In 1989, plaintiffs hired the law firm of Futterman & Howard ("F & H") to file a class action challenging a reorganization plan adopted by defendant Rockford Board of Education. Two months later, the parties entered into a consent decree whereby the Board rescinded most of the plan's segregative elements. The case was not over, however, for the consent decree preserved plaintiffs' right to pursue a liability adjudication, which they did. Following the entry of the consent decree, the district court granted F & H an interim fee award relating to attorney's fees incurred in the implementation and enforcement of the consent decree. Both parties challenged the calculation of that award and Judge Roszkowski certified the question for appeal pursuant to 28 U.S.C. § 1292(b). We dismissed the appeal, however, on the ground that it was not a "final order," People Who Care v. Rockford Bd. of Education, 961 F.2d 1335, 1339 (7th Cir.1992), and F & H thus continued to accept interim fee awards while reserving their objections to those amounts until the entry of a final judgment.

The district court entered a Liability Judgment Order on August 31, 1995, requiring the Board "to eliminate root and branch, through the Rockford public school system, all vestiges of racial, ethnic and national origin segregation discrimination against African American and Hispanic students." In a Final Judgment Order entered on August 31, 1995, the court disposed of all issues regarding fees and expenses for the liability litigation, as well as reconsideration of the interim fee awards, by awarding F & H $1,827,124 in fees and $198,342 in expenses (plus interest on both). F & H appeals, and Rockford cross-appeals, the fees portion of that order on the ground that the district court's calculation was incorrect. For the following reasons, we reverse the district court's fee determination.

Under 42 U.S.C. § 1988, prevailing parties are entitled to "reasonable" fees as a portion of their cost of bringing suit. Clear guidelines have been developed to aid courts in calculating the amount of those fees. The "lodestar" method--reasonable hourly rates multiplied by hours reasonably expended--is the most appropriate starting point. Hensley v. Eckerhart, 461 U.S. 424, 433, 103 S. Ct. 1933, 1939, 76 L. Ed. 2d 40. In determining "reasonable hourly rates," the Supreme Court has repeatedly stressed that "attorney's fees awarded under [Section 1988] are to be based on market rates for services rendered." Missouri v. Jenkins, 491 U.S. 274, 283, 109 S. Ct. 2463, 2469, 105 L. Ed. 2d 229; see also Pressley v. Haeger, 977 F.2d 295, 299 (7th Cir.1992). The attorney's actual billing rate for comparable work is "presumptively appropriate" to use as the market rate. Gusman v. Unisys Corp., 986 F.2d 1146, 1150 (7th Cir.1993). If the court is unable to determine the attorney's true billing rate, however (because he maintains a contingent fee or public interest practice, for example), then the court should look to the next best evidence-the rate charged by lawyers in the community of "reasonably comparable skill, experience, and reputation." Blum v. Stenson, 465 U.S. 886, 892, 895 n. 11, 104 S. Ct. 1541, 1545, 1547 n. 11, 79 L. Ed. 2d 891. Once the court reaches an amount using the lodestar determination, it may then adjust that award in light of factors adopted by Congress in enacting Section 1988, known as the Hensley factors,1 although most of those factors are usually subsumed within the initial lodestar calculation. Hensley, 461 U.S. at 434 n. 9, 103 S. Ct. at 1940 n. 9.

With these principles in mind, we must determine whether the district court's calculation in this case was an abuse of its discretion. Bankston v. State of Illinois, 60 F.3d 1249, 1255 (7th Cir.1995). We conclude that it was. The court properly began by using the lodestar method and a historical rate-plus-interest approach to compensate F & H for the delay in receiving its payments. See Smith v. Village of Maywood, 17 F.3d 219, 221 (7th Cir.1994). Where it erred, however, was in its calculation of "reasonable hourly rates" and "reasonable hours."

