SYKES

Brendan Shiller, Shiller Preyar, Chicago, IL, for Plaintiff–Appellant. David Decelles, City of Chicago Law Department, Chicago, IL, for Defendants–Appellees.
Before POSNER, FLAUM, and SYKES, Circuit Judges.

SYKES, Circuit Judge.

Andy Montanez won a small jury verdict—just $2,000—in this straightforward excessive-force case against two Chicago police officers, but his lawyers racked up enormous legal fees during the course of the litigation. Invoking the fee-shifting statute applicable in civil-rights cases, 42 U.S.C. § 1988(d), they submitted a request for more than $400,000 in fees, but the district court awarded only a fraction of that amount. The award will be paid by the City of Chicago, and although it's much less than requested, it's still a huge sum—almost $110,000—in part because the City adopted a scorched-earth defense strategy. This simple civil-rights claim, overlitigated by both sides, took on all the protracted complexity of high-stakes commercial litigation, replete with hardfought discovery battles and even a mock trial.

The main issue on appeal is whether the City should be required to pay a larger portion of Montanez's legal fees than the district court ordered. The court trimmed the fee request by striking hours as unnecessary or improperly documented, reducing the lawyers' billing rates, and slashing the resulting “lodestar” figure in half to account for the limited success on the merits.

We review an award of attorney's fees deferentially. Trial judges are in a better position to determine what fees are “reasonable” in a given case. See§ 1988(b) (allowing “a reasonable attorney's fee” to the prevailing party in a successful suit under 42 U.S.C. § 1983). This is especially true when the plaintiff is only partially successful; setting a reasonable fee for limited success is necessarily imprecise. We find no abuse of discretion in the district court's painstaking analysis of the billing records or in its lodestar reduction; the court's approach reflects the perfectly sensible conclusion that the case was overstaffedand much of the billed time was unjustified.

We take this opportunity to remind trial judges that where a fee-shifting statute is in play, the court has the opportunity and the discretion to check runaway attorney's fees while the litigation is underway, not just when reviewing a fee request after the case has concluded. Early and active use of the court's case-management authority can help prevent excessive fees before they accrue.

I. Background

Andy Montanez sued the City of Chicago and Police Officers Vincent Fico and Joseph Simon, alleging that Fico used excessive force while arresting him (for drinking on a public way) and Simon failed to intervene to stop it. He sustained minor injuries during the arrest, for which he sought damages under 42 U.S.C. § 1983 (for violation of his Fourth Amendment rights) and several state-law theories. He also brought a claim against the City based on its duty to indemnify the police officers. The case was assigned to a magistrate judge, see28 U.S.C. § 636(c), and the state-law claims were dismissed as time-barred. The City conceded its obligation to indemnify, so only the § 1983 claims proceeded to trial. Officer Fico was found liable, Officer Simon was cleared, and the jury awarded $1,000 in compensatory damages and $1,000 in punitive damages.

Having secured this $2,000 judgment for their client, Montanez's lawyers submitted a bill for more than $426,000 in attorneys' fees and about $6,500 in costs and expenses. The City challenged most of the request as unreasonable. The judge resolved the dispute by meticulously scrutinizing the bill line by line, striking entries that she determined were unnecessary, duplicative, excessive, or improperly documented.

Seven lawyers had billed approximately 1,021 hours on the case. Three of the lawyers were only tangentially involved, cumulatively billing less than 10 hours; the judge excluded their time entirely. Carefully reviewing the remaining hours, the judge discounted entries where more than one partner oversaw the same activities, or where the lawyers researched or drafted motions that were never filed. She also excluded the hours the lawyers spent on a full-day mock trial. She struck entries related to matters that were essentially administrative—such as time spent deciding which attorneys would handle the lawsuit and hours billed for “trying to find” Montanez—on the rationale that the City should not be billed for these case-management problems. The judge discounted or disallowed other vaguely or improperly billed entries, such as time spent by partners on tasks that could have been delegated to paralegals, unspecified “call[s] to client” and his family, and more than three hours a partner billed while shopping for clothes for a witness.

