JOHN M. WALKER, JR., Circuit

Mitchell A. Karlan (Mark E. Bini and Michelle Craven, on the brief), Gibson, Dunn & Crutcher, LLP, New York, New York, for Plaintiffs-Appellants.

Thomas J. O'Connor, Napierski, Vandenburgh & Napierski, LLP, Albany, New York, for Defendants-Appellees.

Before: JACOBS, Chief Judge, WALKER, Circuit Judge, O'CONNOR, Associate Justice Retired.

JOHN M. WALKER, JR., Circuit Judge.

In this appeal from the district court's disposition of their motion for an award of attorney's fees, plaintiffs-appellants ("plaintiffs"), who prevailed in a suit brought under the Voting Rights Act of 1965 ("VRA"), seek a recalculation of the amount that they may recoup. The fee — historically known as the "lodestar" — to which their attorneys are presumptively entitled is the product of hours worked and an hourly rate. Plaintiffs argue that the district court applied an unnecessarily strict "forum rule": The district court, they contend, required them to show extraordinary special circumstances before it would use in its "lodestar" calculation an hourly rate greater than the hourly rate charged by attorneys in the district where the district court sits.

We agree that the district court may have applied the forum rule in too unyielding a fashion. We therefore clarify its proper application in this circuit: While the district court should generally use the prevailing hourly rate in the district where it sits to calculate what has been called the "lodestar" — what we think is more aptly termed the "presumptively reasonable fee" — the district court may adjust this base hourly rate to account for a plaintiff's reasonable decision to retain out-of-district counsel, just as it may adjust the base hourly rate to account for other case-specific variables.

Moreover, this dispute concerning the "forum rule" is but a symptom of a more serious illness: Our fee-setting jurisprudence has become needlessly confused — it has come untethered from the free market it is meant to approximate. We therefore suggest that the district court consider, in setting the reasonable hourly rate it uses to calculate the "lodestar," what a reasonable, paying client would be willing to pay, not just in deciding whether to use an out-of-district hourly rate in its fee calculation. A plaintiff bringing suit under the Voting Rights Act, pursuant to which fees can be recovered from the other side, has little incentive to negotiate a rate structure with his attorney prior to the litigation; the district court must act later to ensure that the attorney does not recoup fees that the market would not otherwise bear. Indeed, the district court (unfortunately) bears the burden of disciplining the market, stepping into the shoes of the reasonable, paying client, who wishes to pay the least amount necessary to litigate the case effectively.

Bearing these background principles in mind, the district court should, in determining what a reasonable, paying client would be willing to pay, consider factors including, but not limited to, the complexity and difficulty of the case, the available expertise and capacity of the client's other counsel (if any), the resources required to prosecute the case effectively (taking account of the resources being marshaled on the other side but not endorsing scorched earth tactics), the timing demands of the case, whether the attorney had an interest (independent of that of his client) in achieving the ends of the litigation or initiated the representation himself, whether the attorney was initially acting pro bono (such that a client might be aware that the attorney expected low or non-existent remuneration), and other returns (such as reputation, etc.) the attorney expected from the representation.

Although we clarify the application of the forum rule, we affirm the judgment of the district court in this case. It is clear that the district court would adhere to its fee award were we to vacate the district court's judgment and remand for reconsideration. Indeed, we believe that a reasonable, paying resident of Albany would have made a greater effort to retain an attorney practicing in the Northern District of New York, whether in Syracuse, Binghamton, Utica, or Kingston, than did plaintiffs. The rates charged by attorneys practicing in the Southern District of New York would simply have been too high for a thrifty, hypothetical client — at least in comparison to the rates charged by local attorneys, with which he would have been familiar.

BACKGROUND

On April 22, 2003, plaintiffs filed a complaint against Albany County and its Board of Elections ("Albany defendants") alleging that Albany County's 2002 legislative redistricting plan violated § 2 of the Voting Rights Act of 1965. See 42 U.S.C. § 1973. On August 22, 2003, the District Court for the Northern District of New York (Mordue, Judge) enjoined Albany County from conducting its scheduled November 2003 election pending adoption by the Albany County Legislature of a revised redistricting plan.

