delivered the opinion of the Court.
Federal Rule of Civil Procedure 54(d)(1) gives district courts discretion to award costs to prevailing defendants “[u]nless a federal statute . . . provides otherwise.” The Fair Debt Collection Practices Act (FDCPA), 91 Stat. 881,15 U. S. C. § 1692k(a)(3), provides that “[o]n a finding by the court that an action under this section was brought in bad faith and for the purpose of harassment, the court may award *374to the defendant attorney’s fees reasonable in relation to the work expended and costs.” This case presents the question whether § 1692k(a)(3) “provides otherwise” than Rule 54(d)(1). We conclude that § 1692k(a)(3) does not “provid[e] otherwise,” and, thus, a district court may award costs to prevailing defendants in FDCPA cases without finding that the plaintiff brought the case in bad faith and for the purpose of harassment.
I
Petitioner Olivea Marx defaulted on a student loan guaranteed by EdFund, a division of the California Student Aid Commission. In September 2008, EdFund hired respondent General Revenue Corporation (GRC) to collect the debt. One month later, Marx filed an FDCPA enforcement action against GRC.1 Marx alleged that GRC had violated the FDCPA by harassing her with phone calls several times a day and falsely threatening to garnish up to 50% of her wages and to take the money she owed directly from her bank account. Shortly after the complaint was filed, GRC made an offer of judgment under Federal Rule of Civil Procedure 68 to pay Marx $1,500, plus reasonable attorney’s fees and costs, to settle any claims she had against it. Marx did not respond to the offer. She subsequently amended her complaint to add a claim that GRC unlawfully sent a fax to her workplace that requested information about her employment status.
Following a 1-day bench trial, the District Court found that Marx had failed to prove any violation of the FDCPA. As the prevailing party, GRC submitted a bill of costs seeking $7,779.16 in witness fees, witness travel expenses, and deposition transcript fees. The court disallowed several *375items of costs and, pursuant to Federal Rule of Civil Procedure 54(d)(1), ordered Marx to pay GRC $4,543.03. Marx filed a motion to vacate the award of costs, arguing that the court lacked authority to award costs under Rules 54(d)(1) and 68(d) because 15 U. S. C. § 1692k(a)(3) sets forth the exclusive basis for awarding costs in FDCPA cases.2 Section 1692k(a)(3) provides, in relevant part: “On a finding by the court that an action under this section was brought in bad faith and for the purpose of harassment, the court may award to the defendant attorney’s fees reasonable in relation to the work expended and costs.” Marx argued that because the court had not found that she brought the case in bad faith and for the purpose of harassment, GRC was not entitled to costs. The District Court rejected Marx’s argument, concluding that § 1692k(a)(3) does not displace a court’s discretion to award costs under Rule 54(d)(1) and that costs should also be awarded under Rule 68(d).
The Tenth Circuit affirmed but agreed only with part of the District Court’s reasoning. In particular, the court disagreed that costs were allowed under Rule 68(d). 668 F. 3d 1174, 1182 (2011). It explained that “Rule 68 applies only where the district court enters judgment in favor of a plaintiff” for less than the amount of the settlement offer and not where the plaintiff loses outright. Ibid, (citing Delta Air Lines, Inc. v. August, 450 U. S. 346, 352 (1981)). Because the District Court had not entered judgment in favor of Marx, the court concluded that costs were not allowed under Rule 68(d). 668 F. 3d, at 1182. Nevertheless, the court found that costs were allowed under Rule 54(d)(1), which *376grants district courts discretion to award costs to prevailing parties unless a federal statute or the Federal Rules of Civil Procedure provide otherwise. Id., at 1178, 1182. After describing the “venerable” presumption that prevailing parties are entitled to costs, id., at 1179, the court concluded that nothing in the text, history, or purpose of § 1692k(a)(3) indicated that it was meant to displace Rule 54(d)(1), id., at 1178-1182. Judge Lucero dissented, arguing that “[t]he only sensible reading of [§ 1692k(a)(3)] is that the district court may only award costs to a defendant” upon finding that the action was brought in bad faith and for the purpose of harassment and that to read it otherwise rendered the phrase “and costs” superfluous. Id., at 1187 (emphasis in original).