Because the court's treatment of hourly rates was essentially the same for each of F & H's personnel, this opinion need only discuss in detail its calculation of Mr. Howard's rate.2 Initially, Mr. Howard submitted an affidavit indicating that his requested rates over the five-year period were the rates that he actually billed. For 1989-90, he presented actual statements reflecting 33.4 hours billed at $210. He further submitted evidence that he was awarded $210 per hour in similar cases during that period (see, e.g., Littlefield v. Mack, 750 F. Supp. 1395 (N.D. Ill.1990); ACLU v. City of Chicago, No. 75 C 3295 (N.D. Ill. filed Feb. 22, 1993), aff'd. sub nom. Alliance to End Repression, No. 74 C 3268, 1994 WL 86690 (N.D.Ill. March 11, 1994)). Finally, he referred the court to affidavits that he had submitted in 1990 (in reference to the Interim Fee Order) indicating that the $210 rate was in line with similarly situated attorneys. For 1991-92, Mr. Howard submitted billing statements indicating that he billed one hour at $240 and that his partner, Mr. Futterman, billed 21.9 hours at that rate. He further submitted evidence that he was awarded a $250 per hour rate in ACLU v. City of Chicago, supra. For 1992-93, he presented no statements relating to hours that he had billed paying clients because he "was required to devote substantially all of his time to [this case]." He did, however, present evidence that elsewhere his partner, Mr. Futterman, had billed 36.75 hours at $260 per hour and had been awarded rates of $275 and $280 in numerous other cases. See, e.g., Jaffee v. Redmond, 855 F. Supp. 244 (N.D.Ill.1994).

The district court acknowledged that an attorney's actual billing rate is "presumptively appropriate." Gusman, 986 F.2d at 1150. However, for each of the above years, the district court concluded that Mr. Howard's evidence failed to establish that his requested rate was his actual billing rate. It concluded that the number of hours on his billing statements was "insufficient"; that it was "not bound by [other fee award] decisions"; and that the evidence of Mr. Futterman's billing rates was "entirely irrelevant." Furthermore, looking to the next best evidence, the court considered the statements indicating Mr. Futterman's rate "not enough" to establish that it was the rate similar attorneys charged for similar work; and it concluded that the affidavits of other Chicago civil rights lawyers were relevant only to the years in which they were submitted (1989-90 and 1993-94) because the affiants spoke only in current terms. Thus the court held that "under Gusman, effectively, there is no presumption."

The court was correct that the Gusman approach requires a plaintiff's attorney to do more than merely request an hourly rate; he must present evidence to establish that the requested rate is his actual billing rate. The court erred, however, in setting the amount of evidence required at a nearly unattainable level. The evidence F & H submitted was more than sufficient to establish its rates, and the district court's contrary conclusion was an abuse of its discretion. On remand, the court should take into account the following considerations:First, there is nothing inappropriate about establishing an attorney's overall rate using only a small sample of the hours that he has billed to other clients. Compiling years of invoices and billing statements is not inexpensive and a rule forcing attorneys to do so would add little to the integrity of the fee-collecting process.

Second, evidence of fee awards received in similar cases should not be dismissed as "non-binding." While each court should certainly arrive at its own determination as to a proper fee, rates awarded in similar cases are clearly evidence of an attorney's market rate. Tolentino v. Friedman, 46 F.3d 645, 652 (7th Cir.1995).

Third, the billing rates of other attorneys in the same firm should not be dismissed as simply "irrelevant." If a court chooses to look to the next best evidence of what similarly situated attorneys charge, there is no reason that it should refuse consideration of those who practice in conjunction with the attorney at issue. The billing correlation between attorneys in the same firm is likely to be much stronger than that of attorneys who are merely in the same practice region. And while there may be occasions to distrust the rates of a related attorney, disproving all potential reasons for doing so is not part of the attorney's initial burden.

Fourth, an attorney's billing rate for a certain year need not be viewed in a vacuum. Evidence of his rate in Year 3 can still act as relevant evidence as to his rate in Year 2 and Year 4. For example, based upon affidavits signed in 1990, the court accepted as reasonable Mr. Howard's $210 rate for 1989-90.3 Based upon affidavits signed in 1994, it accepted as reasonable his $275 rate for 1993-94. However, it refused to view these affidavits as relevant evidence for establishing a rate for the years in between. Such a refusal is illogical: If a $210 rate was reasonable in 1989-90 and a $275 rate was reasonable in 1993-94, then the gradual increases requested for the years in between are most likely reasonable as well.