The judge also reduced the hourly billing rates. The two lead lawyers—partners in the firm with nine and thirteen years' experience, respectively—sought a rate of $400 per hour for the first two years of work on the case, $425 per hour for the third year, and $450 per hour for the last year. The judge concluded that these rates could not be justified by reference to the billing rates of comparably qualified lawyers in the Chicago market for § 1983 litigation. After conducting her own assessment of the market, the judge settled on an hourly rate of $385 for the two lead lawyers. Similarly, the judge set a rate of $175 for the firm's second- and third-year associates rather than the requested rates between $250 and $300 per hour. Multiplying each attorney's allowed hours by these lower rates produced a total adjusted bill of $217,110.50.

The judge then reduced this figure in two ways. First, and less significantly, the lawyers had requested $881.25 less than the original billing totals supported, so the judge deducted that amount from the adjusted total. Second, and more importantly, the judge concluded that the limited success on the merits warranted a substantial reduction in the fee award. Noting that Montanez lost four of his six claims and won just $2,000 in damages, and also that the case lacked public or precedential importance, the judge reduced the lodestar amount by 50% and awarded $108,350.87 in fees.

The judge then turned to the bill of costs and again addressed each entry individually. The judge struck entries for unnecessary expenses, like fees for a witness who was never subpoenaed, copies of deposition transcripts for witnesses who were never called, and costs of printing untimely discovery requests. In addition, Montanez's lawyers had paid more for certain transcripts than allowed under the local rules, so the judge capped recovery for these transcripts at the highest per-page rate permitted. For some of these documents, the lawyers failed to supply the page lengths; the judge disallowed these requests outright. After these reductions, the final award of costs was $3,051.94, down from an initial request of $4,696.84.

II. Discussion

Montanez's lawyers challenge just about every dime the district court disallowed. One example will illustrate the lack of perspective that pervades this case. The lawyers dedicate an entire section of their opening brief to claims of error in the calculation of printing costs, amounting to at most $35.20—about 0.0082% of the approximately $430,000 they hope to recover. No reasonable client would countenance spending even a tenth of an hour arguing over $35. 20. The willingness to fight so hard for so little goes a long way toward explaining why there is a $426,000 bill for attorneys' fees in a $2,000 case.

The lack of private restraint in this case underscores a point that deserves brief attention before we address the substance of this appeal. Trial judges have substantial case-management authority to control the course of litigation in their courts. In cases lacking private incentives to limit the scope of litigation, active judicial oversight can help prevent straightforward cases like this one from spiraling out of control. The Federal Rules of Civil Procedure authorize judges to monitor and influence the scope of litigation. See, e.g., Fed.R.Civ.P. 16 (authorizing courts to order pretrial conferences to “establish[ ] early and continuing control” and to “discourag[e] wasteful pretrial activities”); Fed. R. Civ. P. 26(b)(2)(C) (requiring the court to limit discovery if “the burden or expense ... outweighs its likely benefit, considering the needs of the case”). As we will explain, trial judges have broad discretion to adjust bloated bills for attorney's fees after the fact, but judicious use of the court's case-management authority during the litigation can also help to check overlawyering. Reasonable limits on discovery and trial preparation can effectively channel the efforts of counsel before excessive time and resources are expended.

A. Attorney's Fees Under § 1988(b)

We review an award of attorney's fees under a “highly deferential abuse of discretion standard.” Pickett v. Sheridan Health Care Ctr., 664 F.3d 632, 639 (7th Cir.2011) (quoting Estate of Borst v. O'Brien, 979 F.2d 511, 514 (7th Cir.1992)). Section 1988(b) allows prevailing parties in § 1983 litigation to recover “a reasonable attorney's fee,” and the district court is in the best position to make the “contextual and fact-specific” assessment of what fees are reasonable. Sottoriva v. Claps, 617 F.3d 971, 975 (7th Cir.2010). The Supreme Court has consistently held that the interest in uniformity of fee awards is not great enough to justify closer appellate scrutiny—indeed, the Court has said that there is hardly any “sphere of judicial decisionmaking in which appellate micromanagement has less to recommend it.” Fox v. Vice, ––– U.S. ––––, 131 S.Ct. 2205, 2216, 180 L.Ed.2d 45 (2011).

That doesn't mean that the district court's discretion is boundless. The court “must apply the correct standard,” id., and “bears the responsibility of justifying its conclusions,” Sottoriva, 617 F.3d at 975. To this end, the district court generally begins the fee calculation by computing a “lodestar”: the product of the hours reasonably expended on the case multiplied by a reasonable hourly rate. See id. Although the lodestar yields a presumptively reasonable fee, Pickett, 664 F.3d at 639, the court may nevertheless adjust the fee based on factors not included in the computation, see Hensley v. Eckerhart, 461 U.S. 424, 434, 103 S.Ct. 1933, 76 L.Ed.2d 40 (1983).

