from account balances) while giving participants the same rights as retail investors. And it could be hard to establish a separate class of shares, limited to Exelon. That might run afoul of the 1940 Act’s rule against senior securities, 15 U.S.C. § 80a-18(f), or the Internal Revenue Code’s rule against preferential dividends from investment companies, 26 U.S.C. § 562(c). (A mutual fund’s failure to charge expenses against certain investors would be economically equivalent to a preferential dividend.)

Pension plans’ sponsors could get around these limits by creating in-house or captive mutual funds, which then would have only one class of shares and one set of rights. But captive funds run into the sort of problems we discussed above. They offer less choice (participants would have 1 or 2 options, not the 32 Exelon currently offers); they also are less liquid, less diversified, and may be harder to value. And a captive fund also would be smaller, so the expense ratio per dollar under management could be higher, especially if the fund had some expenses that do not vary with the amount under management. (The cost of writing a registration statement and prospectus, for example, is largely fixed, so the smaller the fund the larger this expense looms as a percentage of invested capital.)

■ Plaintiffs’ theory is paternalistic. They appear to believe that participants should prefer captive funds, even with loss of liquidity, and should not be allowed to invest in the funds from the Fidelity Group that Exelon’s Plan now offers. According to plaintiffs, participants like these mutual funds ■ for “the wrong reasons,” such as advertising. Since the seminars that Exelon offers have not dissuaded the participants from continuing to commit what plaintiffs call mistakes, they want the judiciary to force Exelon to make these investments impossible. Hostility to advertising has a long history, reflecting a belief that advertising is costly and thus must drive price up; but available data suggest that advertising promotes competition, which drives price down by more than the costs of the ads. See, e.g., Lee Benham, The Effect of Advertising on the Price of Eyeglasses, 15 J.L. & Econ. 337 (1972); Craig A. Depken II & Dennis P. Wilson, Is Advertising Good or Bad?, 77 J. Business S61 (April 2004); John Rizzo, Advertising and Competition in the Ethical Pharmaceutical Industry, 42 J.L. & Econ. 89 (1999).

For current purposes, it does not matter whether advertising is good or bad; all that matters is the absence from ERISA of any rule that forbids plan sponsors to allow participants to make their own choices. Far from reflecting a paternalistic approach, the safe harbor in § 1104(c) encourages sponsors to allow more choice to participants in defined-contribution plans. Exelon offered participants a menu that includes high-expense, high-risk, and potentially high-return funds, together with low-expense index funds that track the market, and low-expense, low-risk, modest-return bond funds. It has left choice to the people who have[*674] the most interest in the outcome, and it cannot be faulted for doing this.

This concludes our discussion of the merits. Plaintiffs have filed a second appeal, No. 10-1755, from the district court’s award of some $42,000 in costs to Exelon. 2010 U.S. Dist. Lexis 24405 (N.D.Ill. Mar. 11, 2010). The district court relied on Fed.R.Civ.P. 54(d), which says that prevailing parties presumptively recover their costs. Plaintiffs reply that there is an exception. Rule 54(d)(1) begins: “Unless a federal statute, these rules, or a court order provides otherwise”. They contend that 29 U.S.C. § 1132(g)(1) “provides otherwise”. It reads: “In any action under this subchapter (other than an action described in paragraph (2) [to enforce § 1145]) by a participant, beneficiary, or fiduciary, the court in its discretion may allow a reasonable attorney’s fee and costs of action to either party.” Section 1132(g)(1) gives the district judge more discretion than does Rule 54(d), plaintiffs contend, and therefore supersedes the rule. Plaintiffs then assert that an award of attorneys’ fees under § 1132(g)(1) depends on a finding that the plaintiff sued in bad faith or in order to harass; an award of costs must depend on the same standard, the argument continues. The district court did not order plaintiffs to pay Exelon’s attorneys’ fees under § 1132(g)(1) and therefore, the argument wraps up, cannot properly order plaintiffs to pay costs either.

One court of appeals has rejected this line of argument, and none has accepted it. Quan v. Computer Sciences Corp., 623 F.3d 870, 888-89 (9th Cir.2010), holds that § 1132(g)(1) does not “provide otherwise” than Rule 54(d) because it never forbids an award of costs. The ninth circuit wrote: “To ‘provide otherwise’ than Rule 54(d)(1), the statute or rule would have to bar an award of costs to a prevailing party.” 623 F.3d at 888 (emphasis in original). We are skeptical about this conclusion. Rule 54(d) establishes a presumption in favor of an award to the prevailing party. A statute that established a presumption against an award of costs, but without forbidding one, would provide “otherwise” than the rule; similarly a statute establishing a presumption that the winner pays the loser’s costs would provide “otherwise” than Rule 54(d), even though it did not forbid an award to the winner.

Decisions in this circuit could be read both to support and to reject the conclusion in Quan. Compare Nichol v. Pullman Standard Inc., 889 F.2d 115, 121 (7th Cir.1989), with Quinn v. Blue Cross & Blue Shield Association, 161 F.3d 472, 478-79 (7th Cir.1998). Our court has never grappled directly with the subject, and it is not appropriate to read oblique remarks as answering a question not squarely posed. We need not resolve this question definitively today, because one of the minor premises in plaintiffs’ syllogism is wrong. Plaintiffs believe that only a litigant who proceeds in bad faith, or to harass, can be required to pay attorneys’ fees under § 1132(g)(1), and that bad faith therefore must be essential to an award of costs. That’s not what § 1132(g)(1) says.

Section 1132(g)(1) authorizes a district “court in its discretion [to] allow a reasonable attorney’s fee and costs of action to either party.” The Supreme Court discussed the meaning of this language in Hardt v. Reliance Standard Life Insurance Co., — U.S. -, 130 S.Ct. 2149, 176 L.Ed.2d 998 (2010), and held that “a court ‘in its discretion’ may award fees and costs ‘to either party’ as long as the fee claimant has achieved ‘some degree of success on the merits.’ ” 130 S.Ct. at 2152 (citations omitted). In other words, even the ultimate loser could receive an award of attorneys’ fees and costs, if on the way to defeat the litigant won a skirmish that[*675] conferred some legal benefit. See Ruckelshaus v. Sierra Club, 463 U.S. 680, 694, 103 S.Ct. 3274, 77 L.Ed.2d 938 (1983). A district judge need not find that the party ordered to pay fees has engaged in harassment or otherwise litigated in bad faith. Language in some appellate opinions declaring “bad faith” vital to an award under § 1132(g)(1) did not survive Hardt. (Whether other approaches, such as Bittner v. Sadoff & Rudoy Industries, 728 F.2d 820 (7th Cir.1984), which analogized § 1132(g)(1) to the Equal Access to Justice Act, survived Hardt, is yet another issue we can avoid until the answer matters.)

Both the rule and the statute give the district judge discretion to decide whether an award of costs is appropriate. Plaintiffs did not succeed on any issue in this litigation, so the award could not run in their favor under Hardt’s standard. Doubtless § 1132(g)(1) gave the district judge discretion to deny Exelon’s request for costs — but then so did Rule 54(d). If the district judge had understood Rule 54(d) to make an award in Exelon’s favor mandatory, then a remand would be necessary, but the judge recognized that he possessed discretion. Plaintiffs stake their all on the proposition that, under § 1132(g)(1), attorneys’ fees and costs must be awarded (or denied) together, and may be awarded only to penalize misconduct by the losing side. Because that’s not the statutory standard, we can leave to another day the question whether § 1132(g)(1) supersedes Rule 54(d)(1) in some other situation.

Affirmed