Fifth, the fact that an attorney spends nearly every hour of a certain year on the case at hand should not act to his detriment. Here the district court concluded that because many of the F & H attorneys billed very few hours to other clients in certain years, they had simply failed to establish their overall rate. But the goal is not to determine what an attorney actually made in a certain year. The goal is to determine what the attorney could have made "if he were not representing this plaintiff in this case." Gusman, 986 F.2d 1146. The fact that an attorney bills other clients very few hours at his requested rate does not necessarily imply that the requested rate is not the rate that he would have earned had he not been working on this case.

Sixth, atypical billing arrangements with certain clients should enter a court's calculation in a manner that is consistent with the attorney's overall practice. We have recognized that attorneys often provide services to some of their clients on terms that decrease their "true" billing rate:

Consider two cases: (1) An attorney who ordinarily works 2,000 hours in a year sells 1,900 of these hours to clients who pay $250 per hour and devotes the other 100 hours to civil rights litigation in which the court will fix the fee; (2) An attorney who ordinarily works 2,000 hours in a year sells 100 of these hours to clients who pay $250 per hour and devotes the other 1,900 to civil rights litigation in which the court will fix the fee. The willingness of clients to buy almost all of the attorney's time in Case 1 at $250 per hour implies that he could have sold the rest at that rate or close to it. Any discount needed to fill the remainder of his calendar would not have been steep. Thus $250 per hour is a good estimate of the opportunity cost of the civil rights case. In Case 2 the inference is not nearly so strong. Selling the first 5% of one's time is much easier than filling the remaining 95% of the timesheet. Too, most work in fee-shifting cases is compensated only if the client prevails. Even a lawyer who wins 80% of his cases is working, in effect, for 80% of the stated hourly wage.... Unless this lawyer is donating the other 20% of the value of his time to a favored cause, his behavior tells us that his market-clearing price is less than $250 per hour, perhaps considerably less. Thus we do not require the district court to leap directly from the willingness of some persons to pay $X to Lawyer Y that $X is "the" hourly rate of Lawyer Y. But this is, as we have said in many other cases, and reiterate today, the starting point.

Gusman, 986 F.2d at 1150-1151. However, reduced-rate hours should be considered only in proportion to the percentage of the attorney's practice they represent.

Here F & H had provided services to one of its largest clients, the Chicago Board of Education ("CBOE"), at rates that were lower than those charged to its other clients.4 However, the evidence indicated that the services provided the CBOE represented only 5 percent of F & H's total billable hours for this period. Despite this low percentage, the district court concluded that the CBOE rate should be used as the overall rate for those years in which the CBOE hours represented a large proportion of an attorney's "non-People Who Care" hours. This was an incorrect approach. The work completed for the CBOE may have represented 50 percent, 60 percent, or some greater amount of an attorney's "non-People Who Care" hours, but that is not the proper inquiry. The focus should be on the amount the attorneys would have made if they had not been representing the plaintiffs in this case. And there is no indication that F & H would have billed any more hours to the CBOE than it did had it spent the rest of its time working for clients other than these plaintiffs. Thus the CBOE rate should have been viewed, at most, as 5 percent of F & H's billing rate.5

Finally, and most important, the second half of the Gusman burden-shifting approach must be followed. Once an attorney provides evidence of his billing rate, the burden is upon the defendant to present evidence establishing "a good reason why a lower rate is essential." Gusman, 986 F.2d at 1151. A defendant's failure to do so is essentially a concession that the attorney's billing rate is reasonable and should be awarded. Here, other than an affidavit regarding paralegal rates, the Board presented absolutely no rebuttal evidence. And while the district court recognized this fact, "[n]otably, ... Defendant has not submitted any affidavits or other evidence to contradict the $275 per hour rate ... and Mr. Howard's evidence ... remains unrebutted," it only applied it to the last of the five years in question. This was improper. See Pressley v. Haeger, 977 F.2d 295, 299 (7th Cir.1992) (holding that court was required to award an attorney's billing rate where defendant had submitted no evidence on fees).