Perhaps the most important of these factors is the degree of success on the merits, especially “where a plaintiff is deemed ‘prevailing’ even though he succeeded on only some of his claims for relief.” Id. Where the hours spent on successful claims can easily be distinguished from those spent on unsuccessful claims, the court can simply strike the latter entries when computing the lodestar. But when the work on each set of claims is difficult to disentangle, the lodestar calculation will likely include some time spent on unsuccessful claims, and in these cases the court ought to reduce the lodestar figure. Ultimately, the guiding inquiry is whether “the plaintiff achieve[d] a level of success that makes the hours reasonably expended a satisfactory basis for making a fee award.” Id.

B. The Lower Hourly Rates

A reasonable hourly rate is based on the local market rate for the attorney's services. Pickett, 664 F.3d at 640. The best evidence of the market rate is the amount the attorney actually bills for similar work, but if that rate can't be determined, then the district court may rely on evidence of rates charged by similarly experienced attorneys in the community and evidence of rates set for the attorney in similar cases. See Johnson v. GDF, Inc., 668 F.3d 927, 933 (7th Cir.2012). The party seeking a fee award bears the burden of establishing the market rate for the work; if the lawyers fail to carry that burden, the district court can independently determine the appropriate rate. See id.

Montanez's lawyers tried to establish hourly rates between $400 and $450 for the two most experienced lawyers on the team and between $250 and $300 for the other attorneys. The primary support for these rates was a collection of the attorneys' retainer agreements with other clients. Most of these were contingency agreements, however, and the judge reasonably found this evidence unhelpful. Although the clients “acknowledged” that the attorneys would charge up to $450 per hour, the agreed-upon rate was really a percentage of the winnings (usually 40%); clients would only pay the $450 rate if they changed lawyers before final judgment. The judge did not abuse her discretion by giving little weight to these agreements as evidence of market hourly rates for the attorneys' services. See Pickett, 664 F.3d at 640 (recognizing “the difficulty of determining the hourly rate of an attorney who uses contingent fee agreements”).

The judge also disregarded other agreements offered by Montanez's lawyers because they involved different markets—namely, the markets for immigration, contract, and criminal-law services. If the record contained evidence that the prevailing market rate for this type of legal work was similar to rates for representation in civil-rights cases, we might question the district court's decision to distinguish between these markets. See Johnson, 668 F.3d at 933 (“It was an abuse of discretion for the district court to decide that the market must distinguish between FLSA and Title VII cases. Either it does or it doesn't, but it is not the court's job to say that it should.”). Without such evidence, however, “[t]he district court's discretion in determining what is a reasonable attorney's fee applies to its determination of what constitutes a market,” Moriarty v. Svec, 233 F.3d 955, 966 (7th Cir.2000), and that discretion was not abused here.

The court also questioned other evidence Montanez's lawyers submitted to support their requested rates, including the so-called Laffey Matrix and affidavits from other lawyers in Chicago. The Laffey Matrix is a chart of hourly rates published by the U.S. Attorney's Office for the District of Columbia, which some circuits use to help determine a reasonable fee under fee-shifting statutes. We've expressed some skepticism about applying the Laffey Matrix outside Washington, D.C., and have left it to trial judges to exercise their discretion in evaluating its usefulness in any particular case. See Pickett, 664 F.3d at 649–51. Here, the judge properly considered the Laffey Matrix and in her discretion found it unpersuasive in this case. Similarly, the court disregarded the affidavits from other Chicago attorneys, which only attested that Montanez's lawyers were well qualified and that their fees were reasonable. We've held that conclusory affidavits from attorneys “merely opin [ing]” on the reasonableness of another attorney's fee—unlike affidavits describing what “comparable attorneys charge for similar services”—have little probative value. Id. at 647. The judge properly disregarded this evidence.

With little record evidence to support the requested rates, the judge looked to the fees awarded in some of the lawyers' previous cases and the publicly available information about rates charged for similar work in the community. The prior fee awards varied considerably, from $225 per hour in one case to $400 in another; both rates were based on compromises between parties, so the judge discounted their probative value. The judge also considered rates awarded to similarly experienced Chicago attorneys in other civil-rights cases in the district. This “next-best” evidence was properly considered after the court found insufficient evidence of the attorneys' actual market rates. See Johnson, 668 F.3d at 933.

