KOZINSKI, Chief Judge:
We consider various issues pertaining to the district court's award of attorneys' fees under 42 U.S.C. § 1988.
Moreno sued the City of Sacramento and several other defendants, alleging that they violated his civil rights by seizing and destroying his property without due process. After lengthy pre-trial proceedings and a previous appeal, a jury awarded Moreno $717,000 in compensatory and punitive damages. Moreno's principal trial counsel, Andrea Miller, sought an award of attorneys' fees under 42 U.S.C. § 1988. Miller requested $704,858.07 for herself and her staff, including compensation for 1,973.6 hours of her own time, at a rate of $300 per hour. This request excluded around 9 percent of the total hours actually spent on the case.
The district court reduced the hours further, concluding that around a quarter to a third of the time spent on research, appeal and trial preparation and half the time spent on investigation was unnecessary.
Lawyers must eat, so they generally won't take cases without a reasonable prospect of getting paid. Congress thus recognized that private enforcement of civil rights legislation relies on the availability of fee awards: "If private citizens are to be able to assert their civil rights, and if those who violate the Nation['s] fundamental laws are not to proceed with impunity, then citizens must have the opportunity to recover what it costs them to vindicate these rights in court." S.Rep. No. 94-1011, at 2 (1976), as reprinted in 1976 U.S.C.C.A.N. 5908, 5910.
In making the award, the district court must strike a balance between granting sufficient fees to attract qualified counsel to civil rights cases, City of Riverside v. Rivera, 477 U.S. 561, 579-80, 106 S.Ct. 2686, 91 L.Ed.2d 466 (1986), and avoiding a windfall to counsel, see Blum v. Stenson, 465 U.S. 886, 897, 104 S.Ct. 1541, 79 L.Ed.2d 891 (1984) (quoting S.Rep. No. 94-1011, at 6 (1976)). The way to do so is to compensate counsel at the prevailing rate in the community for similar work; no more, no less.
In this case, the district court used the lodestar method to calculate fees. Under this method, a district court must start by determining how many hours were reasonably expended on the litigation, and then multiply those hours by the prevailing local rate for an attorney of the skill required to perform the litigation. See Blum, 465 U.S. at 895, 104 S.Ct. 1541. The district court may then adjust upward or downward based on a variety of factors. Hensley, 461 U.S. at 434, 103 S.Ct. 1933. The number of hours to be compensated is calculated by considering whether, in light of the circumstances, the time could reasonably have been billed to a private client. Id. We review the district court's calculation of the reasonable hours and hourly rate for abuse of discretion. See Camacho v. Bridgeport Fin., Inc., 523 F.3d 973, 977-78 (9th Cir.2008).
When the district court makes its award, it must explain how it came up with the amount. The explanation need not be elaborate, but it must be comprehensible. As Hensley described it, the explanation must be "concise but clear." 461 U.S. at 437, 103 S.Ct. 1933 (emphasis added). Where the difference between the lawyer's request and the court's award is relatively small, a somewhat cursory explanation will suffice. But where the disparity is larger, a more specific articulation of the court's reasoning is expected. See Bogan v. City of Boston, 489 F.3d 417, 430 (1st Cir.2007). We review the legal principles underlying the fee award de novo. Ferland v. Conrad
The court may reduce the number of hours awarded because the lawyer performed unnecessarily duplicative work, but determining whether work is unnecessarily duplicative is no easy task. When a case goes on for many years, a lot of legal work product will grow stale; a competent lawyer won't rely entirely on last year's, or even last month's, research: Cases are decided; statutes are enacted; regulations are promulgated and amended. A lawyer also needs to get up to speed with the research previously performed. All this is duplication, of course, but it's necessary duplication; it is inherent in the process of litigating over time. Here, there was a previous appeal (of the district court's grant of summary judgment) which would have added to the delay and rendered much of the research stale. One certainly expects some degree of duplication as an inherent part of the process. There is no reason why the lawyer should perform this necessary work for free.