The district court also abused its discretion in reducing F & H's hours. First, the court reduced Mr. Howard's request by 161.3 hours. The reason for the reduction was that Mr. Howard would have had to bill 42.5 hours per week to reach the figure he requested. The court concluded that this amount was "patently incredible," and because it found 32.5 hours per week to be a "more reasonable" number, it awarded Mr. Howard that amount. This reduction was an abuse of discretion. There is nothing "incredible" about billing 42.5 hours per week. This equates to roughly 2,100 hours per year, which is below some of the minimum requirements of Chicago law firms and 600 hours below the annual hours awarded to other attorneys in this case. Moreover, the evidence indicated that Mr. Howard's opposing counsel billed more hours during this period than Mr. Howard did.

An attorney's hours are subject to the scrutiny of the court and unreasonable hours should not be compensated. However, "[t]he record ought to assure us that the district court did not 'eyeball' the fee request and 'cut it down by an arbitrary percentage because it seemed excessive to the court.' " Spellan v. Bd. of Educ. for Dist. 111, 59 F.3d 642, 647 (7th Cir.1995) (quoting Tomazzoli v. Sheedy, 804 F.2d 93, 97 (7th Cir.1986)). Here the court did not conclude that Mr. Howard had failed to achieve a successful result, see Anderson v. Secretary of Health & Human Serv., 80 F.3d 1500, 1505 (10th Cir.1996); that he had been inefficient or duplicitous; or that he had inadequately documented his time, see Hensley, 461 U.S. at 437, 103 S.Ct. at 1941. Rather, it essentially found that he had worked too many of his hours in too short a period. That is not a proper basis for finding requested hours unreasonable.

Additionally, the court reduced by half all of the hours F & H had spent preparing for the plaintiffs' appeal of the 1990 Interim Fee Order. The court concluded that because we dismissed the appeal as premature, the plaintiffs had not been "successful" and attorney's fees were therefore unwarranted. See Bankston, supra. We disagree with this conclusion. A court's focus should not be limited to the success/failure of each of the attorney's actions. Rather, it should be upon whether those actions were reasonable. In other words, the court should not look to whether F & H "won" the 1990 appeal, but should instead look to whether the fees F & H requests for that appeal were reasonably incurred. We conclude that they were. First, because the Board also appealed the 1990 Order, F & H would have been forced to incur substantially the same amount of fees in briefing and arguing the issue even if the plaintiffs had not chosen to appeal. Second, the decision to appeal was not unreasonable at the time it was made. Despite the fact that we dismissed it as premature, both parties, as well as Judge Roszkowski, felt that an immediate appeal of the issue was in everyone's best interest. Thus F & H's fees in relation to the 1990 appeal were not unreasonable.

Finally, we address two remaining concerns. First, we do not mean to suggest that there are never occasions where a court should depart from the lodestar determination. For example, in its original 1990 Interim Fee Award Order, the district court determined that the billing statements and affidavits established that the $210 per hour rate Mr. Howard requested for 1989-90 was the presumptive market rate. However, it did not award that rate for the following reason:

The court finds that [F & H] lacks the requisite experience in school desegregation cases to support the Chicago rates charged in their fee petition. For example, Attorney Robert Howard stated in his deposition that he has never, with the exception of a Chicago Board of Education case, handled a school desegregation case prior to the instant case. Furthermore, Mr. Howard knows of no other partner in his firm that has handled any other school desegregation cases.... The court believes that an attorney asking $210 to $225 per hour in attorney's fees should be very familiar with school desegregation law. People Who Care v. Rockford Bd. of Educ., 1990 WL 25883 at * 5 (N.D.Ill.).