Weighing all this evidence, the court settled on an hourly rate of $385 for the two more experienced lawyers on the case and $175 for the two junior associates. Those rates are within the upper middle of the range supported by the evidence from the lawyers' past cases and the rates awarded to other Chicago attorneys in civil-rights cases. We find no abuse of discretion.

C. The Reduced Hours

The judge found that dozens of the hours billed should not be included in the lodestar computation. On appeal Montanez's lawyers challenge nearly every one of these deductions. But they also argue that the court lacked power to strike any of the entries because the City listed its objections line by line on a spreadsheet of the bills instead of describing its objections in a legal memorandum. We can quickly put this argument to rest: A district court may strike billing entries so long as the party requesting fees has an opportunity to respond to any objections. Cf. Spellan v. Bd. of Educ., 59 F.3d 642, 646 (7th Cir.1995) (holding that a court may not strike billing entries sua sponte without giving the party an opportunity to defend them). The City's line-by-line objections were admittedly brief, but they were specific enough to allow a response, and the format was considerably more practical and economical than a long memorandum detailing each objection.

Regarding the specific exclusions, we note that “[t]he essential goal in shifting fees (to either party) is to do rough justice, not to achieve auditing perfection.” Fox, 131 S.Ct. at 2216. We decline the plaintiff's invitation to reexamine every line-item deduction. The district court is in a better position to evaluate the reasonableness of specific requests, and fee-shifting “should not result in a second major litigation.” Id. (quoting Hensley, 461 U.S. at 437, 103 S.Ct. 1933). Our job is to determine whether the district court applied the correct standards and avoided arbitrary decisionmaking, not to decide for ourselves how much litigation was truly reasonable.

The judge concluded that the case had been overlitigated and on that basis carefully scrutinized the billing records to exclude any time that was unnecessary, duplicative, or insufficiently documented. To take just a few examples, she disallowed time spent by two partners simultaneously doing the same thing, explaining that on such a straightforward case, one partner was enough. She concluded that other hours were wasted, like time spent drafting untimely requests for admission that were never filed and time spent researching a Freedom of Information Act claim as an alternative to simple discovery. It's not unexpected that some legal research will prove fruitless on an ultimately successful claim, and a prevailing party may in appropriate circumstances recover for time spent going down roads that seemed promising but turn out to be dead ends. See Kurowski v. Krajewski, 848 F.2d 767, 776 (7th Cir.1988). But the district court has the discretion to decide what research was likely to contribute to a successful claim, and here the court reasonably determined that many of the lawyers' projects were needlessly esoteric in the context of such a simple case. Montanez's lawyers belatedly offer an explanation for some of this research (for example, the time they spent researching the Prison Litigation Reform Act), but the arguments are new on appeal and do not establish that the judge abused her discretion.

The judge also deducted some hours because they were improperly or inadequately recorded. For example, the court rejected time billed for vaguely described phone calls; hours billed as “call to client,” without more, were disallowed. The judge disallowed time billed for clerical work that was nonetheless recorded at an attorney'srate. For example, some scanning and faxing was billed at $450 per hour because a partner did the work; one associate billing at $275 per hour spent many hours doing nothing but filing motions in limine. The district court has broad discretion to strike such vague or unjustified billing entries. See, e.g., Harper v. City of Chicago Heights, 223 F.3d 593, 605 (7th Cir.2000). Appellate deference to the trial judge's discretionary judgment in striking specific billing entries encourages candor in fee requests and relieves the burden on district courts faced with vague or poorly documented fee claims.

D. The Lodestar Reductions

Finally, Montanez's lawyers challenge the judge's two adjustments to the lodestar: the $881.25 deduction to account for a discrepancy between the bills and the fee request, and the 50% reduction to reflect Montanez's limited success on the merits. The discrepancy is hard to understand. The judge added up all of the hours billed, multiplied them by the requested rates, and found that the lawyers had asked for $881.25 less than the billing records might have supported. Montanez's attorneys haven't explained this discrepancy on appeal; instead they submitted a new chart with new arithmetic errors. Our own review of the bills reveals that the discrepancy was caused by occasionally billing attorneys at a lower rate—below even the reduced rates set by the district court. For example, on several occasions the more experienced attorneys were billed at $225 per hour instead of the $450 they ultimately requested. We don't know whether this was intentional or an oversight, but it was not an abuse of discretion to reduce the adjusted lodestar calculation by the amount of the discrepancy.

