Lloyd N. Cutler, Washington, D. C., also entered an appearance for appellees.

Lutz A. Prager, Washington, D. C., for amicus curiae, Equal Employment Opportunity Commission.

Charles R. Halpern, Washington, D. C., was on brief for amici curiae, American Civil Liberties Union, et al.

David M. Dorsen, Richard Seymour, Richard S. Kohn, Roderic V. O. Boggs, and Ann K. Macrory, Washington, D. C., were on brief for amicus curiae, Lawyers' Committee for Civil Rights Under Law.

James M. Nabrit III, Charles Stephen Ralston, Bill Lann Lee, and Eric Schnapper, New York City, were on brief for amicus curiae, NAACP Legal Defense and Educational Fund, Inc.

Bruce J. Terris, Washington, D. C., was on brief for amicus curiae, Law Office of Bruce J. Terris.

James R. Richards, Washington, D. C., was on brief for amicus curiae, Capital Legal Foundation.

Before WRIGHT, Chief Judge, and McGOWAN, TAMM, LEVENTHAL,* ROBINSON, MacKINNON, ROBB, WILKEY and MIKVA, Circuit Judges.

Opinion filed by Circuit Judge McGOWAN, in which Chief Judge WRIGHT and Circuit Judges SPOTTSWOOD W. ROBINSON, III, MacKINNON, ROBB and MIKVA join.

Opinion concurring filed by Circuit Judge MacKINNON, in which Circuit Judge ROBB joins.

Opinion dissenting filed by Circuit Judge WILKEY, in which Circuit Judge TAMM joins.

The court en banc has before it for review an order of the District Court awarding an attorney's fee of $160,000 for the successful prosecution of a gender-discrimination class suit against the United States Department of Labor.1 A panel of this court earlier reversed the District Court's award and remanded for reconsideration under the novel standards described in its opinion (Copeland I ).2 The panel denied rehearing, but issued a second opinion (Copeland II )3 clarifying the first. We granted rehearing en banc.4

At issue in this appeal are (1) the standards to be applied in awarding attorney's fees in Title VII suits against the government, and (2) the reasonableness of the District Court's fee award in this case. For the reasons set forth below, we affirm the District Court's award.

* We cannot determine whether the District Court's fee award was reasonable without examining in some detail the history of this employment discrimination litigation. This chronicle is necessarily lengthy because the lawsuit involved numerous and complex proceedings and maneuverings. We think the very intricacy of the litigation-which was a product, in part, of the government's vigorous and long-continued resistance to the claim asserted against it-is highly relevant to the reasonableness of the fee award.

Appellee Dolores Copeland, a black woman trained in data processing, joined the Department of Labor (the Department) in 1967. She worked for several years in the Department's Directorate of Data Automation and its predecessor unit (the Directorate) as a GS-13 computer specialist. Copeland thought that her supervisors were unfairly denying her training, promotions, and interesting work. Moreover, she believed that other female Directorate employees were treated similarly.

Pursuant to regulations, Copeland explained her suspicions to a Department Equal Employment Opportunity (EEO) counselor in April, 1973, but no action was taken. She therefore formally complained of discrimination in June. Between July and September, 1973, the Department investigated her complaint. Copeland filed comments and criticisms with respect to the investigation and report.

The Department reopened its investigation in November, and submitted supplemental reports in January and February, 1974. Copeland thought that this supplemental investigation also was inadequate, and therefore began her own investigation. She interviewed numerous current and former Directorate employees, and sought affidavits that would support her allegations of discrimination. In April, 1974, Copeland submitted her findings, and her comments on the Department's supplemental investigation, to the EEO Director.

Assistant Secretary Fred G. Clark submitted his proposed disposition of the complaint in June, 1974. That disposition would have removed all adverse references from her personnel file, but it proposed no other significant relief.

Copeland, still dissatisfied, requested a formal hearing. Her file was sent to the Civil Service Commission for that purpose, but no hearing was held. The file was returned to the Labor Department without explanation.

Assistant Secretary Clark resubmitted his proposed disposition of Copeland's complaint in September, 1974. According to Copeland, she was assured that because she had already requested a hearing, she need not repeat that request.

No hearing was held, however, and the Department issued its final decision on November 7, 1974. The final decision conceded "that a pattern of sex discrimination exists" in the Directorate, and that such discrimination "manifests itself in the lack of leadership responsibility assignments given to qualified women professionals." The decision, however, denied that the Department's refusal to promote Copeland resulted from sex discrimination and asserted that Copeland's personal disagreements with her supervisors were the true cause of her grievances.

The Department in its decision agreed, inter alia, to (1) consider her fairly for future work assignments; (2) clarify her responsibilities and objectively assess her performance of them, (3) expunge adverse evaluations from her personnel file, and (4) monitor future promotion decisions to insure fair treatment for her and other minority employees and women. The Department did not, however, offer retroactive promotion and back pay, or priority consideration for future promotions.

The Department sent Copeland a copy of its decision. However, Copeland's attorneys were not served with a copy, in violation of Department regulations.

Copeland filed this class suit in the District Court on December 13, 1974. The complaint, as amended, alleged three gender discrimination counts, namely, violations of (1) Title VII of the Civil Rights Act of 1964, (2) Executive Order 11478, and (3) rights under the first and fifth amendments to the constitution and 42 U.S.C. section 1985. The complaint also alleged a count of race discrimination under the first and fifth amendments and 42 U.S.C. sections 1981 and 1985.

1. The Government's Motion for Judgment on the Pleadings

The government promptly moved for judgment on the pleadings under a variety of theories. Judgment on the Title VII count was sought because the suit was filed 31 days after Copeland received notice of the final agency decision, not within the 30-day period established by statute. See 42 U.S.C. § 2000e-16(c) (1976). The District Court held, however, that the government's failure to serve Copeland's attorneys with the agency decision tolled the running of the 30-day period.5

2. The Government's Opposition to Class Certification

Copeland next moved that she represent a class of all past, present, and future female data processing employees in the Directorate. The government, however, moved to remand the case to the Civil Service Commission for additional hearings and, in the alternative, opposed class certification for a variety of reasons.

The District Court denied the motion to remand. The court also certified the case as a class suit, covering all females employed by the Directorate in data processing positions after June 11, 1971.

Copeland's attorneys meanwhile had propounded a congeries of discovery requests, including interrogatories and requests for production of documents. These discovery requests prompted an acrimonious flurry between the plaintiff class (plaintiff) and the defendant.

The government initially did not comply with these requests. Plaintiff moved to compel discovery. The government then answered some of the interrogatories, but objected to certain others that it thought called for privileged information. The government, accordingly, opposed the motion to compel.

Plaintiff pointed out to the court that the Department had destroyed certain relevant documents6 and that, in any event, the government's responses to many interrogatories were inadequate. The question of the adequacy of the government's response to discovery requests generally was ultimately resolved by negotiation.

Meanwhile, the government had initiated discovery of its own. The government propounded interrogatories, requested documents, and took depositions. Plaintiff continued the discovery battle by noticing the deposition of an Assistant Secretary of Labor. The government moved for a protective order; this motion was denied.

Discovery continued for several additional weeks. Plaintiff answered defendant's numerous interrogatories, served additional interrogatories of its own, and noticed further depositions. The government again sought a protective order; the District Court ordered the government to supply any requested documents and information that were relevant and nonprivileged.

The District Court all along had envisioned that discovery would be completed in time for the liability trial to begin February 16, 1976. The government, however, asked in January for a one-month delay in trial, in part "due to the extreme complexity of the issues and evidence in the case."7 The District Judge, however, insisted that parties promptly finish discovery and prepare for the liability trial on February 16 as originally planned.

Plaintiff orally complained to the court on January 26 about additional discovery difficulties. Plaintiff alleged that the government failed to identify and produce certain highly relevant documents, and requested that the court grant judgment on the merits as the sanction for nondisclosure.

Plaintiff has ample ground to complain. Her systematic discovery efforts initiated months ago have been impeded unnecessarily and she has been forced to expend time and effort to fill in gaps in the proof which the documents would have largely avoided had they been produced as they should have been.

The court nevertheless denied the motion for sanctions, without prejudice, "as representing too extreme a sanction on the basis of facts presently available." The parties at this point continued to plan for a February 16 trial.

Instead of going to trial, however, the parties settled the liability issue. Now three years after Copeland first complained of discrimination, the government finally conceded that the Directorate had

subjected (Copeland) and the other members of the class to sex-based discrimination in assignments, training, performance evaluations, promotions and working conditions, all in violation of Title VII ....8

The government also agreed to develop and put into effect a court-approved affirmative action program.9

The stipulation provided for a trial on relief to each of the individual plaintiffs. In those trials, the government would carry the burden of proving that the conceded sex discrimination had not "monetarily or otherwise" affected the particular plaintiff.

5. Trial on Copeland's Claim for Retroactive Promotion and Back Pay

Shortly after the government stipulated it had discriminated on the basis of sex, a six-day trial ensued on the relief vel non due plaintiff Copeland. The government contended that Copeland in any event would not have been promoted to GS-14, because Copeland's failure to receive promotions and training was attributable to her poor work, lack of qualifications, and personality problems.

The District Court found, however, that the government had failed to prove that sex discrimination did not play a part in Copeland's lack of advancement. The court, accordingly, awarded her a promotion to GS-14 and $6,169.80 in back pay. The court also ordered the Department to provide Copeland with training and assignments commensurate with her position.

6. Litigation Before a Special Master on Retroactive Promotion and Back Pay for Other Class Members

The parties stipulated to the appointment of a Special Master to receive evidence and report to the District Court on the relief due the other members of the class.

Each side initiated a new round of discovery on the issues presented to the Special Master. Plaintiff propounded additional interrogatories, requested admissions from defendant, and noticed further depositions. Defendant also propounded more interrogatories, requested admissions, and sought more documents.

After this substantial additional discovery, the parties settled the remaining individual claims. The settlements generally required promotions, back pay, the opportunity to participate in a training program, or some combination of the above. Approximately $33,000 in back pay was obtained.

Meanwhile, the parties haggled over the terms of the affirmative action program. The government proposed a plan; plaintiff criticized it as inadequate. The District Court held a hearing to discuss problems with the government's plan.10

Plaintiff later proposed its own affirmative action program. The government criticized it, and the District Court held another hearing.11 The following day, the District Court ordered the parties to negotiate a mutually satisfactory plan, using defendant's draft as the starting point, but incorporating various modifications sought by plaintiff.12 On August 1, 1976, the District Court approved a 36-page affirmative action plan negotiated by the parties.

8. Plaintiff's Application for an Attorney's Fee

On November 30, plaintiff filed a documented request for costs and an attorney's fee. The documentation revealed that plaintiff's attorneys had spent 3,602 hours on the case and that, if that time were billed at the law firm's customary hourly rates, the legal fee would be about $206,000. In papers filed December 20, 1976, the government opposed "an(y) award even approaching" $206,000. Apparently content to submit the attorney's fee issue to the judge on the papers, the government did not ask the District Court to hold a hearing.

On January 6, 1977, the District Court entered an order awarding a $160,000 fee, an amount approximately 22% less than that envisioned by plaintiff's papers. The order was accompanied by a four-page memorandum analyzing the fee request. The District Court wrote, in pertinent part:

The Secretary apparently believes a fee award in a case of this type should be based primarily upon the monetary results achieved. This is an erroneous approach to the fee problem. While the actual cash awards to individual members of the class were in this instance relatively small in relation to the total fee claim, this was basically an equity action which was intended to and did achieve benefits that cannot be measured solely in monetary terms. The judgment, which has not been appealed, among other things established an entirely new pattern of training and promotion for female employees in an important segment of the Department of Labor which had blatantly discriminated against women. The benefits of the litigation will be felt for many years to come.

While the Secretary now suggests that there were really no serious issues at stake, this is not borne out by the facts. The litigation went forward in a relatively civilized manner but it was hard fought. The Government offered firm, persistent resistance throughout the litigation and concessions developed only as it became apparent there was little prospect of Government success. Indeed, the Government moved to dismiss at the outset, and it opposed discovery. There were many difficulties encountered during the discovery process which were caused, in part, by the Department's inadvertent destruction of certain records contrary to Court direction and the intentional withholding of other documents by some officials of the Department of Labor, as well as by the complexity of the issues.

The 3,602 hours were logged almost entirely by associates of the firm with varying degrees of experience. The average rate of $57.17 an hour is well within the local range for associates of larger firms .... What plaintiffs' counsel lacked in seasoned trial experience was offset by other factors. They were always well prepared, effective and knowledgeable. No time was deliberately wasted and counsel proceeded with full recognition of the congressional directive to expedite litigation of this type.