Although the court awarded Mr. Howard his requested rate one year later, it set his rate for 1989-90 at $150 per hour. The experience (or inexperience) of an attorney is a permissible reason to depart from the presumptive rate. See Hensley, 461 U.S. at 430 n. 3, 103 S. Ct. at 1938 n. 3. However, courts must be careful in defining "experience": It would be possible to define it so narrowly that no attorney has experience in the type of case in front of him, which would be clearly an incorrect approach. Although this is a close case--Mr. Howard presented evidence of his work in other civil rights cases, but also admitted that he had never handled a desegregation case--we cannot say that the court's refusal to award Mr. Howard the market rate for experienced desegregation lawyers in 1989-90 was an abuse of discretion.

Next is the issue of paralegal rates. The district court refused to award F & H its requested $75 paralegal rate for 1991 through 1994, although it did conclude that the rate was within the range charged by other Chicago law firms. The Board's rebuttal evidence on this issue took the form of an affidavit from Cathy Heath, a former Vice President of the Illinois Paralegal Association. Ms. Heath testified that while $75 was within the range of Chicago paralegal rates for those years, the median rate for Chicago paralegals was $53 per hour. She further testified that the work done by F & H paralegals on this case was not sufficiently complex to command a higher, $75 per hour rate. In her opinion, complex paralegal work consists of duties such as "factual investigation, conducting legal research, summarizing depositions, checking citations, and compiling statistical and financial data." The district court agreed with Ms. Heath's conclusion, noting that the paralegal work on this case consisted mostly of "document preparation, coding and organizing." Thus he awarded F & H $60 for each paralegal hour.

When evaluating the hourly rates of paralegals, it is improper to focus on whether the work at issue was at the upper or lower range of complexity for paralegal-type work. Attorneys do staff their law firms with various "levels" of employees, and it is not unreasonable to require those attorneys to delegate work to the correct level, according to its complexity. However, requiring attorneys to staff their offices with paralegals representing the spectrum of the hourly rate range, and to assign each "paralegal task" to the correct paralegal takes the concept too far. The only inquiry for requested paralegal fees should be whether the work was sufficiently complex to justify the efforts of a paralegal, as opposed to an employee at the next rung lower on the pay-scale ladder. With this focus in mind, it is unclear whether "document preparation, coding and organizing" is reasonably considered paralegal work. On remand, if the court determines these tasks to be sufficiently complex to justify the services of a paralegal, as opposed to clerical staff, then it should have no problem in awarding paralegal rates.

In conclusion, F & H provided ample evidence to establish the billing rates of each of its personnel: It introduced affidavits of its rates; it corroborated these by submitting actual billing statements; it provided evidence of other cases in which it had been awarded those same rates; and finally, it provided next best evidence indicating that its rates were in line with those of similarly situated attorneys. Given that the district court reached an opposite conclusion regarding the establishment of these rates, a remand is necessary to allow it to recalculate the reasonable hourly rates for each F & H employee in light of this opinion. It should also reconsider the work completed by F & H paralegals to determine whether the requested paralegal rate was appropriate. Furthermore, if the court believes that additional evidence is necessary for it to achieve rate determinations that are supported by the record, then further evidentiary submissions should be required of the parties.

For the foregoing reasons, the judgment of the district court is REVERSED and the case is REMANDED for a recalculation of F & H's fees in a manner consistent with this opinion.

COFFEY, Circuit Judge, dissenting.

"People Who Care," the prevailing plaintiffs in a school desegregation case, appeal the district court's reduction of their fee petition. Despite the extremely deferential standard of review for such petitions and the magistrate judge's comprehensive and thoughtful 82 page report and recommendation, approved by the experienced trial court judge, the majority reverses the fee petition and remands for recalculation. I respectfully dissent.

In 1989, "People Who Care" (an association of concerned residents of Rockford, Illinois) sued the Rockford School District under 42 U.S.C. § 1983, alleging that the district had engaged in a continual pattern of intentional segregation and discrimination. After two interim agreements between the parties,1 the district court transferred the case to a magistrate judge pursuant to 28 U.S.C. § 636(b)(1)(B) to make proposed findings of fact and recommendations for the district court's disposition of the matter. In April 1993, The magistrate judge held a hearing, spanning some twenty-four days and heard testimony from over 40 witnesses, reviewed 150 depositions, and thousands of pages of documentary evidence. Some six months thereafter, in November 1993, the magistrate judge concluded in his report and recommendation that the Rockford School District had engaged in systematic segregation and discrimination against minority school children. In February 1994, the district court, after having had the opportunity to review the record, adopted the magistrate's report and granted the plaintiffs' motion for a permanent injunction against the school district.