We granted certiorari, 566 U. S. 1021 (2012), to resolve a conflict among the Circuits regarding whether a prevailing defendant in an FDCPA case may be awarded costs where the lawsuit was not brought in bad faith and for the purpose of harassment. Compare 668 F. 3d, at 1182 (case below), with Rouse v. Law Offices of Rory Clark, 603 F. 3d 699, 701 (CA9 2010). We now affirm the judgment of the Tenth Circuit.
II
As in all statutory construction cases, we “ ‘assum[e] that the ordinary meaning of [the statutory] language accurately expresses the legislative purpose.’” Hardt v. Reliance Standard Life Ins. Co., 560 U. S. 242, 251 (2010) (quoting Gross v. FBL Financial Services, Inc., 557 U. S. 167, 175 (2009); alteration in original). In this case, we must construe both Rule 54(d)(1) and § 1692k(a)(3) and assess the relationship between them.
A
Rule 54(d)(1) is straightforward. It provides, in relevant part: “Unless a federal statute, these rules, or a court order provides otherwise, costs—other than attorney’s fees— should be allowed to the prevailing party.”
*377As the Tenth Circuit correctly recognized, Rule 54(d)(1) codiñes a venerable presumption that prevailing parties are entitled to costs.3 Notwithstanding this presumption, the word “should” makes clear that the decision whether to award costs ultimately lies within the sound discretion of the district court. See Taniguchi v. Kan Pacific Saipan, Ltd., 566 U. S. 560, 565 (2012) (“Federal Rule of Civil Procedure 54(d) gives courts the discretion to award costs to prevailing parties”). Rule 54(d)(1) also makes clear, however, that this discretion can be displaced by a federal statute or a Federal Rule of Civil Procedure that “provides otherwise.”
A statute “provides otherwise” than Rule 54(d)(1) if it is “contrary” to the Rule. See 10 J. Moore, Moore’s Federal Practice § 54.101[l][c], p. 54-159 (3d ed. 2012) (hereinafter 10 Moore’s). Because the Rule grants district courts discretion to award costs, a statute is contrary to the Rule if it limits that discretion. A statute may limit a court’s discretion in several ways, and it need not expressly state that it is displacing Rule 54(d)(1) to do so. For instance, a statute providing that “plaintiffs shall not be liable for costs” is contrary to Rule 54(d)(1) because it precludes a court from awarding costs to prevailing defendants. See, e. g., 7 U. S. C. § 18(d)(1) (2006 ed., Supp. V) (“The petitioner shall not be liable for costs in the district court”). Similarly, a statute providing that plaintiffs may recover costs only under certain condi*378tions is contrary to Rule 54(d) because it precludes a court from awarding costs to prevailing plaintiffs when those conditions have not been satisfied. See, e. g., 28 U. S. C. § 1928 (“[N]o costs shall be included in such judgment, unless the proper disclaimer has been filed in the United States Patent and Trademark Office”).
Importantly, not all statutes that provide for costs are contrary to Rule 54(d)(1). A statute providing that “the court may award costs to the prevailing party,” for example, is not contrary to the Rule because it does not limit a court’s discretion. See 10 Moore’s § 54.101[l][c], at 54-159 (“A number of statutes state simply that the court may award costs in its discretion. Such a provision is not contrary to Rule 54(d)(1) and does not displace the court’s discretion under the Rule”).