Billing for legal services, however, should not be a merely mechanical exercise. Where a fee is sought from the United States, which has infinite ability to pay, the Court must scrutinize the claim with particular care. When an application such as this is filed by a large law firm computing a proposed award by use of "customary rates," the firm has obviously made little, if any, effort to exercise billing judgment. Thus an important ingredient is lacking. A reasonable fee can only be fixed by the exercise of judgment, using the mechanical computations simply as a starting point to reach a higher or lower figure. The Court must perform this function.

In considering what is a reasonable fee in this instance a number of factors deserve special mention. The proposed fee absorbs not only expensive overhead such as rent and secretarial services, but no charge has been made for what was undoubtedly a substantial amount of time spent by paralegals who play such a useful role in large documentary cases. On the other hand, there was practically no partner time expended on this case and the associates lacked experienced trial direction. The Court must also take into account the fact that not all of the work proved productive. Some issues which were joined in the complaint were dropped, as were some individual defendants. Taking into account each of the factors itemized in Evans v. Sheraton Park Hotel, 503 F.2d 177 (D.C.Cir.1974), including the matters specifically mentioned, the Court has concluded that a reasonable fee in this litigation, weighing the results achieved, the novelty of the issues, the difficulties encountered and the effectiveness of the excellent representation given is $160,000.II

Title VII of the Civil Rights Act of 1964 allows the prevailing party to receive from the loser a reasonable attorney's fee in addition to other relief. The statute provides:

In any action or proceeding under ... (Title VII) the court, in its discretion, may allow the prevailing party, other than the (Equal Employment Opportunity) Commission or the United States, a reasonable attorney's fee as part of the costs, and the Commission and the United States shall be liable for costs the same as a private person.

42 U.S.C. § 2000e-5(k) (1976).

The availability of an attorney's fee encourages individuals injured by discrimination to seek judicial redress.13 As the Supreme Court explained:

When the Civil Rights Act of 1964 was passed, it was evident that enforcement would prove difficult and that the Nation would have to rely in part upon private litigation as a means of securing broad compliance with the law. A Title II suit is thus private in form only.... If (a plaintiff) obtains an injunction, he does so not for himself alone but also as a "private attorney general," vindicating a policy that Congress considered of the highest priority. If successful plaintiffs were routinely forced to bear their own attorneys' fees, few aggrieved parties would be in a position to advance the public interest by invoking the injunctive powers of the federal courts. Congress therefore enacted the provision for counsel fees not simply to penalize litigants who advance arguments they know to be untenable but, more broadly, to encourage individuals injured by racial discrimination to seek judicial relief ....

Newman v. Piggie Park Enterprises, Inc., 390 U.S. 400, 401-02, 88 S. Ct. 964, 966-67, 19 L. Ed. 2d 1263 (1968) (footnotes omitted)14; accord, New York Gaslight Club, Inc., v. Carey, 447 U.S. 54, 60-66, 100 S. Ct. 2024, 2029-2032, 64 L. Ed. 2d 723 (1980).

Confronted by the explicit language of the statute and its accompanying legislative history, the government in the instant case concedes that plaintiff is entitled to an attorney's fee. Indeed, the parties so stipulated during the course of the lawsuit. At issue in this appeal is whether the District Court's fee award was reasonable.

The Court of Appeals for the Fifth Circuit explained, in general terms, how the fee should be calculated under Title VII in Johnson v. Georgia Highway Express, Inc., 488 F.2d 714 (1974). In Johnson, the court suggested that district courts base fee awards on the following criteria: (1) the time and labor required; (2) the novelty and difficulty of the questions; (3) the skill requisite to perform the legal services properly; (4) the preclusion of other employment; (5) the customary fee in the community for similar work; (6) the fixed or contingent nature of the fee; (7) time limitations imposed by the client or the circumstances; (8) the amount involved and the results obtained; (9) the experience, reputation, and ability of the attorneys; (10) the undesirability of the case; (11) the nature and length of the professional relationship with the client; and (12) awards in similar cases. Id. at 717-19.

We recognized the importance of considering the twelve Johnson factors in awarding fees in Evans v. Sheraton Park Hotel, 503 F.2d 177, 187-88 (1974). Many other courts have applied the Johnson factors in subsequent cases, and those factors remain central to any fee award.15

Simply to articulate those twelve factors, however, does not itself conjure up a reasonable dollar figure in the mind of a district court judge. A formula is necessary to translate the relevant factors into terms of dollars and cents. This is particularly true because the twelve factors overlap considerably. For example, largely subsumed under the factor "time and labor required" is an assessment of the "difficulty of the questions." That is so because the more difficult the problem, the longer it will take adequately to solve it. Similarly, the customary hourly fee (Johnson factor # 5) is likely to be influenced by (# 3) the level of skill necessary to perform the services, (# 6) whether the fee is fixed or contingent, (# 7) time limitations, (# 8) the amount to be obtained, (# 9) the reputation of the attorneys, and (# 10) the undesirability of the case.

For these reasons, scholars have noted that the twelve Johnson factors, without more, cannot guarantee a rational setting of fees. One commented:

The fundamental problem with an approach that does no more than assure that the lower courts will consider a plethora of conflicting and at least partially redundant factors is that it provides no analytical framework for their application. It offers no guidance on the relative importance of each factor, whether they are to be applied differently in different contexts, or, indeed, how they are to be applied at all.

Berger, Court Awarded Attorneys' Fees: What is "Reasonable" ?, 126 U.Pa.L.Rev. 281, 286-87 (1977) (footnotes omitted); accord, Dawson, Lawyers and Involuntary Clients in Public Interest Litigation, 88 Harv.L.Rev. 849, 927 & n.327 (1975); Note, Promoting the Vindication of Civil Rights Through the Attorney's Fees Awards Act, 80 Colum.L.Rev. 346, 372-73 & nn. 164-69 (1980).

District court judges for this reason have had difficulty applying the Johnson factors. A common, yet understandable, fault is for the trial judge to make the conclusory statement, "After considering each of the twelve factors in Johnson, I find that a reasonable fee is X dollars." This very often leads to reversal and remand. See, e. g., Gay v. Board of Trustees, 608 F.2d 127, 128 (5th Cir. 1979); Davis v. Fletcher, 598 F.2d 469, 470-71 (5th Cir. 1979).

Appellate courts have recognized that the Johnson factors, despite their substantial conceptual value, also are imprecise.16 Some courts, therefore, have incorporated the twelve factors into an analytical framework that can be easily applied by trial courts and that will make possible meaningful appellate review.

Any fee-setting formula must produce an award sufficient to fulfill the primary purpose of awarding fees in Title VII cases, namely, "to encourage individuals injured by ... discrimination to seek judicial relief." Piggie Park, 390 U.S. at 402, 88 S.Ct. at 966. An award of fees provides an incentive to competent lawyers to undertake Title VII work only if the award adequately compensates attorneys for the amount of work performed. The Court of Appeals for the Third Circuit was the first to develop a fee-setting formula that reflects this principle. In Lindy Bros. Builders, Inc. v. American Radiator & Standard Sanitary Corp., 487 F.2d 161 (1973) (Lindy I), and its successor case, Lindy II, 540 F.2d 102 (1976) (en banc), the Third Circuit articulated a formula that considered all the relevant factors but eliminated the redundancy and imprecision that many have identified in other fee-setting schemes.

Lindy recognized that the starting point in fee setting-what it characterized as the "lodestar" fee-should be computed by multiplying a reasonable hourly rate by the number of hours reasonably expended on the lawsuit. 487 F.2d at 167. Adjustments in this figure are appropriate, the court recognized, but the "lodestar" provides "the only reasonably objective" starting point for awarding a fee. Id.

In National Treasury Employees Union v. Nixon, 521 F.2d 317 (D.C.Cir. 1975), this court had occasion again to consider fee calculation. We recognized Lindy's important analytical contribution to the inquiry, and we adopted its framework for use in this circuit. We said:

The inquiry begins with a determination of the time devoted to the litigation. This figure in turn is multiplied by an hourly rate for each attorney's work component, a rate which presumably would take into account the attorney's legal reputation and experience. The resulting figure represents an important starting point because it "provides the only objective basis for valuing an attorney's services" (citing Lindy ).

Id. at 322 (footnote omitted).

Myriad cases involving court-awarded fees continue to come before the district court judges and, ultimately, before this court.17 We therefore take this opportunity en banc to elaborate, to a greater extent than we have in the past, on the appropriate mechanism for calculating an attorney's fee pursuant to statutes like Title VII.

A. The "Lodestar"

Any fee-setting inquiry begins with the "lodestar": the number of hours reasonably expended multiplied by a reasonable hourly rate. The figure generated by that computation is the basic fee from which a trial court judge should work. We examine below some of the problems that arise in calculating the "lodestar."

The fundamental purpose of the fee award is to compensate the attorney for his efforts. The first task for the trial court judge, therefore, is determining the amount of time reasonably expended.

When a law firm seeks a fee, it should document the amount of work performed. The District Court then will be able to do more than merely lump together all the hours spent by the various attorneys associated with the enterprise; the judge instead can segregate into categories the kinds of work performed by each participating attorney. This project need not be unduly burdensome:

It is not necessary to know the exact number of minutes spent nor the precise activity to which each hour was devoted nor the specific attainments of each attorney. But without some fairly definite information as to the hours devoted to various general activities, e. g., pretrial discovery, settlement negotiations, and the hours spent by various classes of attorneys, e. g., senior partners, junior partners, associates, the court cannot know the nature of the services for which compensation is sought.

Lindy I, 487 F.2d at 167.

Compiling raw totals of hours spent, however, does not complete the inquiry. It does not follow that the amount of time actually expended is the amount of time reasonably expended. In the private sector, "billing judgment" is an important component in fee setting. It is no less important here. Hours that are not properly billed to one's client also are not properly billed to one's adversary pursuant to statutory authority. Thus, no compensation is due for nonproductive time. For example, where three attorneys are present at a hearing when one would suffice, compensation should be denied for the excess time. Similarly, no compensation should be paid for time spent litigating claims upon which the party seeking the fee did not ultimately prevail.18

At this point in the computation, the District Judge might usefully construct a table that looks something like this example.

Attorney & Type of Work                 Hours
-----------------------                 -----
Senior Partner: Court Appearances       17.3
Senior Partner: Review of pleadings     39.2
Junior Associate: Research & drafting   87.6
Junior Associate: Depositions           35.5

The remaining element in fixing a "lodestar" fee is the reasonable hourly rate.

The reasonable hourly rate is that prevailing in the community for similar work.19 As we noted a reasonable hourly rate is the product of a multiplicity of factors. Evans itself listed several of the relevant considerations: the level of skill necessary, time limitations, the amount to be obtained in the litigation, the attorney's reputation, and the undesirability of the case. See Evans, 503 F.2d at 187-88. It follows that there may be more than one reasonable hourly rate for each of the attorneys, and for each of the kinds of work, involved in the litigation. After receiving documentation and other submissions,20 and perhaps holding a hearing,21 the trial judge might complete the fee table in the following manner.

Thus, the "lodestar" fee in this hypothetical is $9,899.50.

B. Adjustments to the "Lodestar"

The "lodestar" fee may be adjusted to reflect other factors. We discuss herein those applicable in Title VII and similar fee-setting cases.22 The burden of justifying any deviation from the "lodestar" rests on the party proposing the deviation. Lindy II, 540 F.2d at 118.

Under statutes like Title VII, only the prevailing party is eligible for a court-awarded fee. An attorney contemplating representation of a Title VII plaintiff must recognize that no fee will be forthcoming unless the litigation is successful. An adjustment in the lodestar, therefore, may be appropriate to compensate for the risk that the lawsuit would be unsuccessful and that no fee at all would be obtained.

It is important to recognize that the contingency adjustment is designed solely to compensate for the possibility at the outset that the litigation would be unsuccessful and that no fee would be obtained. Contingency adjustments of this sort are entirely unrelated to the "contingent fee" arrangements that are typical in plaintiffs' tort representation. In tort suits, an attorney might receive one-third of whatever amount the plaintiff recovers. In those cases, therefore, the fee is directly proportional to the recovery. Such is not the case in contingency adjustments of the kind we describe herein. The contingency adjustment is a percentage increase in the "lodestar" to reflect the risk that no fee will be obtained. The contingency adjustment is not a percentage increase based on the amount of recovery. Merola v. Atlantic Richfield Co., 515 F.2d 165, 169 (3d Cir. 1975).