Further proceedings below culminated in the district court's rejection of plaintiffs' request that it order Albany County to hold a special election to take the place of the enjoined November 2003 election; plaintiffs then appealed to this court. On January 28, 2004, we vacated the district court's judgment and ordered the County to hold the special election on March 2, 2004. See Arbor Hill Concerned Citizens Neighborhood Ass'n v. County of Albany, 357 F.3d 260 (2d Cir.2004) ("Arbor Hill I").

Plaintiffs then moved in this court for an award of attorney's fees under 42 U.S.C. § 1973l(e). While we acknowledged the merit of the motion in principle, we remanded for a determination of the appropriate fee. See Arbor Hill Concerned Citizens Neighborhood Ass'n v. County of Albany, 369 F.3d 91 (2d Cir. 2004) ("Arbor Hill II"). We noted that plaintiffs had not demonstrated that "special circumstances existed" that would justify the use of higher rates than those prevailing in the Northern District of New York in calculating that fee. Arbor Hill II, 369 F.3d at 96 (quoting In re "Agent Orange" Prods. Liab. Litig., 818 F.2d 226, 232 (2d Cir.1987)).

During the course of this litigation, three entities have rendered legal services to the plaintiffs: (1) the Albany law firm of DerOhannesian & DerOhannesian ("D&D"), as local counsel; (2) the Washington, D.C.-based non-profit Lawyer's Committee for Civil Rights Under Law ("LCCRUL"), selected for its voting rights expertise; and (3) the Manhattan law firm of Gibson, Dunn & Crutcher ("Gibson Dunn"), chosen because of the firm's practice before the Second Circuit and the firm's "muscle," specifically, its ability to quickly prepare the appeal on an abbreviated briefing schedule.

Gibson Dunn sought in the district court to recoup attorney's fees calculated on the basis of the hourly rate charged by most attorneys in the Southern District of New York (and the hourly rate usually charged by Gibson Dunn). The district court denied Gibson Dunn's request that it adjust the hourly rate it would use to calculate the fees due from that prevalent in the Northern District of New York. The district court explained, "it is undisputed that plaintiffs did not even attempt to contact attorneys or law firms in the Northern District of New York outside of Albany County insofar as obtaining representation in this matter." Noting that "it was plaintiffs' obligation to submit factual support for their claim that there were no law firms in Syracuse, Binghamton, Utica or Kingston ready, willing or able to take their case," the district court held that plaintiffs had not adequately justified their request for higher fees.

In addition, the district court reduced the fee award proposed by Gibson Dunn in various other respects not relevant to this appeal. Plaintiffs then timely appealed the fee award, challenging only the district court's decision to award Gibson Dunn a fee based on the hourly rate commonly charged in the Northern District.

ANALYSIS

I. A Brief History of Attorney's Fees Awards

Courts in the United States have historically applied the "American Rule," under which each party is to bear its own costs of litigation, unmitigated by any fee-shifting exceptions. See Alyeska Pipeline Serv. Co. v. Wilderness Soc'y, 421 U.S. 240, 247, 95 S.Ct. 1612, 44 L.Ed.2d 141 (1975). In 1976, however, Congress enacted the Civil Rights Attorney's Fees Awards Act, which, like the provision of the VRA at issue in this appeal, provided that prevailing parties could recoup "reasonable attorney's fees." See 42 U.S.C. § 1988(b); cf. 42 U.S.C. § 1973l(e) ("In any action or proceeding to enforce the voting guarantees of the fourteenth or fifteenth amendment, the court, in its discretion, may allow the prevailing party . . . a reasonable attorney's fee. . . . ").