Marx and the United States as amicus curiae suggest that any statute that specifically provides for costs displaces Rule 54(d)(1), regardless of whether it is contrary to the Rule. Brief for Petitioner 17; Brief for United States as Amicus Curiae 11-12 (hereinafter Brief for United States). The United States relies on the original 1937 version of Rule 54(d)(1), which provided, “‘Except when express provision therefor is made either in a statute of the United States or in these rules, costs shall be allowed as of course to the prevailing party unless the court otherwise directs.’” Id., at 12 (quoting Rule). Though the Rules Committee updated the language of Rule 54(d)(1) in 2007, the change was “stylistic only.” Advisory Committee’s Notes, 28 U. S. C. App., p. 734 (2006 ed., Supp. V). Accordingly, the United States asserts that any “express provision” for costs should displace Rule 54(d)(1).
We are not persuaded, however, that the original version of Rule 54(d) should be interpreted as Marx and the United States suggest. The original language was meant to ensure that Rule 54(d) did not displace existing costs provisions that were contrary to the Rule. Under the prior language, *379statutes that simply permitted a court to award costs did not displace the Rule. See 6 J. Moore, Moore’s Federal Practice §54.71[1], p. 54-304 (2d ed. 1996) (“[W]hen permissive language is used [in a statute regarding costs] the district court may, pursuant to Rule 54(d), exercise a sound discretion relative to the allowance of costs”). Rather, statutes had to set forth a standard for awarding costs that was different from Rule 54(d)(1) in order to displace the Rule. See Friedman v. Ganassi, 853 F. 2d 207, 210 (CA3 1988) (holding that 15 U. S. C. § 77k(e) is not an “express provision” under Rule 54(d) because it does not provide an “alternative standard” for awarding taxable costs). The original version of Rule 54(d) is consistent with our conclusion that a statute must be contrary to Rule 54(d)(1) in order to displace it.4
*380B
We now turn to whether § 1692k(a)(3) is contrary to Rule 54(d)(1). The language of § 1692k(a)(3) and the context surrounding it persuade us that it is not.
1
The second sentence of § 1692k(a)(3) provides: “On a finding by the court that an action under this section was brought in bad faith and for the purpose of harassment, the court may award to the defendant attorney’s fees reasonable in relation to the work expended and costs.” 5
GRC contends that the statute does not address whether costs may be awarded in this case—where the plaintiff brought the ease in good faith—and thus it does not set forth a standard for awarding costs that is contrary to Rule 54(d)(1). In its view, Congress intended §1692k(a)(3) to deter plaintiffs from bringing nuisance lawsuits. It, therefore, expressly provided that when plaintiffs bring an action in bad faith and for the purpose of harassment, the court may award attorney’s fees and costs to the defendant. The statute does address this type of case—i. e., cases in which the plaintiff brings the action in bad faith and for the purpose of harassment. But it is silent where bad faith and purpose of harassment are absent, and silence does not displace the background rule that a court has discretion to award costs.
Marx and the United States take the contrary view. They concede that the language does not expressly limit a court’s discretion to award costs under Rule 54(d)(1), Brief for Petitioner 10; Brief for United States 19, but argue that it does so by negative implication. Invoking the expressio unins *381canon of statutory construction, they contend that by specifying that a court may award attorney’s fees and costs when an action is brought in bad faith and for the purpose of harassment, Congress intended to preclude a court from awarding fees and costs when bad faith and purpose of harassment are absent. They further argue that unless §1692k(a)(3) sets forth the exclusive basis on which a court may award costs, the phrase “and costs” would be superfluous. According to this argument, Congress would have had no reason to specify that a court may award costs when a plaintiff brings an action in bad faith if it could have nevertheless awarded costs under Rule 54(d)(1). Finally, the United States argues that § 1692k(a)(3) is a more specific cost statute that displaces Rule 54(d)(l)’s more general rule.
The context surrounding §1692k(a)(3) persuades us that GRC’s interpretation is correct.