To the extent, of course, that an hourly rate underlying the "lodestar" fee itself comprehends an allowance for the contingent nature of the availability of fees in Title VII litigation against the Government, no further adjustment duplicating that allowance will be made. The district judge has ample powers of inquiry into the make-up of hourly rates to assure that the Government will not suffer from any such duplication or, indeed, from any excessive allowance for this purpose.

The delay in receipt of payment for services rendered is an additional factor that may be incorporated into a contingency adjustment. The hourly rates used in the "lodestar" represent the prevailing rate for clients who typically pay their bills promptly. Court-awarded fees normally are received long after the legal services are rendered. That delay can present cash-flow problems for the attorneys. In any event, payment today for services rendered long in the past deprives the eventual recipient of the value of the use of the money in the meantime, which use, particularly in an inflationary era, is valuable. A percentage adjustment to reflect the delay in receipt of payment therefore may be appropriate. Lindy II, 540 F.2d at 117.23

To the district court judge falls the task of calculating as closely as possible a contingency adjustment with which fairly to compensate the successful attorney. We have not, however, lost sight of the fact that this adjustment is inherently imprecise and that certain estimations must be made. For example, it is difficult in hindsight to determine the risk of failure at the commencement of a lawsuit that ultimately proved to be successful. Thus, we ask only that the district court judges exercise their discretion as conscientiously as possible, and state their reasons as clearly as possible.24

Next, the "lodestar" may be adjusted up or down to reflect "the quality of representation." It is important to make clear precisely the analysis that must accompany such an adjustment. A quality adjustment is appropriate only when the representation is unusually good or bad, taking into account the level of skill normally expected of an attorney commanding the hourly rate used to compute the "lodestar." In other words,

the court must recognize that a consideration of "quality" inheres in the "lodestar" award: counsel who possess or who are reputed to possess more experience, knowledge and legal talent generally command hourly rates superior to those who are less endowed. Thus, the quality of an attorney's work in general is a component of the reasonabl(e) hourly rate; this aspect of "quality" is reflected in the "lodestar" and should not be utilized to augment or diminish the basic award under the rubric of "the quality of an attorney's work."

Lindy I, then permits an adjustment to the "lodestar"-up or down-based on the all-around performance of counsel in the specific case: "Any increase or decrease in fees to adjust for the quality of work is designed to take account of an unusual degree of skill, be it unusually poor or unusually good." 487 F.2d at 168. By this is meant simply that the district court may determine that the lawyer discharged the professional burden undertaken with a degree of skill above or below that expected for lawyers of the caliber reflected in the hourly rates.

Lindy II, 540 F.2d at 117-18 (emphasis in original).

Until now the calculations have entirely ignored the results of the litigation. Success was a threshold inquiry relevant to the entitlement vel non to a fee, but the amount or nature of recovery was not considered in setting the "lodestar." These latter factors should be considered now, under the rubric of "quality of representation."

Where exceptional results are obtained-taking into account the hourly rate commanded and number of hours expended-an increase in fee is justifiable. However, it is important again to emphasize that a huge dollar recovery does not itself justify a huge fee award. The "lodestar" itself generally compensates lawyers adequately for their time. An upward adjustment for quality is appropriate only when the attorney performed exceptionally well, or obtained an exceptional result for the client. For example, if a substantial monetary judgment was to be expected, that expectation normally is reflected in the hourly rate used to compute the "lodestar," and no further adjustment would be necessary.

Quality adjustments may be upward or downward. Thus, if a high-priced attorney performs in a competent but undistinguished manner, a decrease in the "lodestar" may be necessary under the "quality of representation" rubric because the hourly rate used to calculate the "lodestar" proved to be overly generous.

Copeland I and Copeland II, however, took an entirely different view from that expressed in this opinion. The fee approach we have described rests on compensating attorneys for the market value of services rendered. The panel had the notion that, at least where the government is the losing defendant, the fee should be the amount representing the "actual cost to the law firm plus a reasonable and controllable profit" for the legal work done. Copeland II, slip op. at 5 (emphasis deleted).

We think, however, that the approach articulated earlier in this opinion represents the proper formula for the setting of fees regardless of the defendant's identity. We explain below why we think that fees should be calculated no differently when the government (rather than a private party) is the losing defendant. We then explain the difficulties we have, in any event, with the panel's "cost-plus" approach.

The panel opinions suggested that, where the government is the losing defendant, a fee award should be subject to greater scrutiny-i. e., the fee should be lower-than one against a private defendant. E. g., Copeland II, slip op. at 3. We agree that a judge setting any award should scrutinize the amount with care. But we do not think that the amount of the fee should depend on the identity of the losing party. Our conclusion is based on both the language of the statute and the policies that underlie it.

Our starting point, of course, is the statutory text. The attorney's fee section provides that, in any Title VII action,

the court, in its discretion, may allow the prevailing party ... a reasonable attorney's fee as part of the costs, and the ... United States shall be liable for costs the same as a private person.

42 U.S.C. § 2000e-5(k) (1976) (emphasis added).25

The language of the statute indicates that the calculation of the fee should not vary with the identity of the losing defendant, and the policies underlying the attorney's fee provision are fully consistent with this plain language. Those policies, as we have seen, are two. The primary purpose is to help persons obtain competent counsel with which to vindicate civil rights through litigation. E. g., Piggie Park, 390 U.S. at 402, 88 S.Ct. at 966. Nothing in the statute suggests that the incentive to ferret out discrimination, provided by the prospect of an attorney's fee, should be any less when the government is the defendant. If anything, it is even more important to provide adequate fees to employment discrimination litigants who prevail against the government. In Parker v. Califano, 561 F.2d 320 (D.C. Cir. 1977), this court observed that Title VII litigants against the government face greater obstacles than litigants against private defendants. We noted:

Unlike private sector employees, federal employee complainants are not merely private attorneys general; they are the only attorneys general under the enforcement scheme adopted in Section 717, 42 U.S.C. § 2000e-16 (Supp. V 1975). Suits in behalf of federal employees by the Attorney General or EEOC are not authorized against federal agencies. Indeed, the Attorney General is frequently counsel for the other side. Also unlike private sector employees, federal employees must first bring their employment discrimination grievances, not to an independent state or local administrative body or to EEOC, but to the very agency about whose practices they are complaining.

Id. at 331 (footnotes omitted).

A second policy also underlies fee awards. As we have noted, the prospect of liability for an attorney's fee may help deter discrimination26 and thereby obviate litigation. We do not think that the incentive for the government to refrain from discrimination should be any less than for private employers.

Finally, we note that various courts have expressed views consistent with those we express today. For example, in an age discrimination suit against the City of Philadelphia, the Court of Appeals for the Third Circuit observed:

The fact that the City of Philadelphia's tax revenues must pay the fees award does not warrant special standards for public and private employers. The reasonable value of an attorney's time does not depend on who his or her adversary is.

Rodriguez v. Taylor, 569 F.2d 1231, 1249 n.32 (3d Cir. 1977), cert. denied, 436 U.S. 913, 98 S. Ct. 2254, 56 L. Ed. 2d 414 (1978); see Dennis v. Chang, 611 F.2d 1302, 1304-07 (9th Cir. 1980); King v. Greenblatt, 560 F.2d 1024, 1025-26 n.2 (1st Cir. 1977), cert. denied, 438 U.S. 916, 91 S. Ct. 3146, 57 L. Ed. 2d 1161 (1978).27

In sum, we agree with the panel opinions that the government has a "deep pocket" and that any fee request should be examined with care. But we think, for the reasons stated above, that fees should be neither lower, nor calculated differently, when the losing defendant is the government.

B. Difficulties with the "Cost-Plus" Approach

The panel opinions in any event suggested that the "cost-plus" method of calculating fees might usefully be applied in all cases, regardless of the identity of the defendant. See Copeland II, slip op. at 18-19. The panel opinions, in brief, thought that a fee should be based on "the sums paid out to (the) attorneys as personal income and to defray overhead costs attributable to the maintenance of the attorneys in the firm," plus a "reasonable and controllable margin for profit." Copeland I, 594 F.2d at 251.

We think, however, that the standards we discussed earlier in this opinion are those that should govern all fee-setting cases under the statute. The "lodestar," or "market value," method of fee setting has the virtue of being relatively easy to administer. We do not want

a district court, in setting an attorneys' fee, (to) become enmeshed in a meticulous analysis of every detailed facet of the professional representation. It ... is not our intention that the inquiry into the adequacy of the fee assume massive proportions, perhaps even dwarfing the case in chief.

Lindy II, 540 F.2d at 116. We fear that the proposed "cost-plus" method of calculating fees would indeed become the inquiry of "massive proportions" that we strive to avoid. The problems associated with administering a "cost-plus" calculus are multifarious. How might a firm allocate its overhead costs to a particular piece of litigation? In what manner does one calculate the costs associated with the "imputed salaries" of firm partners? What is a "reasonable" profit to be awarded?28 The necessity, under "cost-plus," of answering these and other questions creates the specter of a monumental inquiry on an issue wholly ancillary to the substance of the lawsuit.

To address questions like these, considerable discovery would be necessary to obtain documentary evidence. A law firm's financial structure is highly relevant to a "cost-plus" inquiry, so the firm's financial records would be discoverable. Third-party and expert testimony would have to be proffered. Because time spent litigating the fee request is itself compensable,29 the depth of the inquiry ironically might lead to an increase, rather than a diminution, in fee awards.30

The prospect of enduring an inquiry of this scope might discourage competent counsel from undertaking Title VII representation at all. This possibility cannot be tolerated in light of Title VII's purpose "to encourage individuals injured by ... discrimination to seek judicial relief." Piggie Park, 390 U.S. at 402, 88 S.Ct. at 966.

In sum, the "cost-plus" system poses considerable administrative difficulties. But our rejection of its thesis does not depend on administrative inconvenience alone. We think the theoretical basis of "cost-plus" is fundamentally inconsistent with Congress' purpose in providing for statutory fee-shifting. A fee should be based on the market value of services rendered, not on some notion of "cost" incurred by the law firm. That is the conclusion both of the courts that have spoken on the issue, and also Congress itself.

As we have seen, courts attempting to be faithful to Congress' wishes have taken as the starting point in fee calculation the prevailing hourly rate multiplied by the number of hours reasonably expended. An amount calculated in that manner is adequate to entice competent counsel to undertake representation.

Courts had articulated this principle often, and consistently, when Congress passed the Civil Rights Attorney's Fees Awards Act of 1976, 42 U.S.C. § 1988 (1976). That Act was passed in the wake of the Supreme Court's decision in Alyeska Pipeline Service Co. v. Wilderness Society, 421 U.S. 240, 95 S. Ct. 1612, 44 L. Ed. 2d 141 (1975), which sharply restricted courts' ability, in the absence of statutory authority, to award an attorney's fee to the prevailing party.31 Congress understood and noted with approval that in the Title VII context courts used a "market value" approach to the award of fees. According to the Senate Report accompanying the 1976 Act, that approach provides the "appropriate standards" for awarding fees. S.Rep. No. 1011, 94th Cong., 2d Sess. 6 (1976), U.S.Code Cong. & Admin.News 1976, p. 5908. See H.R.Rep. No. 1558, 94th Cong., 2d Sess. 8 & n.16 (1976), U.S.Code Cong. & Admin.News 1976, p. 5908.

Even if we were free to ignore Congress' intent, we see no reason now to abandon the "market value" approach to fee calculation in favor of "cost-plus." In the instant case, for example, nothing in the record suggests that "cost-plus" would be a preferable method of setting fees. The government's primary contention all along in this case has been simply that too many hours were spent by plaintiff's attorneys. The District Court judge agreed that too much time had indeed been spent, and he reduced the requested award accordingly.32 The panel opinions are fraught with concern that too many hours were spent on this case; ironically, however, nothing in the "cost-plus" scheme is more effective than the "market value" approach in reducing the fee to reflect such wasted hours. As the government recognizes, "cost-plus" in effect is simply a new method for setting the rate of compensation.33 Nothing in it is of any value to trial judges in determining the number of hours properly spent.34

More fundamentally, nothing in the panel opinions explains why, in the typical case, rates established by the pressures of the market mechanism will differ from those resulting under "cost-plus."35 If "cost-plus" is simply another method of reaching the same result as under "market value" calculations, the added administrative burden seems not worth the candle. On the other hand, if "cost-plus" somehow produces different results from those obtaining under "market value" calculations, "cost-plus" is inconsistent with the wishes of Congress and the overwhelming view of courts that have considered the matter.