In the accompanying Senate Report, Congress implicitly endorsed two existing methods of calculating the "reasonable fee" that were developed in the 1970s by the circuit courts. See Hensley v. Eckerhart, 461 U.S. 424, 429-30 & n. 3, 103 S.Ct. 1933, 76 L.Ed.2d 40 (1983). The first, developed by the Third Circuit, was the "lodestar" method. See Lindy Bros., Builders, Inc. v. Am. Radiator & Standard Sanitary Corp., 487 F.2d 161 (3d Cir.1973). The lodestar was the product of the attorney's usual hourly rate and the number of hours worked. See id. at 167 (directing district courts to calculate the lodestar using the attorney's "normal billing rate"); see also City of Burlington v. Dague, 505 U.S. 557, 559, 112 S.Ct. 2638, 120 L.Ed.2d 449 (1992). After determining the lodestar, the district court could adjust it in setting the reasonable fee. See generally Hensley, 461 U.S. at 433, 103 S.Ct. 1933 ("The most useful starting point for determining the amount of a reasonable fee is the number of hours reasonably expended on the litigation multiplied by a reasonable hourly rate. This calculation provides an objective basis on which to make an initial estimate. . . .") (emphasis added); Lindy, 487 F.2d at 168-69. Thus, the lodestar method involved two steps: (1) the lodestar calculation; and (2) adjustment of the lodestar based on case-specific considerations.

The second method, developed by the Fifth Circuit, was for district courts to consider twelve specified factors to establish a reasonable fee. See Johnson v. Ga. Highway Express, Inc., 488 F.2d 714 (5th Cir.1974), abrogated on other grounds by Blanchard v. Bergeron, 489 U.S. 87, 92-93, 96, 109 S.Ct. 939, 103 L.Ed.2d 67 (1989) (declining to limit fee award to amount stipulated in attorney-client agreement). The Johnson method differed from the lodestar method in that it contemplated a one-step inquiry.

These two circuits had sought to channel the district court's discretion in different ways. The lodestar method was consistent with the law firm practice of accounting for each billable hour. See Lindy, 487 F.2d at 167 ("The first inquiry of the court should be into the hours spent by the attorneys. . . ."); see also Gisbrecht v. Barnhart, 535 U.S. 789, 800-01, 122 S.Ct. 1817, 152 L.Ed.2d 996 (2002) ("As it became standard accounting practice to record hours spent on a client's matter, attorneys increasingly realized that billing by hours devoted to a case was administratively convenient. . . ."). When the lodestar did not accurately reflect the market, the district court retained authority to adjust the lodestar to ensure that the fee ultimately awarded was reasonable. By contrast, under the Johnson method, the "hours claimed or spent on a case" were not "the sole basis for determining a fee." Johnson, 488 F.2d at 717. Rather than depending on market forces, the Johnson method relied on the district court's experience and judgment. See id. at 718 ("The trial judge's expertise gained from past experience as a lawyer and his observation from the bench of lawyers at work become highly important"); id. at 720 (discussing the necessary "balancing process"). Compare id. ("By this discussion we do not attempt to reduce the calculation of a reasonable fee to mathematical precision."), with Lindy, 487 F.2d at 167.

In theory, therefore, a district court that adopted the lodestar method was expected to consider fewer variables than a district court utilizing the Johnson method. In practice, however, both considered substantially the same set of variables — just at a different point in the fee-calculation process. A district court using the lodestar method would set the lodestar and then consider whether, in light of variables such as the difficulty of the case, it should adjust the lodestar before settling on the reasonable fee it was ultimately inclined to award. See, e.g., Silberman v. Bogle, 683 F.2d 62, 64 (3d Cir.1982); Baughman v. Wilson Freight Forwarding Co., 583 F.2d 1208, 1217-18 (3d Cir.1978) (permitting the district court to multiply the lodestar by a "contingency factor" and accepting, in theory, that obtaining an exceptional result might justify a further upward departure from the lodestar). By contrast, a district court employing the Johnson method would consider factors, such as the difficulty of the case, earlier in the fee-calculation process by weighing them in setting its tentative reasonable fee, from which there would seldom be a need to depart. See, e.g., In re First Colonial Corp. of Am., 544 F.2d 1291, 1299-1300 (5th Cir.1977) (outlining a process whereby first, the attorney seeking fees would document the hours devoted to the case; second, the district court would consider the Johnson factors and set a reasonable hourly rate; and third, the district court would explain how it balanced the Johnson factors to arrive at the reasonable hourly rate).