Since the plaintiffs were "prevailing parties" in their civil rights suit against the school district, section 1988(b) of Chapter 42 of the U.S. Code provides that: "the court in its discretion, may allow the prevailing party, other than the United States, a reasonable attorney's fee as part of the costs." Accordingly, the plaintiffs submitted a fee petition to the magistrate judge, who held pretrial conferences, discovery conferences and allowed an extended briefing schedule concerning the award of attorney's fees. The plaintiffs requested a total of $3,755,999.00 in attorney's fees and costs. After a thorough review, the magistrate judge rejected the majority of the school board's objections to the fee petition and recommended to the district judge, in a detailed 82-page report, reducing the award to $2,025,486.00. The magistrate judge found that (1) the amount of preparation hours claimed for trial was excessive; (2) the evidence failed to support the plaintiffs' lawyer's contention that his requested hourly rate was in fact his normal market rate; and (3) the evidence concerning the plaintiffs' lead lawyer's experience in 1989 as an expert in the field of school desegregation law failed to qualify him as an expert and thus his requested rate for 1989 was unreasonable. The district court accepted the magistrate's report and the plaintiffs appeal.

In reviewing the district court's determination of "reasonable" attorney's fee awards, our standard is the highly deferential "abuse of discretion" standard. Spellan v. Board of Educ., 59 F.3d 642, 645 (7th Cir.1995); Dunning v. Simmons Airlines Inc., 62 F.3d 863, 872 (7th Cir.1995). "We will not find an abuse of discretion if reasonable persons could differ over the district court's view." Leffler v. Meer, 936 F.2d 981, 984 (7th Cir.1991) (citing Lightfoot v. Walker, 826 F.2d 516, 520 (7th Cir.1987)).

The Supreme Court has emphasized that "the district court has discretion in determining the amount of the fee award." Hensley v. Eckerhart, 461 U.S. 424, 437, 103 S. Ct. 1933, 1941, 76 L. Ed. 2d 40 (1983). The Court went on to state that discretionary review is "appropriate in view of the district court's superior understanding of the litigation and the desirability of avoiding frequent appellate review of what essentially are factual matters." Id. The district court is afforded significant deference in fee matters because the desirability of avoiding a second major litigation over attorneys' fees outweighs the need for uniformity in attorneys' fee awards. Spellan, 59 F.3d at 645 (citing Hensley, 461 U.S. at 437, 103 S.Ct. at 1941).

The authoring judge of today's decision has observed in a prior case that "there is no one correct formula for determining a fee award" and that the district court's calculation requires deference because it is "anything but an arithmetical exercise." Spanish Action Comm. of Chicago v. City of Chicago, 811 F.2d 1129, 1134 (7th Cir.1987) (Cummings, J.) (quoting Tomazzoli v. Sheedy, 804 F.2d 93, 97 (7th Cir.1986)). Further, as the other judge in the majority has previously stated: "the district judge can, of course, and should, disallow particular expenses that are unreasonable whether because excessive in amount or because they should not have been incurred at all." Henry v. Webermeier, 738 F.2d 188, 192 (7th Cir.1984). Following the law articulated in Henry, the magistrate judge, as described above, reduced the rates and hours of attorney's fees requested by the plaintiffs' lawyers, finding them to be unreasonable and excessive.