2
The argument of Marx and the United States depends critically on whether § 1692k(a)(3)’s allowance of costs creates a negative implication that costs are unavailable in any other circumstances. The force of any negative implication, however, depends on context. We have long held that the ex-pressio unius canon does not apply “unless it is fair to suppose that Congress considered the unnamed possibility and meant to say no to it,” Barnhart v. Peabody Coal Co., 537 U. S. 149, 168 (2003), and that the canon can be overcome by “contrary indications that adopting a particular rule or statute was probably not meant to signal any exclusion,” United States v. Vonn, 535 U. S. 55, 65 (2002). In this case, context persuades us that Congress did not intend §1692k(a)(3) to foreclose courts from awarding costs under Rule 54(d)(1).
First, the background presumptions governing attorney’s fees and costs are a highly relevant contextual feature. As already explained, under Rule 54(d)(1) a prevailing party is entitled to recover costs from the losing party unless a fed*382eral statute, the Federal Rules of Civil Procedure, or a court order “provides otherwise.” The opposite presumption exists with respect to attorney’s fees. Under the “bedrock principle known as the ‘“American Rule,’”” “[e]ach litigant pays his own attorney’s fees, win or lose, unless a statute or contract provides otherwise.” Hardt, 560 U. S., at 253 (quoting Ruckelshaus v. Sierra Club, 463 U. S. 680, 683 (1983)). Notwithstanding the American Rule, however, we have long recognized that federal courts have inherent power to award attorney’s fees in a narrow set of circumstances, including when a party brings an action in bad faith. See Chambers v. NASCO, Inc., 501 U. S. 32, 45-46 (1991) (explaining that a court has inherent power to award attorney’s fees to a party whose litigation efforts directly benefit others, to sanction the willful disobedience of a court order, and to sanction a party who has acted in bad faith, vexatiously, wantonly, or for oppressive reasons); Alyeska Pipeline Service Co. v. Wilderness Society, 421 U. S. 240, 257-259 (1975) (same).
It is undisputed that § 1692k(a)(3) leaves the background rules for attorney’s fees intact. The statute provides that when the plaintiff brings an action in bad faith, the court may award attorney’s fees to the defendant. But, as noted, a court has inherent power to award fees based on a litigant’s bad faith even without § 1692k(a)(3). See Chambers, supra, at 45-46. Because § 1692k(a)(3) codifies the background rule for attorney’s fees, it is dubious to infer congressional intent to override the background rule with respect to costs. The statute is best read as codifying a court’s pre-existing authority to award both attorney’s fees and costs.6
*383Next, the second sentence of § 1692k(a)(3) must be understood in light of the sentence that precedes it.7 The first sentence of § 1692k(a)(3) provides that defendants who violate the FDCPA are liable for the plaintiffs attorney’s fees and costs. The second sentence of §1692k(a)(3) similarly provides that plaintiffs who bring an action in bad faith and for the purpose of harassment may be liable for the defendant’s fees and costs.
If Congress had excluded “and costs” in the second sentence, plaintiffs might have argued that the expression of costs in the first sentence and the exclusion of costs in the second meant that defendants could only recover attorney’s fees when plaintiffs bring an action in bad faith. By adding “and costs” to the second sentence, Congress foreclosed that argument, thereby removing any doubt that defendants may recover costs as well as attorney’s fees when plaintiffs bring suits in bad faith. See Ali v. Federal Bureau of Prisons, 552 U. S. 214, 226 (2008) (explaining that a phrase is not superfluous if used to “remove . .. doubt” about an issue); Fort *384Stewart Schools v. FLRA, 495 U. S. 641, 646 (1990) (explaining that “technically unnecessary” examples may have been “inserted out of an abundance of caution”). The fact that there might have been a negative implication that costs are precluded, depending on whether Congress included or excluded the phrase “and costs,” weighs against giving effect to any implied limitation.
Finally, the language in §1692k(a)(8) sharply contrasts with other statutes in which Congress has placed conditions on awarding costs to prevailing defendants. See, e. g., 28 U. S. C. § 1928 (“[N]o costs shall be included in such judgment, unless the proper disclaimer has been filed in the United States Patent and Trademark Office prior to the commencement of the action” (emphasis added)); 42 U. S. C. § 1988(b) (“[I]n any action brought against a judicial officer . . . such officer shall not be held liable for any costs . . . unless such action was clearly in excess of such officer’s jurisdiction” (emphasis added)).