The one circumstance in which "cost-plus" almost certainly will yield a different fee from that under "market value" is where the successful plaintiff was represented by a "public interest" law firm. Such groups often represent their clients for low fees, or for no fee at all. Consequently, the individual attorneys at those organizations typically are compensated at rates far below those prevailing in the marketplace. "Cost-plus" calculations, focusing as they do on lawyers' salaries rather than on the value of services rendered, in this instance will yield lower fee awards than those under a "market value" system.

Copeland II argued that this result was entirely reasonable. The opinion pointed out that a fee-setting mechanism that awarded a firm its costs, plus a reasonable profit, could not be condemned as penurious because any such amount, almost by definition, provided the attorneys the same compensation that they normally received. Copeland II, slip op. at 14, 15. We think, nevertheless, that the proper focus is the market value of services rendered, regardless of the notion of "costs" that the panel seemed to think so important.

We should have thought this issue resolved in this circuit as far back as the seminal decision in Evans. In that case, we agreed with the Fifth Circuit that the applicable rate was that prevailing "for similar work in the community." Johnson, 488 F.2d at 718; see Evans, 503 F.2d at 187-88. Nowhere in Johnson or Evans are salaries mentioned as relevant to the calculus. Moreover, this court en banc had earlier explained:

It may well be that counsel serve organizations like appellants for compensation below that obtainable in the market because they believe the organizations further a public interest. Litigation of this sort should not have to rely on the charity of counsel .... The attorneys who worked on this case should be reimbursed the reasonable value of their services ....

Wilderness Society v. Morton, 495 F.2d 1026, 1037 (1974), rev'd on other grounds sub nom. Alyeska Pipeline Service Co. v. Wilderness Society, 421 U.S. 240, 95 S. Ct. 1612, 44 L. Ed. 2d 141 (1975); accord, National Treasury Employees Union v. Nixon, 521 F.2d 317, 322-23 (D.C.Cir. 1975). Indeed, to appraise the reasonable value of an attorney's time by

(r)eference to absolute salary levels is about as reasonable as deriving the reasonable value of a federal judge's time from his or her salary.

Rodriguez v. Taylor, 569 F.2d 1231, 1248 (3d Cir. 1977), cert. denied, 436 U.S. 913, 98 S. Ct. 2254, 56 L. Ed. 2d 414 (1978).

Despite this precedent, the panel opinions contended that the market value of time expended should be irrelevant. Because the purpose of the fee award is to provide a sufficient incentive to counsel to litigate, a formula based on "cost plus reasonable profit will guarantee (public interest practitioners) a return that will be at least equal to that received in their usual practice." Copeland II, slip op. at 15. (emphasis in original). Thus, the panel correctly concluded that paying low-salaried attorneys the prevailing market rate normally will yield a larger fee than that to which they are accustomed.

For a variety of reasons, we see this as no flaw. First, Congress has indicated that public interest lawyers in these sorts of cases should be compensated by using a market value approach. The Senate Judiciary Committee, in its report endorsing the 1976 fees act, draws no distinction in awarding fees between public interest law firms and private attorneys. S.Rep. No. 1011, 94th Cong., 2d Sess. 6 (1976). That report cites with approval Davis v. County of Los Angeles, 8 Empl.Prac.Dec. 5047 (C.D.Cal.1974), a Title VII case in which the court said:

(I)t is not legally relevant that plaintiffs' counsel ... are employed by the Center for Law In The Public Interest, a privately funded non-profit public interest law firm. It is in the interest of the public that such law firms be awarded reasonable attorneys' fees to be computed in the traditional manner ....

Id. at 5048-49. Similarly, the House Report endorsed other cases that said awards of fees to civil rights law firms should be equal to those awarded to members of the private bar. See H.R.Rep. No. 1558, 94th Cong., 2d Sess. 8 n. 16 (1976), and cases cited therein.

Second, the purpose of the legislative scheme of the Civil Rights Act of 1964 will be served by computing fees based on a "market value" approach. The purpose of Title VII's fee award provision, as we have seen, is to encourage the private enforcement of the civil rights laws. While some lawyers would assist in the private enforcement of Title VII for a reduced fee, Congress has recognized that payment of full fees will provide greater enforcement incentives. Full fee awards to public interest law firms help finance their work, both in the instant case, and in others. Indeed, fee awards (paid by proven discriminators) may help reduce the subsidies (paid from the public fisc) that some of these organizations receive.

Third, to compute fees differently depending on the identity of the successful plaintiff's attorney might result in two kinds of windfalls to defendants. The incentive to employers not to discriminate is reduced if diminished fee awards are assessed when discrimination is established. Moreover, where a public interest law firm serves as plaintiff's counsel (a law firm that, under the panel's approach, will not obtain the full value of its services from the losing defendant) the defendant will be subject to a lesser incentive to settle a suit without litigation than would be the case if a high-priced private firm undertook plaintiff's representation. Dennis v. Chang, 611 F.2d 1302, 1307 (9th Cir. 1980). That is so because the marginal cost of each hour of continued litigation would be reduced. Defendant's counsel could inundate the plaintiff with discovery requests without fear of paying the full value of the legal resources wasted in response. We do not think that Title VII intended that defendants should have an incentive to litigate imprudently simply because of the fortuity of the identity of plaintiff's counsel.

Fourth, we note that the vast majority of courts that have considered this issue agrees with us that attorney's fees should not be based on the costs of the successful party. Instead, fees should be based on the market value of the legal services rendered. Oldham v. Ehrlich, at 168-169 (8th Cir. March 12, 1980); Palmigiano v. Garrahy, at 599-603 (1st Cir. 1980); Dennis v. Chang, 611 F.2d 1302, 1309 (9th Cir. 1980); Carey v. New York Gaslight Club, Inc., 598 F.2d 1253, 1255 n. 1 (2d Cir. 1979), aff'd, 447 U.S. 54, 100 S. Ct. 2024, 64 L. Ed. 2d 723 (1980); Reynolds v. Coomey, 567 F.2d 1166, 1167 (1st Cir. 1978); Rodriguez v. Taylor, 569 F.2d 1231, 1248 (3d Cir. 1977), cert. denied, 436 U.S. 913, 98 S. Ct. 2254, 56 L. Ed. 414 (1978); Torres v. Sachs, 538 F.2d 10, 13 (2d Cir. 1976); Fairley v. Patterson, 493 F.2d 598, 606-07 (5th Cir. 1974), disapproved on other grounds, Alyeska Pipeline Service Co. v. Wilderness Society, 421 U.S. 240, 270 n. 46, 95 S. Ct. 1612, 1628, 44 L. Ed. 2d 141 (1975); Meisel v. Kremens, 80 F.R.D. 419, 422-23 (E.D.Pa.1978).36 Indeed, the Supreme Court very recently commented:

We also reject petitioners' argument, not suggested in the petition for certiorari, that respondent's representation by a public interest group is a "special circumstance" that should result in denial of counsel fees. Federal Courts of Appeals' decisions are to the contrary. See, e. g., Reynolds v. Coomey, 567 F.2d 1166 (1st Cir. 1978); Torres v. Sachs, 538 F.2d 10, 13 (2d Cir. 1976). Congress endorsed such decisions allowing fees to public interest groups when it was considering, and passed, the Civil Rights Attorney's Fees Awards Act of 1976, 90 Stat. 2641, 42 U.S.C. § 1988, which is legislation similar in purpose and design to Title VII's fee provision. See H.R.Rep. No. 94-1558, pp. 5 and 8, n.16 (1976).

New York Gaslight Club, Inc. v. Carey, 447 U.S. 54, 70 n.9, 100 S. Ct. 2024, 2034, 64 L. Ed. 2d 723 (1980). This passage is significant not only for its discussion of the public interest lawyer's entitlement vel non to a fee; the approving reference to Reynolds and Torres, both cited supra, is highly relevant also to the calculation of the fee. The Second Circuit opinion in Torres, for example, stated:

Litigation to secure the law's protection has frequently depended on the exertions of organizations dedicated to the enforcement of the Civil Rights Acts. We agree with the Courts which have held that the "allowable fees and expenses may not be reduced because (the prevailing party's) attorney was employed ... by a civil rights organization or because the attorney does not exact a fee."

538 F.2d at 13 (ellipsis and brackets in original) (emphasis added) (citations omitted).

Nor is it relevant that a law firm, as in this case, originally undertook representation pro bono publico. We see "nothing inconsistent in prosecuting a case in the public interest, agreeing not to charge one's own client a fee and thereafter seeking fees" from the losing defendant. Keyes v. School Dist. No. 1, 439 F. Supp. 393, 406-07 (D.Colo.1977) (emphasis in original); see National Treasury Employees Union v. Nixon, 521 F.2d 317, 322-23 (D.C. Cir. 1975), quoting Wilderness Society v. Morton, 495 F.2d 1026, 1037 (D.C. Cir. 1974) (en banc). Similarly, the fee calculus does not change simply because the law firm representing plaintiff in this case, see note 1 supra, may choose to donate its fee to a "public interest" law entity. See, e. g., Tillman v. Wheaton-Haven Recreation Ass'n, 517 F.2d 1141, 1148 (4th Cir. 1975).

For all these reasons, we decline to adopt the panel's "cost-plus" method of calculating fees.

The preceding explication of the proper criteria for awarding an attorney's fee permits us now to consider the District Court's award in this case.

It is common learning that an attorney's fee award by the District Court will be upset on appeal only if it represents an abuse of discretion.37 We customarily defer to the District Court's judgment because an appellate court is not well situated to assess the course of litigation and the quality of counsel. The District Court judge, by contrast, closely monitors the litigation on a day-to-day basis. The Supreme Court long ago observed that a trial judge "has far better means of knowing what is just and reasonable than an appellate court can have." Trustees v. Greenough, 105 U.S. 527, 537, 26 L. Ed. 1157 (1882). Accordingly, we think "it is better to have th(e) discretion (to award fees) exercised by the court which has been most intimately connected with case."38

In this case, for example, the District Court Judge was intimately familiar with the barrage of pleadings, memoranda, and documents filed, and he observed the proficiency of counsel in court. Our inspection of the cold record cannot substitute for his first-hand scrutiny. Under these circumstances, we are most hesitant to upset the product of his judgment.39

It is readily apparent that the District Court's fee-setting calculations do not precisely conform to the procedures identified in earlier cases and elaborated upon in this opinion.40 We do not believe, however, that it would be productive now to remand this case for new computations. It was more than seven years ago that Dolores Copeland alleged the gender discrimination that resulted in this lawsuit.41 This very opinion, regretably, for clarity must be identified as Copeland III. However, the end of this protracted litigation finally is in sight, and we do not wish unnecessarily to prolong it when the only remaining issue is that of the attorney's fee. The Court of Appeals for the Third Circuit, en banc, commented in a similar fee-award situation:

(I)n the interest of terminating the lengthy proceedings at bar-now in their fifth year-we will not require the district court here to reconsider its determination. Although we do not disturb the district court's treatment ..., it should be apparent that we do not necessarily endorse the methods or the reasoning employed to reach its result.

Lindy II, 540 F.2d at 118.

Moreover, the time spent litigating the fee award normally is itself compensable, see note 29 supra, so it would be anomalous to require plaintiff's lawyers to spend additional time unless the remand were likely to be productive.42 For the reasons set forth below, we do not think that a remand in this case now would be useful. The fee award appears to us to be a reasonable one. Other circuit courts of appeals have observed that appellate judges are themselves experts in assessing the reasonableness of an attorney's fee award,43 and that the appellate court, in a pinch, may independently review the record,44 or itself set the fee.45

Our own examination of the fee award in this case leads us to believe that the award is fully justifiable and that affirmance of the award, under the special circumstances of this case, therefore would not be inappropriate.46 Accordingly, we affirm.

1. Calculation of the "Lodestar"

Plaintiff's attorneys submitted the following materials documenting their fee request:

(1) a memorandum reviewing the history of the litigation;

(2) affidavits revealing that they spent 3,602 hours on the case;

(3) statistics revealing that the firm's normal hourly rates for these attorneys ranged from about $52 per hour to about $90 per hour;

(4) a letter from the Lawyers' Committee for Civil Rights Under Law stating that the typical fee charged by large Washington firms on employment discrimination cases ranged from $35 to $100 per hour; and

(5) an itemization of the back pay awards obtained by the lawsuit.

The average hourly rate, weighted for the number of hours spent by each attorney, was $57.17. A "lodestar" fee, computed by multiplying $57.17 by the 3,602 hours spent, would be $205,916.50.

The District Court properly inquired whether that hourly rate was reasonable, and whether all the hours were reasonably expended. It found that, although the rate was indeed reasonable, some of the hours were nonproductive. The District Court noted that "the firm has obviously made little, if any, effort to exercise billing judgment.... (T)here was practically no partner time expended on this case and the associates lacked experienced trial direction."