More significantly, the judge did not abuse her discretion by slashing the lodestar by 50%. In setting a reasonable fee, the district court must determine whether “the plaintiff achieve[d] a level of success that makes the hours reasonably expended a satisfactory basis for making a fee award.” Hensley, 461 U.S. at 434, 103 S.Ct. 1933. A plaintiff who achieves “excellent results” should receive the entire lodestar, but where “a plaintiff has achieved only partial or limited success,” the lodestar “may be an excessive amount.” Id. at 435–36, 103 S.Ct. 1933.

We do not need to belabor the point: Montanez did not achieve “excellent results.” The district court properly noted that he lost four of his six claims, including all of his claims against Officer Simon. Montanez's lawyers argue that some of the claims were simply different routes to achieving the same recovery. A plaintiff who brings multiple theories to remedy the same harm “is not to be penalized just because some, or even all but one, are rejected, provided that the one or ones that succeed give him all that he reasonably could have asked for.” Lenard v. Argento, 808 F.2d 1242, 1246 (7th Cir.1987). But Montanez did not get all that he asked for—far from it. His failure to prevail on the claims against Simon, for instance, foreclosed additional punitive damages against another defendant. And even on the claims Montanez won, his success was severely limited: just $2,000 in damages, a minimal victory in light of the time expended to achieve it.

We don't mean to suggest that Montanez's victory was purely nominal, in which case he would not be entitled to attorney's fees at all. See Farrar v. Hobby, 506 U.S. 103, 115, 113 S.Ct. 566, 121 L.Ed.2d 494 (1992). Montanez won a meaningful sum and established both a compensable injury and an entitlement to punitive damages against one of the officers. Nonetheless, we agree with the district court that his limited success meant that the lodestar was not “a satisfactory basis for making a fee award.” Hensley, 461 U.S. at 434, 103 S.Ct. 1933.

“No algorithm is available” for adjusting a lodestar to reflect partial or limited success. Richardson v. City of Chicago, 740 F.3d 1099, 1103 (7th Cir.2014). When the judge cannot easily separate the successful and unsuccessful work, “there is nothing to do but make an across-the-board reduction that seems appropriate in light of the ratio between winning and losing claims.” Id. Here again, the district court has broad discretion to determine the appropriate reduction. The trial judge is in a better position to assess whether the unsuccessful claims were important or trivial; whether a $2,000 judgment is a spectacular success, a dismal failure, or something in between; and whether the plaintiff's lawyers would have spent substantially less time on the case had they been more realistic. As the magistrate judge properly noted, “a fee request that dwarfs the damages award might raise a red flag.” Anderson v. AB Painting & Sandblasting Inc., 578 F.3d 542, 546 (7th Cir.2009). The disparity between the damages and the fee request could not be more striking here.

The judge did not make the mistake of limiting the fee to some multiple of the judgment, which would have been reversible error. See id. at 545. Instead, after finding that the attorneys' expenditure of time could not be explained by the complexity of the facts or the relevant legal doctrine, or by the vindication of an important public interest, see City of Riverside v. Rivera, 477 U.S. 561, 574, 579, 106 S.Ct. 2686, 91 L.Ed.2d 466 (1986) (plurality opinion), the judge treated the disproportionate fees as an indicator that Montanez and his attorneys unrealistically believed these claims were worth far more than they recovered. The 50% reduction was not an abuse of discretion, and the final fee award of $108,350.87 was quite “generous in relation to [Montanez's] recovery.” Richardson, 740 F.3d at 1103.

E. The Bill of Costs

Rule 54(d) of the Federal Rules of Civil Procedure creates a strong presumption that the prevailing party will be awarded those costs of litigation identified in 28 U.S.C. § 1920. See U.S. Neurosurgical, Inc. v. City of Chicago, 572 F.3d 325, 333 (7th Cir.2009). Recoverable costs include “[f]ees for ... transcripts necessarily obtained for use in the case,” § 1920(2), “[f]ees ... for printing,” § 1920(3), and “the costs of making copies of any materials where the copies are necessarily obtained for use in the case,” § 1920(4). Of the $4,696.84 in costs requested, the district court awarded $3,051.94. Montanez's lawyers challenge the court's deductions.