Although Congress need not use explicit language to limit a court’s discretion under Rule 54(d)(1), its use of explicit language in other statutes cautions against inferring a limitation in § 1692k(a)(3). These statutes confirm that Congress knows how to limit a court’s discretion under Rule 54(d)(1) when it so desires. See Small v. United States, 544 U. S. 385, 398 (2005) (THOMAS, J., dissenting) (explaining that “Congress’ explicit use of [language] in other provisions shows that it specifies such restrictions when it wants to do so”). Had Congress intended the second sentence of § 1692k(a)(3) to displace Rule 54(d)(1), it could have easily done so by using the word “only” before setting forth the condition “[o]n a finding by the court that an action . . . was brought in bad faith and for the purpose of harassment. ...”8
*3853
As the above discussion suggests, we also are not persuaded by Marx’s objection that our interpretation renders the phrase “and costs” superfluous. As noted, supra, at 383, the phrase “and costs” would not be superfluous if Congress included it to remove doubt that defendants may recover costs when plaintiffs bring suits in bad faith. But even assuming that our interpretation renders the phrase “and costs” superfluous, that would not alter our conclusion. The canon against surplusage is not an absolute rule, see Arlington Central School Dist. Bd. of Ed. v. Murphy, 548 U. S. 291, 299, n. 1 (2006) (“While it is generally presumed that statutes do not contain surplusage, instances of surplusage are not unknown”); Connecticut Nat. Bank v. Germain, 503 U. S. 249, 253 (1992) (“Redundancies across statutes are not unusual events in drafting . . . ”), and it has considerably less force in this case.
First, the canon against surplusage “assists only where a competing interpretation gives effect to every clause and word of a statute.” Microsoft Corp. v. i4i Ltd. Partnership, 564 U. S. 91, 106 (2011) (internal quotation marks omitted). But, in this case, no interpretation of § 1692k(a)(3) gives effect to every word. Both Marx and the United States admit that a court has inherent power to award attorney’s fees to a defendant when the plaintiff brings an action in bad faith. Because there was, consequently, no need for Congress to specify that courts have this power, § 1692k(a)(3) is superfluous insofar as it addresses attorney’s fees. In light of this redundancy, we are not overly concerned that the reference to costs may be redundant as well.
Second, redundancy is “hardly unusual” in statutes addressing costs. See id., at 107. Numerous statutes overlap with Rule 54(d)(1). See, e. g., 12 U. S. C. § 2607(d)(5) (“[T]he *386court may award to the prevailing party the court costs of the action”); § 5565(b) (2006 ed., Supp. V) (“[T]he [Consumer Financial Protection] Bureau . . . may recover its costs in connection with prosecuting such action if [it]... is the prevailing party in the action”); 15 U. S. C. § 6104(d) (2006 ed.) (“The court. . . may award costs of suit and reasonable fees for attorneys and expert witnesses to the prevailing party”); § 7706(f)(4) (“In the case of any successful action . . . the court, in its discretion, may award the costs of the action”); § 7805(b)(3) (“[T]he court may award to the prevailing party costs”); § 8131(2) (2006 ed., Supp. V) (“The court may also, in its discretion, award costs and attorneys fees to the prevailing party”); 29 U. S. C. § 431(c) (2006 ed.) (“The court... may, in its discretion, . . . allow a reasonable attorney’s fee to be paid by the defendant, and costs of the action”); 42 U. S. C. § 3612(p) (“[T]he court, ... in its discretion, may allow the prevailing party ... a reasonable attorney’s fee and costs”); § 3613(c)(2) (“[T]he court, in its discretion, may allow the prevailing party ... a reasonable attorney’s fee and costs”); 47 U. S. C. § 551(f)(2) (“[T]he court may award . . . other litigation costs reasonably incurred”).