With this analysis we can find no fault. Our review of myriad attorney's fee cases demonstrates that the $57 average hourly rate certainly is within the bounds of reasonableness.47 We think, also, that the District Court properly refused compensation for nonproductive hours. Hours may be disallowed as nonproductive for at least two reasons. First, no compensation should be given for hours spent litigating issues upon which plaintiff did not ultimately prevail.48 Also not allowable are hours that simply should not have been spent at all,49 such as where attorneys' efforts are unorganized or duplicative. This may occur, for example, when young associates' labors are inadequately organized by supervising partners.50

It is true that the District Court in this case did not identify precisely the hours, or types of work, for which no compensation could be paid. Compare section II-A-1 of this opinion supra. Instead, it simply decreased the "lodestar" by a substantial fixed amount from the $206,000 proposed starting point. Under the special circumstances of this case, this practice cannot be condemned. The reduction in fee resulted primarily from the expenditure of unnecessary time by relatively inexperienced lawyers. It is neither practical nor desirable to expect the trial court judge to have reviewed each paper in this massive case file to decide, for example, whether a particular motion could have been done in 9.6 hours instead of 14.3 hours. In Lindy II, the en banc Court of Appeals for the Third Circuit observed that an appellate court does not

intend that a district court, in setting an attorneys' fee, become enmeshed in a meticulous analysis of every detailed facet of the professional representation. It ... is not our intention that the inquiry into the adequacy of the fee assume massive proportions, perhaps even dwarfing the case in chief. Once the district court determines the reasonable hourly rates to be applied, for example, it need not conduct a minute evaluation of each phase or category of counsel's work.

540 F.2d at 116 (emphasis added).

This approach is particularly appropriate in a case such as this. A pleading-by-pleading examination of the copious files in this case would be unnecessarily burdensome. We think that the District Court Judge in this case-recognizing, as he did, that some duplication or waste of effort had occurred-did not err in simply reducing the proposed "lodestar" fee by a reasonable amount without performing an item-by-item accounting.51

The government in the District Court characterized the number of hours as "litigious overkill" by "relatively junior attorneys." Such vast numbers of hours were wholly unnecessary, according to the government, because the lawsuit was a "mutually cooperative effort" to ferret out and remedy discrimination in the Labor Department. The government points out, for example, that it stipulated that the Labor Department had discriminated, settled virtually all of the individual claims for relief, and developed an affirmative action program. (The government's account of the litigation was proffered, of course, after the heat of battle had cooled.)

The District Court, however, closely monitored the course of this litigation, and therefore is more aware than anyone of the manner in which it was conducted. The court commented:The litigation went forward in a relatively civilized manner but it was hard fought. The Government offered firm, persistent resistance throughout the litigation and concessions developed only as it became apparent there was little prospect of Government success. Indeed, the Government moved to dismiss at the outset, and it opposed discovery. There were many difficulties encountered during the discovery process which were caused, in part, by the Department's inadvertent destruction of certain records contrary to Court direction and the intentional withholding of other documents by some officials of the Department of Labor, as well as by the complexity of the issues.

Our review of the record demonstrates that the District Court's description of the litigation is accurate. The tenor of the case can be sensed simply by examining the District Court docket. Its entries abstract the history of the lawsuit, which was chronicled in detail in section 1 of this opinion, and reflect the truly adversarial attitude with which it was waged. Some of the docket entries, couched in the cryptic argot of that genre, are:

MOTION of defts. for judgment on the pleadings; P & A's; ....

MOTION of defts. to remand; P & A's; ....

OPPOSITION of defts. to motion for certification ....

MEMORANDUM by pltffs relating to defts destruction (of) documents ....

POST-HEARING brief by defts in opposition to pltffs claim for retroactive promotion and back pay; ....

and the like. Under these circumstances, the government's characterization of this litigation as a "mutually cooperative effort" is startling.52 It is true, as we noted above, that the government conceded at the administrative level that the Directorate had discriminated against women "in the lack of leadership responsibility assignments given to qualified women professionals." That was no major concession for purposes of this lawsuit, however, because the concession did not encompass discrimination in treatment of nonprofessional women, or discrimination in promotions and training of professional women. The lawsuit proved to be necessary to resolve those issues.

We do not, of course, criticize the defendant's attorneys for skillful and thorough representation of their client. The government's defense of this suit may well have been a model of effective advocacy. That, however, is not the point. The government's contentious litigation strategy forced the plaintiff to respond in kind. The government cannot litigate tenaciously and then be heard to complain about the time necessarily spent by the plaintiff in response.53

Nor does it help the government that it eventually conceded, or settled, many of the important issues before trial. That judgment is by stipulation, concession, or consent decree normally is irrelevant to the entitlement vel non to an attorney's fee,54 and in this case it also is irrelevant to the calculation of the fee. Where "developments made it apparent that the judge was about to rule for the complainant," a defendant cannot hope to ameliorate the burden of the attorney's fee by making eleventh-hour concessions. Cuneo v. Rumsfeld, 553 F.2d 1360, 1362, 1364 (D.C.Cir.1977). Because the starting point in setting an attorney's fee is the "lodestar"-hours reasonably spent multiplied by a reasonably hourly rate-a last-ditch concession does not prevent compensation for hours spent litigating before the concession was offered.

In its appellate brief, the government argues that its version of the litigation would have been portrayed more fully if the District Court had conducted a hearing on the issue of the fee. Such a hearing in some circumstances may be useful.55 In this case, however, the District Court ruled on the fee question after witnessing the conduct of the entire case, and with the benefit of substantial briefs from both sides.56 We cannot say that failure to hold a hearing under these circumstances was error.57 Moreover, the government never took the elementary step of asking the District Court to hold a hearing.58 We think that "having sought no hearing on (the attorney's fee) motion, counsel could hardly have been surprised when none was held." Kargman v. Sullivan, 589 F.2d 63, 67 (1st Cir. 1978).

The government next contends that the fee request was inadequately documented. Specifically, the government asserts that the District Court erred in awarding the fee because it had before it affidavits that only roughly itemized the hours spent. The District Court should have consulted the original time logs kept by the firm, we are told.

It may, on occasion, be necessary for the District Court to examine the original time logs. See Pete v. UMW Welfare & Retirement Fund, 517 F.2d 1275, 1292 (D.C.Cir.1975) (en banc ). In this case, however, there is no allegation that the affidavits summarizing the time logs are in any respect inaccurate.59 Moreover, although plaintiff's attorneys apparently were willing to produce the time logs,60 the government never requested that they do so. Nor did the government ever object in the District Court that the itemization of the hours was in any respect deficient. We think that objections to documentation, like requests for evidentiary hearings, must be made to the District Court; one cannot, after receiving an unfavorable ruling from the trial judge, concoct objections for use on appeal.

2. Adjustments to the "Lodestar"

A "lodestar" fee may be adjusted, as we have seen, to compensate for the possible contingent nature of success, and in recognition of the quality of representation.

a. Contingency

The "lodestar" fee may be increased to reflect the possibility that the litigation would not be successful and that no fee ultimately would be obtained. The District Court in this case made no contingency adjustment, nor was it requested to do so.61

b. Quality

None of the factors that we must consider under this rubric suggests that the District Court's fee award need be disturbed.

The District Court recognized that plaintiff's attorneys were inexperienced. It found, though, that they were "excellent"; what they "lacked in seasoned trial experience was offset by other factors. They were always well prepared, effective and knowledgeable." Our examination of the record in this case confirms the appraisal of the trial judge.62

Another of the important elements in assessing the quality of counsel is the benefit of counsel's efforts: the success of the litigation. The government points out that the fee award was vastly disproportionate to the monetary relief obtained, and contends that the fee award should be reduced accordingly.

The District Court, however, observed that where, as here, the primary purpose of a lawsuit is to obtain equitable relief, the fee award should not necessarily be reduced simply because monetary relief was small. Moreover, the court found that the nonmonetary relief achieved in this case-an affirmative action program for training and promoting women-was important, and "will be felt for many years to come."

We think that the District Court has accurately summarized both the benefits produced by this lawsuit and the law governing the award of fees in cases that produce primarily nonmonetary relief.

Title VII provides various sorts of remedies once discrimination has been found. Relief can be in form of back pay,63 promotion,64 and other prospective equitable relief such as affirmative action programs.65 This case involved each category of relief. Some $33,000 in back pay for 13 of the plaintiffs was paid. That amount is not insubstantial, but it does not represent the principal relief obtained. Some class members received promotions to the GS-14 level. Others received specialized training and guaranteed promotions upon completion of that training. The prospective value of this kind of relief far exceeds the back pay awarded.

Probably the most important relief obtained by this litigation was the affirmative action plan. The plan, which was broadly aimed at eliminating the "blatant discrimination" the District Court found to have existed in the Directorate, offers substantial prospective relief for the class. The plan provides significant training opportunities for women, including accelerated training for management positions. It requires that work assignments be made on a nondiscriminatory basis. It also establishes goals, timetables, and an annual reporting system to measure the Directorate's progress in remedying the effects of past discrimination.66

Plaintiff's efforts also produced intangible benefits that flow from the elimination of discrimination. Since women no longer should be denied advancement because of their gender, the sense of inferiority resulting from the discrimination will be eliminated.67 Moreover, the exposure of discrimination in one portion of the Department of Labor may lead to heightened awareness of, and intensified efforts to eradicate, possibly discriminatory treatment throughout the Department.

The government insists, nevertheless, that the absence of a substantial monetary recovery ought to preclude a fee award on the order of magnitude of that given by the District Court. The panel agreed, concluding that the existence of a $160,000 fee to obtain about $33,000 in back pay establishes "at least a prima facie case that something is wrong with the previously constructed standards." Copeland II, slip op. at 20. (emphasis in original).

We disagree. That this litigation sought and obtained substantial equitable relief is highly relevant to the award of a large fee. We have noted that the primary purpose of the fee provision is to give persons victimized by discrimination the resources to vindicate their rights through litigation. Where, as here, the relief sought is generally nonmonetary, a substantial fee is particularly important if that statutory purpose is to be fulfilled. It is relatively easy to obtain competent counsel when the litigation is likely to produce a substantial monetary award. It is more difficult to attract counsel where the relief sought is primarily nonmonetary. For this reason, fee awards in cases that produce substantial nonmonetary benefits must not be reduced simply because the litigation produced little cash.68 The Court of Appeals for the Second Circuit approved payment of a substantial fee where, as here,

(t)he plaintiff sought and obtained broad equitable relief which will benefit (persons) in addition to the plaintiff. Although the size of the group affected by this litigation may not be large, the claim involved significant federal rights. Because the plaintiff's claim did not involve substantial monetary damages ... it is highly unlikely that the plaintiff could have retained other counsel to act on her behalf. The cost of this litigation and the time and effort required by plaintiff's counsel in order to bring the case to its conclusion ... plainly were disproportionate to any sum the plaintiff might reasonably have expected to recover on her own behalf.

Holley v. Lavine, 605 F.2d 638, 646 (2d Cir. 1979), quoting and aff'g 464 F. Supp. 718, 725 (W.D.N.Y.1979).

For these reasons, no reduction in fee was appropriate after considering the relevant factors in the category of "quality of representation."

The District Court considered all the relevant factors in awarding an attorney's fee. Although its discount of the proposed "lodestar" amount could perhaps have been computed with greater specificity, that discount was quite substantial however it may have been measured. It was an exercise of judgment informed by both experience and direct observation of what had transpired in the course of the litigation.

We therefore find it unnecessary now to remand the matter for further inquiry, and affirm the District Court's attorney's fee award.

It is so ordered.

MacKINNON, Circuit Judge.

I join in the court's opinion. In my view both the market value and the cost-plus standard can at times lead to exorbitant fees; but I view the market value approach as being more time honored and hence easier of application nationwide where all lawyers do not keep detailed records of overhead costs and other relevant expenditures. The bench and the bar are accustomed to market value in the area as representing reasonable attorney's fees and I believe that standard, when properly applied, includes within it its own elements of reasonableness. The only admonition I would include would be the caveat that the attorney's fee in the ordinary case must bear some reasonable relationship to the amount of money and equitable benefits that are involved. Attorneys in such cases should be forced to be mindful of the monetary and equitable benefits that are being sought and should not be permitted to run up bills that are greatly disproportionate to the ultimate benefits that may be reasonably attainable.

Circuit Judge ROBB concurs in the foregoing opinion.