The Supreme Court adopted the lodestar method in principle, see Hensley, 461 U.S. at 433, 103 S.Ct. 1933; Blum v. Stenson, 465 U.S. 886, 104 S.Ct. 1541, 79 L.Ed.2d 891 (1984), without, however, fully abandoning the Johnson method. Rather than using the attorney's own billing rate to calculate the lodestar and then examining the lodestar in light of case-specific variables to ensure that it was in fact a reasonable fee, as the Third Circuit had suggested, the Supreme Court instructed district courts to use a reasonable hourly rate — which it directed that district courts set in light of the Johnson factors — in calculating what it continued to refer to as the lodestar. See Hensley, 461 U.S. at 434 n. 9, 103 S.Ct. 1933 ("The district court also may consider other factors identified in Johnson though it should note that many of these factors usually are subsumed within the initial calculation of hours reasonably expended at a reasonable hourly rate.") (citation omitted) (emphasis added); Blum, 465 U.S. at 898-900, 104 S.Ct. 1541. The Supreme Court collapsed what had once been a two-step inquiry into a single-step inquiry; it shifted district courts' focus from the reasonableness of the lodestar to the reasonableness of the hourly rate used in calculating the lodestar, which in turn became the de facto reasonable fee.

But the Supreme Court's emphasis on the Third Circuit's economic model, see, e.g., Missouri v. Jenkins, 491 U.S. 274, 283, 109 S.Ct. 2463, 105 L.Ed.2d 229 (1989) ("Our cases have repeatedly stressed that attorney's fees . . . are to be based on market rates for the services rendered."), and its simultaneous invocation of the equitable Johnson factors at an early stage of the fee-calculation process, proved to be in tension, see Blum, 465 U.S. at 895 n. 11, 104 S.Ct. 1541 ("We recognize, of course, that determining an appropriate 'market rate' for the services of a lawyer is inherently difficult . . . since market prices . . . are determined by supply and demand."). While the Third Circuit had expected district courts to correct for market dysfunction, the Supreme Court now asked district court judges to hypothesize that market on the basis of their experience as lawyers within their districts and on the basis of affidavits provided by the parties. Generally speaking, the rates an attorney routinely charges are those that the market will bear; yet the Supreme Court required that the district courts conjure a different, "reasonable" hourly rate.

After Hensley and Blum, circuit courts struggled with the nettlesome interplay between the lodestar method and the Johnson method. Compare Rutherford v. Harris County, Tex., 197 F.3d 173, 192 (5th Cir.1999) ("To decide an appropriate attorney's fee award, the district court was first required to calculate a lodestar fee depending on the circumstances of the case and the Johnson factors. The court was next obligated to consider whether the lodestar amount should be adjusted upward or downward, depending on the . . . Johnson factors.") (emphasis added), with Murray v. Weinberger, 741 F.2d 1423, 1430 (D.C.Cir.1984) ("The reasonable hourly rate which is incorporated into the lodestar figure generally reflects the reputation and ability of the attorney, the attorney's experience, and the level of skill required for the particular case."), and Bebchick v. Wash. Area Metro. Transit Comm'n, 805 F.2d 396, 404 (D.C.Cir.1986) ("Of course, `the actual rate that applicant's counsel can command on the market is itself highly relevant proof of the prevailing community rate.'").

And the Supreme Court has not yet fully resolved the relationship between the two methods. In cases decided after Hensley and Blum, it has both (1) suggested that district courts should use the Johnson factors to adjust the lodestar, see, e.g., Blanchard, 489 U.S. at 94, 109 S.Ct. 939 (stating that the district court should arrive at an initial estimate and then "adjust this lodestar calculation by other factors"); see also id. ("The Johnson factors may be relevant in adjusting the lodestar amount. . . ."); Pierce v. Underwood, 487 U.S. 552, 582-83, 108 S.Ct. 2541, 101 L.Ed.2d 490 (1988) (Brennan, J., concurring) (suggesting that factors might exist "that would justify an enhancement of the lodestar"), and (2) reiterated its holding in Hensley and Blum that "many of the Johnson factors 'are subsumed within the initial calculation.'" Penn. v. Del. Valley Citizens' Council for Clean Air, 478 U.S. 546, 564, 106 S.Ct. 3088, 92 L.Ed.2d 439 (1986).