As we've noted, the lawyers insist that the court erroneously deducted $35. 20 in printing fees, including $22.10 for copying certain deposition transcripts, $3.10 for copying improper requests to admit, and $10 that the court never explained. Their calculations are incorrect. They requested $768.70 in printing fees and received $743.50, so the court deducted $25. 20. As far as we can tell, there was no unexplained $10 deduction.

Turning to the actual deductions, the court struck $22.10 for copies of deposition transcripts of three witnesses who were never identified as potential witnessesat trial. Montanez's lawyers already had the original transcripts, and the district court concluded that they made copies purely for their own convenience. Only fees for copies “necessarily obtained for use in the case” are recoverable, 28 U.S.C. § 1920(4), and the district court has discretion to determine which copies were necessary, Kulumani v. Blue Cross Blue Shield Ass'n, 224 F.3d 681, 685 (7th Cir.2000). The judge disallowed the cost of copying transcripts of witnesses who were never going to be called at trial; that was a sensible decision.

The lawyers also challenge a $3.10 deduction for copies of untimely requests to admit. (Why, we do not know. This lack of perspective is precisely what caused the litigation costs and fees to spiral so wildly out of control.) The judge had already held that the lawyers “should have known better than to file these untimely requests to admit without leave of court” and struck all hours related to drafting these requests. We cannot see how the court abused its discretion in denying costs for copies of improperly filed requests to admit; they could hardly be characterized as “necessarily obtained for use in the case.” 28 U.S.C. § 1920(4).

Finally, the judge denied a portion of the request for transcript fees because in several instances the lawyers paid more for transcripts than the maximum allowed under Local Rule 54.1(b) in the Northern District of Illinois. The rule provides that the cost of any transcript may not exceed the rate set by the United States Judicial Conference unless a higher rate was previously established by court order. Montanez's lawyers argue, for the first time on appeal, that the City chose to use a court reporter who charged more than the local rule allows and that they had to order copies from that reporter; thus, they had no choice but to pay the higher rates. The City responds that Montanez could have brought his own court reporter to the depositions. That's an absurd suggestion, but in the end we don't need to resolve the conflict about compliance with the local rule.

Had Montanez's lawyers explained the problem to the judge, she might have allowed the full cost of the transcripts. Several decisions from the Northern District of Illinois sensibly suggest that Local Rule 54.1(b) does not apply when the party who must bear the costs selected the court reporter—in other words, whoever picked the reporter can't later object to that reporter's rates. See, e.g., Gyrion v. City of Chicago, 454 F.Supp.2d 725, 726 (N.D.Ill.2006). Moreover, the local rule allows parties to seek a court order authorizing higher transcript fees. Montanez's lawyers never sought such an order and never argued that Local Rule 54.1 was improperly applied in this case.

Instead they now argue that the cost of any transcripts not recoverable under Local Rule 54.1 and 28 U.S.C. § 1920(2) should be recoverable as part of the “reasonable attorney's fee” under 42 U.S.C.1988(b). We disagree. Section 1920(2) allows parties to recover only the transcript expenses that can be considered reasonable, that is, “[f]ees for ... transcripts necessarily obtained for use in the case.” Local Rule 54.1 operates as a limit on the amount of fees that will be considered “necessary” for obtaining transcripts within the Northern District of Illinois. The rule might be flawed, but Montanez's lawyers chose not to challenge it. They can't get around the rule now by arguing that § 1988 covers transcript expenses beyond the “[f]ees for ... transcripts necessarily obtained” already covered by § 1920(2).

For most of these transcripts, the judge simply revised the requests downward to bring them within the limits of Local Rule 54.1 by applying the proper per-page rate. In some cases, however, the lawyers failed to document the length of certain transcripts, so revision was impossible and the court simply disallowed the costs entirely. Montanez's lawyers now complain that the court could have determined the page lengths itself by finding the documents in the record. But it was not the judge's responsibility to make up for the lawyers' lack of documentation. Cf. Harper, 223 F.3d at 605 (“[W]hen a fee petition is ... inadequately documented, a district court may ... strike the problematic entries....”). We find no abuse of discretion in the district court's award of costs.

Affirmed.