WILKEY, Circuit Judge, joined by TAMM, Circuit Judge, dissenting:

In our colleagues' "lodestar" opinion, the path of attorney's fees in Title VII litigation is easy to discern. It is Up, Up, and Away! It is Per Calculos Ad Astra.1

Before going to the extraordinary calculos of the majority opinion, it would be well to point out the areas in which we have no disagreement and what precisely we think the issue is. We have no quarrel, of course, with the findings of the trial court, reiterated in exquisite detail by the majority opinion, regarding the labors of the young attorneys who worked for the plaintiff on this case. We have no quarrel with the regular hourly rates charged by this distinguished law firm to its regular corporate clients for the labors of its senior partners, its junior associates, or these particular junior associates. The issue is not whether they shall be paid; as attorneys for the prevailing party, they are rightly entitled to be paid. The issue is not, on this appellate level, even how much they shall be paid. The issue here is only-but unfortunately not simply-the formula by which these attorney's fees should be calculated by the trial court.

This question has required much time and effort on the part of this court to attempt to resolve. All members should have come to grips with the obvious fact, as we analyzed and dissected this particular case, that previous precedents in the field, even in our own circuit, rested on contradictory, overlapping, disharmonious, even spurious and irrelevant "factors." This incoherent melange provides no consistent rationale for a conscientious trial judge to apply, at least in a case in which the Government is the paying defendant; the effort to do so here has highlighted the inadequacies and inapplicability of previous precedents.

Judge McGowan's opinion for the majority is the most strenuous effort so far to pull a basically incoherent rationale (even if of some service in the private sector) together, to smooth putty into the visibly widening cracks of logic and to cover with the lacquer of a polished style. We in the dissent cannot buy the product. Over many months of pondering the arguments for importing the market value hourly rate fee into government Title VII litigation-this is the first case in this circuit-we are more than ever convinced that the market value concept is unworkable where there is no true market, that the effort to apply it inevitably leads to distortions and excesses, far away from the "reasonable attorney's fee" which was Congress's only avowed standard and intent.

First we delineate the differences which, on thoughtful analysis, are necessarily implicit in the setting of a "reasonable attorney's fee" according to the statute, where the Government as contrasted with a private party is the paying defendant and no true market exists. Next we take up the obvious flaws in the applicability of the majority's private sector theory to this case of the Government as defendant-the redundant contingency factor, the care and feeding of lawyers rather than injured plaintiffs, and the evasion of testing the majority formula in this case. Finally we describe the actual cost plus a reasonable and controllable profit method of determining a "reasonable attorney's fee."

                   Outline of Opinion
                                                     Page
  I.  Differences in the Role of the Government
      and Private Parties as Defendants in Title
      VII Litigation ............................... 909
      A. Past Precedents in This Circuit ........... 909
      B. Result When Past Precedents Applied
          to Copeland .............................. 910
      C. How the Attorney's Fee Incentive
          Operates in Government Litigation ........ 911
      D. Failure of Attorney's Fees to Deter
          Discrimination by the Government ......... 912
      E. "Reasonable" Private and Government
          Attorney's Fees in Government
          Litigation ............................... 912
      F. The Majority's "Market Value" Fee
          Where No Market Exists ................... 913
      G. The Statute and the Legislative
          History .................................. 915
 II.  The Majority Opinion Formula--Faulty
      Analysis Produces Skewed Results ............. 917
      A. The Redundant Contingency Factor .......... 917
      B. Encouraging Injured Plaintiffs or
          Encouraging Lawyers? ..................... 920
      C. Quality of Representation ................. 921
      D. Majority Failure to Apply New Formula
          to this Case ............................. 922
III.  "A Reasonable Attorney's Fee"--Actual Cost
      Plus a Reasonable and Controllable Profit .... 923
      A. The Need for Additional Guidelines for
          the District Court ....................... 923
      B. Rationale ................................. 924
      C. Application ............................... 925
      D. Substantive Inequities Feared from the
          New Method ............................... 926
          1. Special Problems with Small Firms
              and Solo Practitioners
          2. Similar Problems for the Public
              Interest and Civil Rights Bar
          3. Contingent Fees
          4. Class Action Practitioners
          5. Counsel Representing Poor Clients
              Generally
          6. Counsel Who Work Long Hours
          7. Requirement of Public Disclosure of
              Financial Information
      E. Fear of Deterring Representation in
          Title VII Cases .......................... 928

I. DIFFERENCES IN THE ROLE OF THE GOVERNMENT AND PRIVATE

The proper method for determination of attorney's fees awards is a recent and developing area of the law, in which courts are still grasping for exact standards consistent with congressional intent. Over the past half dozen years this court has addressed the attorney's fees issue in several cases, each successive case attempting to systematize and sometimes modify the previous precedents.

In Evans v. Sheraton Park Hotel2 we listed twelve factors that should be included in a district court determination of appropriate attorney's fees in Title VII cases. The twelve factors were taken from the Fifth Circuit's decision in Johnson v. Georgia Highway Express, Inc.3 which had in turn drawn the standards from guidelines recommended by the American Bar Association's Code of Professional Responsibility, Ethical Consideration 2-18 and Disciplinary Rule 2-106.

In Kiser v. Huge4 we endorsed these same standards taken from the Code of Professional Responsibility as well as the Manual for Complex Litigation, section 1.47.5 And in Pete v. United Mine Workers6 we incorporated the attorney's fees section of the panel opinion in Kiser.

Our opinion in National Treasury Employees Union v. Nixon, summarizing the analysis of Pete and Huge, listed the primary factors for consideration: hourly rate multiplied by hours, adjusted upward if there is a risk of noncompensation or partial compensation, and adjusted upward or downward on the basis of the quality of work performed as judged by the district court.7

B. Result When Past Precedents Applied to Copeland

In any developing area of the law such as this, courts setting forth general principles must attempt to design them to produce just results in various fact situations to which they will logically apply in the future. The specific situations which arise in future cases will sometimes bring to light deficiencies in the general rules laid down in the past, and will thus require greater elaboration or modification of the standards. The present case illustrates how a mechanical application of the hourly rate times hours formula, which appeared advisable as a general rule in past cases, can actually lead to unreasonable fees entirely out of line with what would happen in private litigation.8

The litigation effort in this case concerned a claim of sex discrimination brought by approximately twenty-four females employed in branches of the Directorate of Data Automation in the Department of Labor. The extensive discovery and numerous pretrial motions in the case were directed at the issue whether the Department of Labor had discriminated in the past against the plaintiff class in assignments, training, performance evaluations, promotions, and working conditions. In compensation for the effects of discrimination found in this case, the Department of Labor paid a sum total of $31,345 in back pay for thirteen of the plaintiffs. In contrast, the legal fee requested by plaintiffs' attorneys was $206,000, and the amount awarded by the district court was $160,000. Plaintiffs' attorneys spent 3,602 hours on the case.

Whether this gross disproportion between the monetary stake of plaintiffs' claim and the cost of litigating that claim is justified by the amount of equitable relief awarded, as the majority apparently believes, does not resolve the problem presented by these facts. It is true that some of the plaintiffs received promotions and prospective future promotions, and that the Department agreed to adopt an affirmative action plan for female employees in positions requiring knowledge of data processing. But the fact remains that this Title VII suit involved only twenty-four class members in a very limited sector of a government agency, and concerned acts of discrimination whose sum total monetary value over a several year period was $31,345; yet plaintiffs managed to throw such resources into the legal battle that they could claim a legal fee of $206,000 and receive from the district court a fee of $160,000. Even if the equitable relief here was worth five times the monetary award, the total amount of relief would never make a $160,000 fee appear reasonable in private litigation. When fees of this magnitude begin to appear in suits against the Government, it is time to ask some serious questions about the reasonableness of the attorney's fees judges are awarding.

This case illustrates the potential result when a large private firm, with high rates customarily charged to wealthy corporate clients, brings a suit against a Government defendant with an unlimited deep pocket, and then proceeds to engage in extensive discovery and numerous pretrial motions, while being assured from the outset that all hours spent on the case will be reimbursed at the firm's customary rate so long as its efforts are relevant to issues on which it ultimately prevails. It is of little use to quibble over whether the amount of equitable relief involved in this particular case can possibly justify the high fee, because the situation in this case foreshadows and points the way toward far greater potential abuses in the future, if the hourly rate times hours formula continues to be mechanically applied.

Though past cases in this area, including those on which we ourselves have sat, have not given specific attention to these potential abuses, it is clearly time to do so now. The reason perhaps why those earlier cases did not consider the great potential for abuse is that they did not involve the Government in Title VII litigation and did not anticipate the possible effect when the resources of a large private law firm are brought to bear against the Government in an employment discrimination suit. We have that case for the first time now. It illustrates how these suits can diverge drastically from the commonly known patterns of private litigation.

C. How the Attorney's Fee Incentive Operates in Government Litigation

When a large firm knows that eventual success will bring it compensation at its customary rate for all relevant hours of work, the firm has a tremendous incentive to expand the pretrial stages of the case to the point where it becomes overwhelmingly in the Government's interest to settle, whether the Government is in the wrong or not. In private litigation the incentive to expand the discovery and pretrial motion stages is counterbalanced by the high cost that this will inflict on the client, because victory does not normally bring a recovery of the litigant's own attorney's fees from the other side. Not so in a case of this sort against the Government, where the law deliberately encourages the litigation by holding out the carrot of attorney's fees-but only to successful plaintiffs.

Furthermore, in private litigation the high cost of extensive discovery serves as an incentive for both sides to settle. But in Title VII cases against the Government, the incentives become entirely lopsided, because expanded litigation costs for plaintiff not only increase his chance of winning, but also greatly increase the sum his lawyer stands to gain if he does win. At the same time, each expansion of the litigation effort will pose a risk of higher and higher liability for the Government. The end result is that the Government faces overwhelming incentives to give in to claims, however unjust, before expanded litigation doubles or triples or quadruples the size of eventual Government liability. When attorney's fee levels come to dwarf the actual monetary amount in controversy, as in the present case, the structure of these incentives is magnified further. In deciding whether to settle a case like this one, the Government is not primarily considering whether its case is strong enough that it should risk an eventual $31,345 judgment to plaintiffs; instead, if Government attorneys are rational they must primarily consider whether their case is strong enough to risk the much greater possibility of a $200,000 plus eventual attorney's fee award.

D. Failure of Attorney's Fees to Deter Discrimination by the Government

Attorney's fees are meant to serve some purpose of deterring discrimination. They doubtless do in the private sector.9 But when attorney's fees come straight out of the United States Treasury, as in the present case, they exert no deterrent effect whatsoever against the persons responsible for the discrimination. In the private sector there is a justifiable punitive element. Attorney's fees impact on the profit picture of the corporation; the same executive management which is responsible for tolerating or encouraging discrimination are the same executives who are responsible for the profit of the corporation, so they are penalized in the pocketbook. No such deterrence applies to the Government, i. e., the Labor Department budget was never touched, will never be touched, by the award of Judge Gesell in this case. Both the back pay and the attorney's fee come out of the general taxpayer contributed funds of the U. S. Treasury. By their strict analogy to the private sector, the majority has validated deterrent or punitive action against the U. S. taxpayer.

E. "Reasonable" Private and Government Attorney's Fees in Government Litigation

The Government operates in a universe of attorney's fees lower than that prevailing in the world of wealthy corporate clients. If the Government were pressing the Title VII case rather than defending, its legal costs would not approximate private "market value." Government legal counsel, whether on the plaintiff or defendant side, traditionally simply have never been compensated at the same scale as in private practice. The young associate makes slightly more than the young Government attorney; the margin becomes really vast when partner status in a prestigious law firm is compared even with the Attorney General of the United States. Since the Government willy-nilly is financing these Title VII suits both by plaintiffs who succeed and do not succeed (giving effect to a contingency factor), Government counsel defending and plaintiffs' counsel bringing the suits should have compensation of roughly the same amplitude.

There is a logical symmetry in this principle. When the Government as a defendant prevails, it has asked for and received only its actual costs, the salaries of the attorneys working on the case plus overhead costs attributable to their work-not the market value going hourly rate of the private sector.10 It is undeniably the same litigation, vindicating the rights of employees discriminated against, whether the Government wins or loses. The work on the facts and the legal issues is of the same complexity on both sides. Indeed, if the Government were not the defendant employer, the private attorneys would not be bringing a pro bono publico lawsuit as private attorneys general. What the private attorneys are doing is essentially Government legal work. Recognizing this, why should the private attorneys be compensated on a scale other than actual costs (salary and overhead) plus a reasonable profit to encourage them to continue accepting employment in this type of litigation? The purpose of Title VII attorney's fees would be fully vindicated by such a policy.

F. The Majority's "Market Value" Fee Where No Market Exists

With reference to attorney's fees in Title VII litigation against the Government, the fundamental verities overlooked by our colleagues are these: first, there is no market in which fees are set by market value forces; second, whatever fee-paying mechanism there is exists as the total creation of the Government itself.