Our court has done little to resolve this confusion. Compare Kassim v. City of Schenectady, 415 F.3d 246, 255-56 (2d Cir. 2005) (affirming the district court's authority to "reduce the fee awarded to a prevailing plaintiff below the lodestar by reason of the plaintiff's 'partial or limited success'") (emphasis added), with Luciano v. Olsten Corp., 109 F.3d 111, 116 (2d Cir.1997) ("The product of the number of reasonable hours times a reasonable hourly rate, however, does not end the inquiry. There remain other considerations, based on the facts of the particular case, that may lead the district court to ultimately make an adjustment to the hourly structure.") (internal citations omitted), and McDonald v. Pension Plan of the NYSA-ILA Pension Trust Fund, 450 F.3d 91, 97 (2d Cir.2006) (lodestar calculated on the basis of "prevailing rate specifically for ERISA practitioners in this Circuit") (emphasis added), and Chambless v. Masters, Mates & Pilots Pension Plan, 885 F.2d 1053, 1058 (2d Cir.1989) (suggesting, in determining the lodestar, that "smaller firms may be subject to their own prevailing market rate").

The net result of the fee-setting jurisprudence here and in the Supreme Court is that the district courts must engage in an equitable inquiry of varying methodology while making a pretense of mathematical precision. See Report of the Third Circuit Task Force, Court Awarded Attorney Fees, 108 F.R.D. 237, 247 (1985) ("The Lindy process creates a sense of mathematical precision that is unwarranted. . . ."). The "lodestar" is no longer a lodestar in the true sense of the word — "a star that leads," Webster's Third International Dictionary 1329 (1981). Nor do courts use it in the way the term was first used by the Third Circuit — as a base amount that is susceptible of ready adjustment; rather, circuit court deference to the district court's estimate of a "reasonable" hourly rate is a "lodestar" only in the sense that it is a guiding jurisprudential principle, see Dague, 505 U.S. at 562, 112 S.Ct. 2638 ("The 'lodestar' figure has, as its name suggests, become the guiding light of our fee-shifting jurisprudence."). What the district courts in this circuit produce is in effect not a lodestar as originally conceived, but rather a "presumptively reasonable fee." See id. (holding that the fee applicant bears the "burden of showing that `. . . an adjustment is necessary to the determination of a reasonable fee'"). The focus of the district courts is no longer on calculating a reasonable fee, but rather on setting a reasonable hourly rate, taking account of all case-specific variables.

The district court's opinion, including the report and recommendation of Magistrate Judge David R. Homer, with which the district court agreed after de novo review, reflects the general confusion surrounding the lodestar calculation. In places, the district court appears to envision a two-step lodestar-calculation process; yet elsewhere it seems to contemplate undertaking the calculation in one step. Likewise, at times, the district court emphasizes its role in approximating the workings of the market, but it also suggests some difference between "rates . . . paid by private retained clients . . . and rates ordered by courts."

The meaning of the term "lodestar" has shifted over time, and its value as a metaphor has deteriorated to the point of unhelpfulness. This opinion abandons its use. We think the better course — and the one most consistent with attorney's fees jurisprudence — is for the district court, in exercising its considerable discretion, to bear in mind all of the case-specific variables that we and other courts have identified as relevant to the reasonableness of attorney's fees in setting a reasonable hourly rate. The reasonable hourly rate is the rate a paying client would be willing to pay. In determining what rate a paying client would be willing to pay, the district court should consider, among others, the Johnson factors; it should also bear in mind that a reasonable, paying client wishes to spend the minimum necessary to litigate the case effectively. The district court should also consider that such an individual might be able to negotiate with his or her attorneys, using their desire to obtain the reputational benefits that might accrue from being associated with the case. The district court should then use that reasonable hourly rate to calculate what can properly be termed the "presumptively reasonable fee."