"Despite the district court's wide latitude in determining a fee award, if the requested hourly rate or number of hours is reduced, a clear explanation must be provided." Leffler, 936 F.2d at 985 (citing Tomazzoli, 804 F.2d at 97). For instance, we remanded a fee reduction in Henry because the district judge had failed to provide any explanation as to his reasoning regarding the reasonableness or unreasonableness of requested fees or hours. 738 F.2d at 192. The judge in Henry erred by relying on a statute to disallow certain expenses, instead of making his own, reasoned decision as to the appropriate attorney's fees and costs to award to the prevailing party. Id. at 195. In contrast, in the case before us, the magistrate judge provided a detailed, thoughtful and well-reasoned explanation for each and every reduction in the hourly rate and requested hours for the plaintiffs' lawyers. See Spanish Action Comm., 811 F.2d at 1134 (affirming reduction in attorney's fees because "[t]he district court's opinion here provides a lengthy and detailed explanation for its decision to reduce substantially the number of hours for which plaintiff requested compensation....") (Cummings, J.). I fail to understand how, in light of the clear statements of law in Spanish Action Committee and Henry and the explanations of the magistrate judge explicitly following the law, the majority has seen fit to remand this case to the trial court for re-evaluation of the fee request.

In People Who Care, the magistrate judge, along with the experienced district court judge (a former very talented and experienced trial counsel), oversaw the six years of protracted litigation in this case, presided at the numerous hearings, personally interacted with the lawyers for both sides, and reviewed their voluminous submissions to the court. The experienced trial judge approved the magistrate judge's clear and reasonable explanation of each one of his findings.

For instance, in 1990, the trial judge, while presiding over this case, reduced the hours for the lead plaintiffs' attorney by 25 percent, finding that the requested number of hours for that time frame "was patently incredible to this court."2 Given each of the judges' intimate familiarity and daily contact with the proceedings and the lawyers, the district court and magistrate judge's explanations that the lead lawyer billed an excessive amount of hours for a particular proceeding was eminently reasonable. I fail to understand how we as appellate judges can second guess a factual determination of this nature based solely on the cold pages of an appellate record. See United States v. Hatchett, 31 F.3d 1411, 1416 (7th Cir.1994) (observing that great deference is given to fact-finders for they have the best opportunity to observe the witnesses).

Further, the trial court reduced the lead lawyer's rate in two of the six years, finding that the plaintiffs' lawyer had submitted a self-serving statement with insufficient documentary evidence to establish that the inflated rate requested would have been paid by another client had he not been involved in the civil rights litigation. See Gusman v. Unisys Corp., 986 F.2d 1146, 1150 (7th Cir.1993) (observing that "the best measure of the cost of an attorney's time is what that attorney could earn from paying clients."). The magistrate judge noted that in the two years in question, the plaintiff's lawyer had submitted evidence of his rates for only a handful of his hours (because, according to the attorney in his self-serving statement, the bulk of his time had been taken up with the school desegregation case). Thus, following Gusman, the magistrate judge found that the plaintiffs' lawyer had not established a presumptive market rate for his services, because if a paying client pays a high rate for a portion of the attorney's time, it does not necessarily follow that all of the attorney's time will be bartered at the same inflated rate. See Id. at 1150-51. Similarly, the magistrate judge observed that the secondary evidence submitted by the plaintiff's lawyer in support of his claimed lost "opportunity costs" (i.e. the billing rates of other lawyers in his firm and the billing rates of other civil rights lawyers in Chicago) was insufficient to support the requested fee rate in two out of the six years. The magistrate judge, who was familiar with the lawyers in the case, found that the rates of Chicago civil rights lawyers for 1993-94 (submitted as evidence by the plaintiff's lawyer) fell short of establishing that the plaintiffs' lawyer's requested rate for 1989-90 was reasonable.

While we as individual judges of this court may come to a different conclusion if we were sitting in the trial court's place, we are obliged to review, according to case precedent, the district court's determinations under the well-settled abuse of discretion standard. Although the magistrate judge made voluminous findings weighing the submissions of the parties and his own knowledge of the litigation, the majority of this court focuses on but one or two isolated findings to declare that the entire trial court's calculations were "unreasonable."

Our job on appeal is not to second-guess the district court's discretionary rulings, but only to review the record in its totality under the abuse of discretion standard. The reductions in the hours and rates that the judge deemed appropriate were supported with clear, well-reasoned, and detailed explanations. Because the majority for reasons unexplained failed to apply the appropriate standard of review, abuse of discretion, I am forced to dissent.