To the latter point, we must remember that the Government need not consent to be sued by its own employees or anyone else. No lawsuit, no lawyer's fee. In almost any other country in the world government employee complaints are adjudicated through the administrative process, as indeed was thought perfectly proper in the United States until 1972, when Congress decided it was more just to give government employees the same rights in court as private employees. Our role is not to choose between administrative or judicial relief; it is to give effect to that selected by Congress. And so we point out that if the majority's pathway to the stars formula for attorney's fees is applied often enough, there may be a popular demand to take government employee Title VII litigation out of the court system altogether and limit employees to administrative relief as previously.

As matters stand now, the Government as defendant pays the cost of its own lawyers in court, the Government pays attorney's fees for those plaintiffs who prevail, and the Government-because it pays not only the true cost of the specific suit but also a contingent factor for attorney's costs in other suits in which the plaintiff does not prevail-pays attorney's fees indirectly for plaintiff lawyers who lose. The Government-the American taxpayer-is financing BOTH sides of ALL Title VII litigation against the Government. This is the chosen mechanism; the court's role is to determine how the fees are to be calculated.

So we now return to the first point above: the Government is financing this litigation under circumstances in which there are no market value forces to restrict or to set "reasonable attorney's fees" ; fundamentally, because there is no market. We all recognize that the government employees who become plaintiffs in Title VII cases do not have the financial resources to bid in the marketplace for attorney's services. Copeland and fellow plaintiffs could never have contemplated this litigation, resulting in a suggested fee of $206,000, and an award of $160,000 plus $12,000 expenses. To vindicate the rights of persons discriminated against, Congress has realistically appraised the situation and recognized that attorney's fees must be taxed to the losing defendant, if there is to be any Title VII litigation brought by the injured persons. We in dissent obviously have no quarrel with this whatsoever; we realistically recognize, however, as our colleagues do not, that here there is no market, and that necessarily there must be a different method of setting a "reasonable attorney's fee" if there are no market forces to determine a "market value" fee.

Whatever the usual merit of "market value," it is anomalous where Congress has specifically sought to complement the market for legal services in Title VII cases with an alternative mechanism for allocating attorney's fees. Rather than relying on the parties in privity-the client and his or her lawyer-to agree among themselves to the value of the services to be rendered, Congress has provided that a trial court, in its discretion, may assign the burden of legal expenses incurred by the winner to the loser.

The emphasized points in this last sentence should underscore the substantial incongruity between market and judicial determinations of the cost of legal services. Furthermore, the market works so well because it works invisibly, almost effortlessly; even the most astute of judges can only mock the market concept when the judge seeks to replicate market results where none exist. Emulating the market, effective as it is, is worthless to the extent that market factors-supply and demand, agreement among market-makers, prospective not retrospective assessments of marginal utility-are more conjectural then real.

So it is plain, then, that attorney's fees in the Title VII context cannot be assigned according to market values because there are quite simply no buyers for the services rendered. Without clients who agree to pay at certain (bargained for) rates, there can be no market. There is missing an a priori pecuniary relation between the legal services undertaken and the willingness of any beneficiary to pay for them. Therefore, charging a losing defendant for the prevailing party's legal expenses at "market" rates which no one would ever have voluntarily assumed is to destroy the market concept by purporting to respect it.

Another inconsistency of the majority's "market value" approach when compared with bona fide markets is with regard to risk. In a market, i. e., a real market, risk is a negative factor which diminishes the "expected value" of the beneficiary's lawsuit to the client. A rational client, the beneficiary of legal services, would pay less to pursue relatively riskier litigation because his anticipated recovery must be discounted for that risk. Naturally, lawyers who will be paid only upon the contingency of success would appreciate and demand a risk premium for their services. Only a true market, of course, could allocate the burden of risk accurately between lawyer and client. This the private sector does every day. In Title VII litigation against the Government we have placed this burden on the judge.

This is indeed a situation which, if recognized in its true proportions, is so completely different from private litigation that it is obvious that the same concepts of market value fees cannot be employed. Instead of the free play of the market of private clients we have the Government consenting to be sued, financing all successful litigation against itself, and also financing all unsuccessful litigation against itself by enlarged fees on a contingent basis. If this is what is to be done, and it is what is to be done by both the majority's concept and ours, then the Government is entitled to put down some rules and regulations on the cost of the legal services, which by the system designed by statute the Government is going to finance in totality. If the Government is going to finance the whole mechanism in totality, then the concept of actual cost plus a reasonable profit to the attorneys is about the only concept that can hold attorney's fees down to a reasonable level.

The tendency toward abusively high fees in this sort of case is real, yet the majority opinion offers absolutely nothing to cope with this problem. We believe there are indeed ways by which we can try to solve this problem, rather than throw up our hands and permit attorney's fees in Title VII cases to continue up to the stars. Two of these factors can be found listed in the Evans case: (1) limiting fees to a level commensurate with awards in similar cases, and (2) limiting fees to a level commensurate with the amount in controversy and results obtained.11 Significantly, the majority opinion makes no mention of these limiting factors from our precedents; apparently the majority actually prefers to encourage legal fees upward toward the stars. We submit that "lodestar" fees were far from the Congressional gaze when Congress specified "a reasonable attorney's fee,"12 so we now turn briefly to what is available in the legislative history.

The majority asserts that the "cost-plus" formula is "fundamentally inconsistent with Congress's purpose in providing for statutory fee-shifting."13 The majority further contends that Congress intended that a fee should be based on the market value of the services rendered.14 Neither the language of the statute nor the legislative history underlying the attorney's fee provision evinces a congressional mandate to use the market value approach in calculating a reasonable attorney's fee award.

The purpose underlying the attorney's fee provision is to encourage deserving litigants to seek judicial relief. To effectuate this purpose, Congress intended that the attorney's fee awards be sufficient to attract competent counsel, but not so unreasonably high as to produce a windfall for the attorney. It is our position, quite simply, that the market value approach is inappropriate at least in cases in which the Government is the losing defendant because, as will be demonstrated in our analysis of the practical flaws in the majority theory in Part II below, it presents the likelihood that the attorney indeed will reap a bonanza. The "cost-plus" formula, on the other hand, will provide adequate compensation to enable litigants to obtain competent counsel without providing a windfall to the attorney.

We begin by looking to the language of the statute. The attorney's fee section provides that "the court, in its discretion, may allow the prevailing party ... a reasonable attorney's fee as part of the costs, and the ... United States shall be liable for costs the same as the private person."15 The language of the statute surely does not mandate a market value approach; it specifies not at all the method of computing the fee, but instead only generally directs that the attorney's fee should be reasonable.

The majority asserts, however, that the language "the United States shall be liable for costs the same as a private person" plainly indicates that the method of calculation of the attorney's fee should not vary with the identity of the losing defendant. To prove that this is wrong we need not advocate a clear dichotomy between the method of calculating fees in the case in which the losing defendant is a private company and in the case in which the losing defendant is the Government. The case before us is one in which the losing defendant is the Government, and we think that it well illustrates how the market value approach leads to an unreasonable result in this type of case. It may be, however, that the market value approach would not be inappropriate in cases in which a private concern is the losing defendant, another case which we do not have before us.

Apart from that caveat, our response to the majority's contention is that we take the language of the statute to mean only that the United States shall also bear the burden of "a reasonable attorney's fee" when it is the losing defendant. The statutory language does not indicate that the identical method of calculation shall be used in computing this fee-what may be a reasonable method of calculating an attorney's fee award in the situation in which a private entity is the defendant may be a totally unreasonable method when the Government is the losing defendant.

We also note that the attorney's fee provision quoted above has been a part of Title VII since the enactment of the statute in 1964. At that time, as the majority indicates in a footnote, Title VII did not permit employment discrimination suits against the United States. Thus attorney's fees were awarded against the Government only when the Government was a losing plaintiff. It was not until 1972 that a new provision was added to Title VII to allow suits against the Government. The language of the attorney's fee section was not changed; rather, the new section providing for suits against the federal government indicated that the attorney's fee provision would also be applicable to claims made by federal employees.16 A fair reading of this sequence of events is that Congress did not contemplate at all whether the method of calculating a reasonable fee should be the same as the case in which the losing defendant was a private entity; instead, it appears that Congress only intended to make clear that the United States also would be assessed for costs when it was a losing defendant.

Turning to the policies and purposes of the attorney's fee provision set forth in Title VII, we find nothing which would dictate a market value approach. As the Supreme Court indicated in Newman v. Piggie Park Enterprises, Inc.,17 the provision for counsel fees in intended "to encourage individuals injured by ... discrimination to seek judicial relief," by enabling these individuals to obtain adequate counsel.18 Nothing in the legislative history of the provision indicates that Congress intended the attorney's fee to be computed according to the market value of the services rendered by the attorney: indeed, nothing in the legislative history indicates that Congress even addressed itself to the details of the method of calculating an award.

In regard to the Civil Rights Attorney's Fee Award Act of 1976,19 an Act similar in design and principle to the attorney's fee provisions set forth in Title VII, the policy underlying the attorney's fee provisions in civil rights cases is elucidated more fully: "(A)warding counsel fees to prevailing plaintiffs in (civil rights) litigation is particularly important and necessary if Federal civil and constitutional rights are to be adequately protected."20 To accomplish this goal, reasonable fees must be awarded "to attract competent counsel ... while avoiding windfalls to attorneys."21 The majority correctly asserts that both the House and Senate Reports to the 1976 Civil Rights Attorney's Fee Award Act cite with approval Title VII cases in which the attorney's fee was calculated according to a market value formula. But Congress found that in those cases cited the "fees (were) adequate to attract competent counsel, but (did) not produce windfalls to (the) attorneys."22

We do not think therefore that Congress intended the market value of the services rendered to be the basis for an award of attorney's fees when that technique would produce a windfall or unreasonable fee. As the court explained in Johnson v. Georgia Highway Express, Inc.,23 one of the cases Congress cited with approval:

The statute was not passed for the benefit of attorneys but to enable litigants to obtain competent counsel worthy of a contest with the caliber of counsel available to their opposition and to fairly place the economical burden of Title VII litigation.24

We think applying the market value approach in cases in which the Government is a defendant, as will be shown below, likely will lead to the award of clearly un reasonable fees, in direct contradiction to congressional intent. To the contrary, the cost-plus formula will best promote the congressional policy of encouraging deserving litigants to bring Title VII suits, without making the attorney's position so lucrative as to ridicule the whole notion of a "reasonable attorney's fee."

II. THE MAJORITY OPINION FORMULA-FAULTY ANALYSIS PRODUCES

In its rigid interpretation of the statute as calling for precisely the same method of calculating attorney's fees in both the public and private sectors the majority errs, and then compounds that error by applying its own formula in a way which precludes ascertaining the congressionally directed "reasonable attorney's fee." Both errors stem from the majority's failure to appreciate the inapplicability of fee setting, as done by the market in private practice, to fixing a reasonable fee in the very different situation when the Government is the defendant. We now turn to specific examples of the extraordinary skewed results which will be the inevitable consequences of the majority's faulty fundamental analysis.

We ourselves in our previous panel opinions recognized the desirability, and indeed necessity where public interest law firms were concerned, of applying a contingency factor to the basic fee awarded. However, (1) we did this to a basic fee calculated on actual cost, and (2) the contingency factor was to be part of the reasonable profit, varied and controlled by the trial judge on his appraisal of several factors, including the contingency nature of the firm's practice and the particular lawsuit. When the contingency factor is applied to the market hourly rate, as the majority would do, the results are confusing and can lead to excessive awards.

Per Calculos, the majority's method, if allowed to stand, will take these and other Title VII attorney's fees Ad Astra. As applied to the total of 3,602 hours of work in this case, the weighted hourly rate of $57.17 results in a calculation of $205,916.50, close to the $206,000 fee "suggested" by the law firm to the court. This is designated by the majority opinion as the "lodestar" or "market value"25 from which all other "adjustments" are to be made.26 It is absolutely vital to see what this $57.17 hourly fee already includes. The regular hourly rates of the law firm, for each lawyer, are necessarily designed to cover the lawyer's individual salary or equivalent partnership pay, his appropriate share of the firm's overhead in every respect, a profit above the actual cost to the firm of his work (which makes up the total firm profit for the partners), and -this must be recognized and kept clearly in mind-an amount necessary for each hour which is billed to cover the numerous hours which for one reason or another cannot be billed, or must be billed at a more modest rate. The firm can never calculate its hourly charge for an attorney on the fallacious theory that every hour of work is going to be productive. There are hours which simply cannot be billed regular paying clients because the are redundant, or too numerous for the character of the task to which they are devoted; and in those instances in which the representation in litigation is on a contingent basis (i. e., as defined by the majority opinion, compensation only if the firm's side prevails), hours may not be compensable at all.