It must also be kept in mind that lawyers are not likely to spend unnecessary time on contingency fee cases in the hope of inflating their fees. The payoff is too uncertain, as to both the result and the amount of the fee. It would therefore be the highly atypical civil rights case where plaintiff's lawyer engages in churning. By and large, the court should defer to the winning lawyer's professional judgment as to how much time he was required to spend on the case; after all, he won, and might not have, had he been more of a slacker.
The district court has a greater familiarity with the case than we do, but even the district court cannot tell by a cursory examination which hours are unnecessarily duplicative. Nevertheless, the district court can impose a small reduction, no greater than 10 percent — a "haircut" — based on its exercise of discretion and without a more specific explanation. Here, however, the district court cut the number of hours by 25 percent, and gave no specific explanation as to which fees it thought were duplicative, or why. While we don't require the explanation to be elaborate, it must be clear, and this one isn't. Plaintiff's counsel had already cut her fees by 9 percent, so an additional 25 percent cut would amount to over one third. The court has discretion to make such an adjustment, but we cannot sustain a cut that substantial unless the district court articulates its reasoning with more specificity. We therefore conclude that the district court's explanation is insufficient to sustain a 25 percent cut based on duplication.
Plaintiff also requested fees for 266.6 hours of preparation for the first two trial dates, July 2002 and February 2005, without indicating how much time was spent preparing for each date. Plaintiff requested fees for 340.7 hours for the third trial date, May 2005. The district court awarded the full hours for the third trial date, but reduced the hours for the first two dates by half, to 133.3. As with the research hours, the cut here is substantial, amounting to 20 percent of the total fees billed for trial, in addition to the 9 percent already cut by plaintiff's counsel.
The district court did not explain the necessity or degree of the cut, other than to say that the amount of time plaintiff's
Of course, the court might have some specific reason for believing that work is excessive or duplicative, but it must explain why. We cannot sustain a 50 percent cut, over and above the 9 percent cut plaintiff's counsel already imposed on herself, without a clear explanation that we can review. The opaque explanation provided here is an insufficient basis for the district court's Draconian cut.
The district court awarded fees for 180 hours of time spent preparing the earlier appeal. Plaintiff requested 269.3 hours. We "look more closely" at fee awards involving appeals, Suzuki v. Yuen, 678 F.2d 761, 762-63 (9th Cir.1982), and can find no justification for a cut of 33 percent, on top of plaintiff's counsel's own cut. The district court noted that plaintiff's counsel spent twice as long on the appeal than on the summary judgment, but this does not mean the additional time spent on appeal was unjustified; after all, plaintiff lost claims at summary judgment that he won on appeal. More fundamentally, preparing summary judgment motions and appeals are not commensurate tasks, though they have some elements in common. What matters is whether spending more time winning on appeal than losing on summary judgment was an imprudent use of hours. The district court points to nothing to support the conclusion that it was.
Cutting fees for "duplication of effort" appears to have been an easy way for the district court to reduce an award it may have felt was too high. But if the court believes the overall award is too high, it needs to say so and explain why, rather than making summary cuts in various components of the award. While we accord deference to the district court's explanation of why a requested fee is excessive, we can only do so if the district court provides an explanation that we can meaningfully review. Findings of duplicative work should not become a shortcut for reducing an award without identifying just why the requested fee was excessive and by how much. As the reduction passes well beyond the safety zone of a haircut, which plaintiff's counsel seems to have given herself already, the district court's justification for the cuts must be weightier and more specific.
But the district court went on to consider impermissible factors. The district court reduced the hourly rate from $300 an hour to $250 an hour, in part because it thought that other firms could have staffed the case differently. The court speculated that other firms would have used a less skilled attorney, rather than the lead counsel, to perform document review. While it is appropriate to consider the skill required to perform a task, Hensley, 461 U.S. at 430 n. 3, 103 S.Ct. 1933, the district court may not set the fee based on speculation as to how other firms would have staffed the case.
The cost effectiveness of various law firm models is an open question,
The district court's inquiry must be limited to determining whether the fees requested by this particular legal team are justified for the particular work performed and the results achieved in this particular case. The court may permissibly look to the hourly rates charged by comparable attorneys for similar work, but may not attempt to impose its own judgment regarding the best way to operate a law firm, nor to determine if different staffing decisions might have led to different fee requests. The difficulty and skill level of the work performed, and the result achieved—not whether it would have been cheaper to delegate the work to other attorneys—must drive the district court's decision.