II. The Forum Rule

We turn now to the particular fee-calculation rule at issue in this case. It was against the muddled legal landscape we have just described that the Second Circuit promulgated what we will call the "forum rule." The Supreme Court directed that district courts should use the "prevailing hourly rate in the community" in calculating the lodestar — or what we are now calling the presumptively reasonable fee. After Blum, we explained that the "community" for purposes of this calculation is the district where the district court sits. See Polk v. N.Y. State Dep't of Corr. Servs., 722 F.2d 23, 25 (2d Cir.1983).

However, district courts — and indeed our court — quickly succumbed to the general confusion surrounding the difference between a "lodestar" and a reasonable hourly rate. Sometimes, they considered the variation between in-district and out-of-district rates in setting the hourly rate (which they then used to calculate the presumptively reasonable fee); but sometimes, they considered that variation only in deciding whether to adjust the presumptively reasonable fee after they had arrived at it (on the basis of in-district rates). Compare Polk, 722 F.2d at 25 ("The rate prevailing in the appropriate community is only one of many factors bearing on determination of a fee award."), with Arbor Hill II, 369 F.3d at 96-97 (intimating that a district court should permit plaintiffs to recover more than a fee calculated on the basis of the hourly rate usually charged by attorneys in the forum district only if plaintiffs could "show . . . that the case required special expertise beyond the competence of forum district law firms").

We now clarify that a district court may use an out-of-district hourly rate — or some rate in between the out-of-district rate sought and the rates charged by local attorneys — in calculating the presumptively reasonable fee if it is clear that a reasonable, paying client would have paid those higher rates. We presume, however, that a reasonable, paying client would in most cases hire counsel from within his district, or at least counsel whose rates are consistent with those charged locally. This presumption may be rebutted — albeit only in the unusual case — if the party wishing the district court to use a higher rate demonstrates that his or her retention of an out-of-district attorney was reasonable under the circumstances as they would be reckoned by a client paying the attorney's bill.

We believe that the district court's assessment of the reasonableness of a prevailing party's decision to retain out-of-district counsel is best considered in setting the hourly rate — rather than in deciding whether to adjust a presumptively reasonable fee — for three reasons. First, our holding comports with the holdings of several sister circuits and with the Supreme Court's focus on reasonable hourly rates rather than reasonable fees. See, e.g., Blum, 465 U.S. at 895, 104 S.Ct. 1541 (emphasizing the importance of using the "market rate" in calculating attorney's fees); Rum Creek Coal Sales, Inc. v. Caperton, 31 F.3d 169, 175 (4th Cir.1994) ("In circumstances where it is reasonable to retain attorneys from other communities . . . the rates in those communities may also be considered."); Maceira v. Pagan, 698 F.2d 38, 40 (1st Cir.1983) ("If a local attorney could perform the service, a well-informed private client, paying his own fees, would probably hire local counsel at the local, average rate."); Chrapliwy v. Uniroyal, Inc., 670 F.2d 760, 769 (7th Cir. 1982) (querying whether "the choice of counsel was improvident").

Second, in Pierce v. Underwood , a case interpreting the attorney's fees provision of the Equal Access to Justice Act ("EAJA"), the Supreme Court hinted that in the "broad spectrum of litigation," the difficulty of obtaining local counsel competent to prosecute a particular case is "little more than a routine reason why market rates are what they are," 487 U.S. 552, 573, 108 S.Ct. 2541, 101 L.Ed.2d 490 (1988) (emphasis added). The Supreme Court distinguished that "broad spectrum of litigation" from the attorney's fees provision of the EAJA, which stipulates that fees "shall be based upon prevailing market rates" but "shall not be awarded in excess of $125 per hour unless the court determines that . . . the limited availability of qualified attorneys for the proceedings involved justifies a higher fee." 28 U.S.C. § 2412(d)(2)(A)(ii); see Pierce, 487 U.S. at 571-72, 108 S.Ct. 2541; see generally Healey v. Leavitt, ___ F.3d ___, ___ (2d Cir.2007).