Finally, the canon against surplusage is strongest when an interpretation would render superfluous another part of the same statutory scheme. Cf. United States v. Jicarilla Apache Nation, 564 U. S. 162, 185 (2011) (“ ‘As our cases have noted in the past, we are hesitant to adopt an interpretation of a congressional enactment which renders superfluous another portion of that same law’ ” (quoting Mackey v. Lanier Collection Agency & Service, Inc., 486 U. S. 825, 837 (1988))). Because § 1692k(a)(3) is not part of Rule 54(d)(1), the force of this canon is diminished.
4
Lastly, the United States contends that § 1692k(a)(3) “establishes explicit cost-shifting standards that displace Rule 54(d)(l)’s more general default standard.” Brief for United States 17; see also EC Term of Years Trust v. United States, *387550 U. S. 429,433 (2007) (“ ‘[A] precisely drawn, detailed statute pre-empts more general remedies’” (quoting Brown v. GSA, 425 U. S. 820, 834 (1976))). Were we to accept the argument that § 1692k(a)(3) has a negative implication, this argument might be persuasive. But the context of § 1692k(a)(3) indicates that Congress was simply confirming the background rule that courts may award to defendants attorney’s fees and costs when the plaintiff brings an action in bad faith. The statute speaks to one type of case—the case of the bad-faith and harassing plaintiff. Because Marx did not bring this suit in bad faith, this case does not “fal[l] within the ambit of the more specific provision.” Brief for United States 13; see also RadLAX Gateway Hotel, LLC v. Amalgamated Bank, 566 U. S. 639, 648 (2012) (“When the conduct at issue falls within the scope of both provisions, the specific presumptively governs ...” (emphasis in original)).9 Accordingly, this canon is inapplicable.
I I I
Because we conclude that the second sentence of § 1692k(a)(3) is not contrary to Rule 54(d)(1), and, thus, does not displace a district court’s discretion to award costs under *388the Rule, we need not address GRC’s alternative argument that costs were required under Rule 68.
The judgment of the Court of Appeals is affirmed.
It is so ordered.
The FDCPA is a consumer protection statute that prohibits certain abusive, deceptive, and unfair debt collection practices. See 15 U. S. C. § 1692. The FDCPA’s private-enforcement provision, § 1692k, authorizes any aggrieved person to recover damages from “any debt collector who fails to comply with any provision” of the FDCPA. § 1692k(a).
Under Rule 68(d), if a defendant makes a settlement offer, and the plaintiff rejects it and later obtains a judgment that is less favorable than the one offered her, the plaintiff must pay the costs incurred by the defendant after the offer was made. See Fed. Rule Civ. Proc. 68(d) (“If the judgment that the offeree finally obtains is not more favorable than the 'unaccepted offer, the offeree must pay the costs incurred after the offer was made”).
Prior to the adoption of the Federal Rules, prevailing parties were entitled to costs as of right in actions at law while courts had discretion to award costs in equity proceedings. See Ex parte Peterson, 253 U. S. 300, 317-318 (1920) (‘"While in equity proceedings the allowance and imposition of costs is, unless controlled by statute or rule of court, a matter of discretion, it has been uniformly held that in actions at law the prevailing party is entitled to costs as of right, except in those few cases where by express statutory provision or by established principles costs are denied” (citation omitted)); Mansfield, C. & L. M. R. Co. v. Swan, 111 U. S. 379, 387 (1884) (“[B]y the long established practice and universally recognized rule of the common law, in actions at law, the prevailing party is entitled to recover a judgment for costs ... ”).