The court also erred by applying what appears to be a de facto policy of awarding a rate of $250 an hour to civil rights cases. At the fees hearing, the district court noted that "300 an hour is a fairly big step for me, and I think for the court generally" and that "the court has pretty much held the line at 250 [an hour] for the past ten years." While the district court's final fee order does not reiterate this reasoning, an effort to adhere to this de facto policy probably influenced the final rate awarded, which was $250 an hour. Nothing else supports the $50 an hour reduction.
District judges can certainly consider the fees awarded by other judges in the same locality in similar cases. But adopting a court-wide policy—even an informal one—of "holding the line" on fees at a certain level goes well beyond the discretion of the district court. One problem with any such policy is that it becomes difficult to revise over time, as economic conditions change; here the rate apparently hadn't changed for 10 years, and even a $50 increase in the hourly rate was considered a "big step ... for the court generally." Unless carefully administered and updated, any such policy becomes a strait-jacket. More fundamentally, such a policy—no matter how well intentioned or administered—is inconsistent with the methodology for awarding fees that the Supreme Court and our court has adopted. The district court's function is to award fees that reflect economic conditions in the district; it is not to "hold the line" at a particular rate, or to resist a rate because it would be a "big step." If the lodestar leads to an hourly rate that is higher than past practice, the court must award that rate without regard to any contrary practice.
The district further erred by double counting the reduction in hourly rate for some tasks; such double counting is impermissible. See Cunningham v. County of Los Angeles, 879 F.2d 481, 489 (9th Cir.1988). It is possible, of course, for a district court to reduce both the hours and hourly rate awarded for some tasks. But the district court must exercise extreme care in making such reductions to avoid double counting. Here, the district court reduced the reasonable hourly rate from $300 to $75 an hour, the paralegal
* * *
The district court has discretion to determine the appropriate fee award, because its familiarity with the case allows it to distinguish reasonable from excessive fee requests. But gut feelings are not enough; if the district court is going to make substantial cuts to a winning lawyer's fee request, it needs to explain why with sufficient specificity that the lawyer can meaningfully object and we can meaningfully review the objection. We can't defer to reasoning that we can't review; if all the district court offers is a conclusory statement that a fee request is too high, then we can't tell if the court is applying its superior knowledge to trim an excessive request or if it is randomly lopping off chunks of the winning lawyer's reasonably billed fees.
We are well aware that awarding attorneys' fees to prevailing parties in civil rights cases is a tedious business. And it may be difficult for the district court to identify the precise spot where a fee request is excessive. But the burden of producing a sufficiently cogent explanation can mostly be placed on the shoulders of the losing parties, who not only have the incentive, but also the knowledge of the case to point out such things as excessive or duplicative billing practices. If opposing counsel cannot come up with specific reasons for reducing the fee request that the district court finds persuasive, it should normally grant the award in full, or with no more than a haircut.
The district court's fee award is vacated and the case is remanded with instructions that the court enter a new fee award consistent with this opinion.
Other courts have adhered more to the pyramid structure in measuring fee awards. See, e.g., Loughner v. Univ. of Pittsburgh, 260 F.3d 173, 180 (3d Cir.2001) ("A claim by a lawyer for maximum rates for telephone calls with a client, legal research, a letter concerning a discovery request, the drafting of a brief, and trial time in court is neither fair nor reasonable. Many of these tasks are effectively performed by administrative assistants, paralegals, or secretaries."); Bee v. Greaves, 669 F.Supp. 372, 377 (D.Utah 1987), rev'd in part on other grounds, 910 F.2d 686 (10th Cir. 1990) (reducing overall rate awarded because less experienced attorneys could have performed much of the work); Mautner v. Hirsch, 831 F.Supp. 1058, 1076 (S.D.N.Y. 1993), rev'd in part on other grounds, 32 F.3d 37 (2d Cir. 1994) (reducing lodestar for using senior attorneys when junior attorneys and paralegals were available).