Third and finally, our holding honors the Supreme Court's emphasis on the need to use the approximate market rate for an attorney's services in calculating the presumptively reasonable fee. See Jenkins, 491 U.S. at 283, 109 S.Ct. 2463. The legal communities of today are increasingly interconnected. To define markets simply by geography is too simplistic. Sometimes, legal markets may be defined by practice area. See A.R. ex rel. R.V. v. New York City Dep't of Educ., 407 F.3d 65, 80 (2d Cir.2005) ("So long as the law provides for or permits fee awards based on geographic markets for services, a lawyer may be paid at different rates for otherwise indistinguishable services."). On the other hand, many cases (including many voting rights cases) are intrinsically local, and the relevant legal market may be coextensive with or smaller than the district itself. By asking what a reasonable, paying client would do, a district court best approximates the workings of today's market for legal services. See Mathur v. Bd. of Trs. of S. Ill. Univ., 317 F.3d 738, 744 (7th Cir.2003) ("The realities of the legal community today mean that though some attorney probably could have represented the plaintiff, one factor or another prevented them from taking the case when he needed a lawyer."). Not incidentally, a reasonable, paying client might consider whether a lawyer is willing to offer his services in whole or in part pro bono, or to promote the lawyer's own reputational or societal goals. Indeed, by focusing on the hourly rate at which a client who wished to pay no more than necessary would be willing to compensate his attorney, the district court can enforce market discipline, approximating the negotiation that might ensue were the client actually required to pay the attorney's fees.

In occasionally permitting a deviation from forum rates in setting the rate that will yield the presumptively reasonable fee, we have in mind no substantial change in circuit law; where circumstances have warranted it, we have not insisted on strict adherence to the forum rule. In Polk, we approved the use of an out-of-district hourly rate. 722 F.2d at 25 (considering whether "counsel might . . . have expected plaintiff's claim to be adjudicated in the Southern District"). In Agent Orange, although we emphasized that district courts should generally use "the hourly rates employed in the district in which the reviewing court sits" in calculating the presumptively reasonable fee, 818 F.2d at 232, we again upheld a district court's decision to use different rates. And since Polk and Agent Orange, we have urged district courts where appropriate to employ out-of-district rates in calculating the fee due. See, e.g., New York City Dep't of Educ., 407 F.3d at 81 & n. 17 ("There is good reason for a district court not be wed to the rates in its own community. If they are lower than those in another district, skilled lawyers from such other district will be dissuaded from taking meritorious cases in the district with lower rates.").

In both Polk and Agent Orange, the touchstone of our analysis was the belief that district courts should award fees just high enough "to attract competent counsel," Lewis v. Coughlin, 801 F.2d 570, 576 (2d Cir.1986). See, e.g., Agent Orange, 818 F.2d at 233 ("Undercompensation could deny counsel their right to fair and just fees; overcompensation would not be consistent with the need to prevent windfalls."); cf. Crescent Publ'g Group, Inc. v. Playboy Enters., Inc., 246 F.3d 142, 151 (2d Cir.2001) (explaining that an attorney-client agreement may provide compelling evidence of the "prevailing market rate"). We adhere to this touchstone, but we would not be true to it by insisting on an overly strict application of the forum rule. Rather, to reiterate, a district court should consider the rate a reasonable, paying client would pay, and use that rate to calculate the presumptively reasonable fee.

III. The District Court's Decision

For the foregoing reasons, we agree with plaintiffs that the district court may have applied the forum rule too strictly. They suggest that the district court calculated the presumptively reasonable fee (on the basis of in-district rates) and then queried whether the plaintiffs had shown sufficient cause to rebut the presumption that it was, in fact, the ultimate reasonable fee.

However, we find no error in the district court's fee award, even when evaluated under the analysis we use. We are confident that a reasonable, paying client would have known that law firms undertaking representation such as that of plaintiffs often obtain considerable non-monetary returns — in experience, reputation, or achievement of the attorneys' own interests and agendas — and would have insisted on paying his attorneys at a rate no higher than that charged by Albany attorneys (and there is no cross-appeal).

Moreover, the considerable deference that we owe to a district court's assessment of the Johnson and other factors, see Farbotko, 433 F.3d at 210 ("The district court is in closer proximity to and has greater experience with the relevant community whose prevailing market rate it is determining."), counsels against remanding this case to the district court for further, likely unnecessary, proceedings.

CONCLUSION

For the reasons set forth above, we AFFIRM the judgment of the district court.