The dissent provides no stable definition of “provides otherwise.” First, it argues that a statute “provides otherwise” if it is “different” from Rule 54(d)(1). Post, at 390 (opinion of Sotomayor, J.). That interpretation renders the Rule meaningless because every statute is “different” insofar as it is not an exact copy of the Rule. Next, it argues that a statute “provides otherwise” if it is an “‘express provision’ relating to costs.” Post, at 389. Under that view, a statute providing that “the court may award costs to the prevailing party” would “provide otherwise.” We do not think such a statute provides otherwise—it provides “same-wise,” and the treatise on which the dissent relies supports our view. See 10 G. Wright, A. Miller, & M. Kane, Federal Practice and Procedure § 2670, p. 258 (3d ed. 1998 and Supp. 2012) (“[Statutes that] are permissive in character . . . are not inconsistent with the discretion given the district court by Rule 54(d)”). Finally, the dissent seems to implicitly accept that “otherwise” means “to the contrary” in the course of arguing that a doctor’s instruction to take medication “ ‘in the morning’ ” would supersede an instruction on the medication label to “‘take [it] twice a day unless otherwise directed,’ ” because the patient would understand the doctor’s advice to mean that he should take the medicine “once a day, each morning.” Post, at 391. If the patient understands the doctor to mean “once a day, each-morning,” we agree that such advice would “provide otherwise,” because the doctor’s order would be “contrary” to the label’s instruction. For the reasons set forth in Part II-B, however, we are not convinced that § 1692k(a)(3) is “contrary” to Rule 54(d)(1).
It is undisputed that GRC is not entitled to costs under § 1692k(a)(3) because the District Court did not find that Marx brought this action in bad faith. But Rule 54(d)(1) independently authorizes district courts to award costs to prevailing parties. The question in this case is not whether costs are allowed under §1692k(a)(3) but whether §1692k(a)(3) precludes an award of costs under Rule 54(d)(1).
Indeed, had Congress intended § 1692k(a)(3) to foreclose a court’s discretion to award costs, it could not have chosen a more circuitous way to do so. The statute sets forth the circumstances in which a court “may” award costs. But under Marx’s and the United States’ view, the only consequence of the statute is to set forth the circumstances in which it may not award costs.
Section 1692k(a) provides:
“Except as otherwise provided by this section, any debt collector who fails to comply with any provision of this subchapter with respect to any person is liable to such person in an amount equal to the sum of—
“(1) any actual damages sustained by such person as a result of such failure;
“(2)(A) in the case of any action by an individual, such additional damages as the court may allow, but not exceeding $1,000; or
“(B) in the ease of a class action, (i) such amount for each named plaintiff as could be recovered under subparagraph (A), and (ii) such amount as the court may allow for all other class members, without regard to a minimum individual recovery, not to exceed the lesser of $500,000 or 1 per centum of the net worth of the debt collector; and
“(3) in the case of any successful action to enforce the foregoing liability, the costs of the action, together with a reasonable attorney’s fee as determined by the court. On a finding by the court that an action under this section was brought in bad faith and for the purpose of harassment, the court may award to the defendant attorney’s fees reasonable in relation to the work expended and costs.”
Marx also suggests that § 1692k(a)(3) is similar to the Pipeline Safety Act, 49 U. S. C. § 60121(b), which provides: “The court may award costs to a prevailing defendant when the action is unreasonable, frivolous, or *385meritless.” We have never had occasion to interpret § 60121(b) and its interaction with Rule 54(d)(1).
Marx, the United States, and GRC also spar over the purpose of § 1692k(a)(3). Brief for Petitioner 14-16; Brief for United States 21-28; Reply Brief 11-14; Brief for Respondent 30-43. Marx and the United States contend that Congress intended to limit a court’s discretion to award costs to prevailing defendants because FDCPA plaintiffs are often poor and may be deterred from challenging unlawful debt collection practices by the possibility of being held liable for the defendant’s costs. This purposive argument cannot overcome the language and context of § 1692k(a)(3), but even if it could, we find it unpersuasive. Rule 54(d)(1) does not require courts to award costs to prevailing defendants. District courts may appropriately consider an FDCPA plaintiff’s indigency in deciding whether to award costs. See Badillo v. Central Steel & Wire Co., 717 F. 2d 1160, 1165 (CA7 1983) (“[I]t is within the discretion of the district court to consider a plaintiff’s indigency in denying costs under Rule 54(d)”).