It is basic common sense that the bill for legal services in successful litigation may have a more comfortable margin than that for a losing effort. That margin, a sort of bonus for winning, acknowledges that litigators adjust their fees in accordance with each fluctuation in their win-loss record. Even in purely private suits, where law firms recover their fees from their own clients, fees and the underlying hourly rates on which they are computed are adapted to the particular circumstances. In the market, a request for fees or hourly rates not conforming to the results of litigation would be outrageous.

The fatal flaw is that our colleagues have taken a standard of values from the marketplace, indeed have referred to their "lodestar" fee at times as a "market value" fee, which it is, and have applied it in the Government sector where there is no real market. What our colleagues fail to realize is that the "market value" fee they have taken as a "lodestar," the starting point to be adjusted for contingencies, already has a substantial contingency factor built into the fee.

This is what a market is for. A market is to place value on commodities or services considering all of the contingencies.27 As we pointed out above, it is absolutely necessary for every private law firm to fix an hourly rate that takes care of the salary of the attorney and overhead attributable to him, provides a profit for the partnership, and also takes care of the contingency of those many hours for which there is no monetary compensation at all (one of the reasons for which is the contingency of not winning the case and not collecting a fee), or for which the firm must make a pragmatic, prudential decision to charge the client at a lower rate because the firm was not successful in litigation and could not rationally expect to recover as large a profit from the case as it might have had it won.

While in the practice of law the fixing of fees and establishment of customary rates is not as volatile as the New York Stock Exchange, the distinguished private firm involved in this litigation is, as our colleagues have rightfully recognized, in a marketplace of sorts. The regular hourly rates fixed by this and other firms reflect in every possible way the contingencies of the marketplace, including the contingency of failure in a litigated case (or the failure of its client in the marketplace, i. e., bankruptcy) and of the receipt of no fee at all, as could have occurred in this Title VII litigation. And so their fee in the marketplace is truly a fee calculated on the "market value" of their services. The majority opinion has thus pointed in numerous places to "market value" as the fairest and most useful starting part for the calculation of the plaintiff's fees in this successful Title VII suit against the Government. Our colleagues have then erroneously specified that this fee be adjusted upward in every case to take care of "contingencies," the primary contingency being that of failure to prevail in some lawsuits and thus the contingency of failure to receive any fee at all.28 Since the "market value" fee in commercial private practice already includes the contingency of failure and receipt of a lower fee than otherwise obtainable, our colleagues create a danger of duplication when they add this contingency factor to the already generous market value regular hourly rate.

The majority appears to recognize that market rates already include a substantial contingency factor, when it acknowledges the possibility that "an hourly rate underlying the 'lodestar fee' itself comprehends an allowance for the contingent nature of the availability of fees in Title VII litigation against the Government ...."29 This is more than just a vague possibility. As we have seen, adjustments for contingency can be expected to be commonplace, given the general nature of law firm billing practices. Consequently, allowances for contingency will generally be comprehended in the hourly rate, and the amount of contingency allowance may well be substantial.

To alleviate this problem, the majority suggests that "(t)he district judge has ample powers of inquiry into the makeup of hourly rates to assure that the Government will not suffer from any such duplication ...."30 If the majority is serious about weeding out redundant contingency allowances, these judicial powers of inquiry will always have to come into play.31 Given the serious possibility that any hourly rate may contain a contingency factor, the district judge will always have to inquire whether in fact it does contain such a factor. And if it does, he must inquire into its magnitude.

Determining the existence and amount of the contingency factor in any hourly rate is a difficult task, and the majority does not suggest how it is to be done. Although there may be several possible ways to do this, the most obvious one is to break down the putative hourly rate into its constituent parts in order to identify that component which reflects the contingency factor. The contingency factor would be that component of the fee in addition to the amount needed to cover costs and to provide the firm with its normally expected overall profit rate. In other words, it would be the amount needed to ensure that in the long run the firm earns its desired profit, after taking account of the proportion of hours spent on a case that must be billed at a lower rate if at all due to lack of success in litigation. It may very well be that a law firm's desired profit is an "unreasonable" one. In the marketplace this is no problem: "reasonable" is whatever the market will bear. Here, with hourly rates applied artificially in a nonmarket context, the potential for unreasonably high desired profits compounds the difficulty of isolating the contingency component.

But the most striking aspect of any technique employed, pursuant to the majority's approach, to identify the contingency factor, is that the elimination of built-in contingency allowances could be achieved more simply from the outset by employing the cost-plus method. Cost-plus provides a base figure that is free of any contingency factor; from that base figure the appropriate adjustments could then be made to reflect contingency of nonsuccess in the case at hand, as well as exceptional quality of work, without duplicating any built-in contingency factor already included in the fee.

Starting from actual cost of services is a far more direct approach than starting with an hourly fee and trying to weed out any built-in contingency allowances. Starting from the hourly rate simply invites confusion and duplication of contingency allowances. Unless district judges are especially diligent in weeding out any built-in contingency allowances, there is no way in which the market value regular hourly rate, the "lodestar" fee, can be taken as the starting point, and a contingency factor for failure applied to that market value fee without totally distorting and exaggerating the compensation awarded to successful plaintiffs' attorneys. We think that an evaluation of the contingency factor is necessary in Title VII cases, in all fairness to the attorneys who bring these suits-sometimes successfully, sometimes unsuccessfully. But the contingency factor can only be applied if the actual cost of services -salary and overhead-is taken as the starting point.

Actual cost of services has no contingency factor built in, as does the regular commercial hourly rate, which is fixed by the customary market, and which is truly a market value fee. Where the Government is the purchaser of services, actual cost plus is a fair and reasonable basis on which to compensate anyone, lawyer or layman. Actual cost of service is a true starting base, to which the factors relevant to Government litigation (and not necessarily relevant in private litigation) can be properly applied. What our colleagues have done here is a horrendous example of miscalculation and inflation of fees chargeable to the Government and to the taxpayer, without even realizing the economics of the marketplace on which they purport to rely.

B. Encouraging Injured Plaintiffs or Encouraging Lawyers?

Aside from the extraordinary financial results from the majority's theory, we have a fundamental disagreement with our colleagues as to the philosophy underlying the award of attorney's fees to prevailing litigants.

It is plain that our nation's policy-both legislative and judicial-is to promote the efforts of so-called "private attorneys general" who vindicate our civil rights laws by seeking legal redress for Title VII plaintiffs' injuries. It is necessary, then, that litigation expenses constitute no barrier that discourages these private plaintiffs from bringing their grievances before the courts. Eliminating the barrier of attorney's fees encourages plaintiffs to assert their legal and civil rights.

But encouraging injured plaintiffs is a goal distinct from that of encouraging lawyers with the lure of attorney's fees bonanzas.32 The majority's philosophy appears to be solicitous of the "sellers" of legal services beyond the needs of the "buyers" of these services. The majority appears to believe that its "market value" formula must award attorney's fees which match the petitioning lawyers' highest opportunity costs. Neither lawyers nor other purveyors of products and services operate on the basis of achieving highest opportunity costs most of the time; by definition, the usual-and entirely satisfactory-reward is less. It seems clear that the mandated encouragement to plaintiffs is achieved by granting the lawyers a sum reflecting their actual cost plus a reasonable profit, as opposed to an award which reflects the highest rates of return that alternative applications of legal manpower and resources could command.

It is simply not invidious to conclude that the fee schedule acceptable to General Motors when confronting a possible billion dollar liability is not necessarily applicable in Title VII attorney's fees determinations. Our "cost-plus" method brooks no disservice to Title VII "private attorneys general." On the other hand, overcoming the legal expense barrier for these private plaintiffs requires no windfall for lawyers-only that it be worth their lawyers' while.33 This is what our view of attorney's fees awards accomplishes: service for the plaintiffs without the need for lawyers to sacrifice.34 The inevitable existence of some other opportunities for lawyers to gain relatively higher remuneration does not mean that all legal services-including purported pro bono work-must be compensated at the very highest figure discoverable.

Under the majority's approach, in calculating an award of attorney's fees, the court should first multiply a reasonable hourly rate by the number of hours reasonably expended on the lawsuit, the so-called "lodestar" fee. Adjustments to this figure then may be appropriate, the majority asserts, to account for the quality of representation in the particular case and the contingent nature of success. Another logical flaw in the majority's formula is that a consideration of the quality of representation in the particular case, like that of contingency, see Part II.A. supra, already inheres in the reasonable hourly rate, one of the two elements used in fixing the "lodestar" fee. The majority reveals this logical gap plainly, yet it does not seem to recognize it.

The majority states that a reasonable hourly rate "is the product of a multiplicity of factors. Evans itself listed several of the relevant considerations (including, inter alia) the level of skill necessary (and) the attorney's reputation."35 The attorney's reputation corresponds to a consideration of the quality of an attorney's work in general. As the court stated in Johnson v. Georgia Highway Express, Inc.,36 (from which the factors in Evans are drawn): "Most fee scales reflect an experience differential with the more experienced attorneys receiving larger compensation. An attorney specializing in civil rights cases may enjoy a higher rate for his expertise than others, providing his ability corresponds with his experience." The level of skill necessary to perform the legal service properly also corresponds to a consideration of the quality of an attorney's representation in the particular case before the court. In Johnson the court explained that this factor required the trial judge to "observe the attorney's work product, his preparation, and general ability before the court."37 So, both the lawyer's skill in general (reputation) and that necessary in the particular case are already accounted for in the reasonable hourly rate.

Although the skill of the lawyer in the particular case thus is already included in the calculation of the reasonable hourly rate, the majority states that "the 'lodestar' may be adjusted up or down to reflect 'the quality of representation.' "38 Under the majority's approach, then, the attorney presumable will be compensated for the quality of his representation twice: once when the court calculates the "lodestar" fee and a second time when an adjustment to the "lodestar" is made.

We might add that we do not find logically persuasive the majority's notion that a decrease in the "lodestar" may result if the quality of the representation was unusually poor. To be entitled to an award of attorney's fees, the attorney must prevail in the lawsuit. It stands to reason that the level of proficiency displayed will always be at least adequate (indicating no adjustment to the "lodestar") if not exceptional (indicating an upward adjustment to the "lodestar") in cases in which the plaintiff prevails. It is thus difficult to conceive how a downward adjustment would ever be justified.

D. Majority Failure to Apply New Formula to this Case

An additional flaw of major proportions in our colleagues' opinion is the faulty application of their new theory and model to the case at bar. They say, avowedly because of the age of the case and the position taken by the parties on appeal, that they will not send this case back to the trial court for an examination on the basis of the revised formula they have devised,39 even though "(i)t is readily apparent that the District Court's fee-setting calculations do not precisely conform to the procedures identified in earlier cases and elaborated upon earlier in this opinion."40 This is really ducking the issue, the only issue on which this case was brought to this court.

In effect, the majority, perhaps influenced by this court's vast administrative law experience, has drifted into rulemaking like an administrative agency; the majority has made a rule for future cases, but declined to perform its primary function as a court-to adjudicate, to apply the rule to the very case at bar.41 We have never been aware that a court may decline to adjudicate simply because a case is long on the docket; what a convenient way of clearing crowded dockets that would be. Nor may a court announce prospective rules but decline to apply them to the litigation before it on the speculation that the district court might have reached a result compatible with the new rule.

The "lodestar" approach would be better understood and supervised if the district judge here had a chance to run through the "lodestar" and exercise his "ample powers of inquiry."42 In order for an accurate fee to be set, the district court should have a chance to apply the majority's formula, looking into the contingency factors in this case and the possibility of a built-in contingency allowance in the hourly rate, regardless of the procedural posture of the parties. A remand is in order so that we may have an indication of how the formula works in practice.43

III. "A REASONABLE ATTORNEY'S FEE"-ACTUAL COST PLUS A

In considering how to achieve the statutory goal of "a reasonable attorney's fee," we have delineated the undeniable distinctions inherent in levying fees against private employers as compared to the Government, where in the latter situation there is no market and therefore no market value constraints on fees, but on the contrary the entire mechanism of affording relief to aggrieved Government employees is created by and financed by the Government on both sides. We have seen how the unthinking utilization of the "market value" fee as the "lodestar" or basic fee results in gross escalation of fee awards against the taxpayers, far from the "reasonable attorney's fee" mandated by statute. Now we turn to outline the actual cost plus a reasonable and controllable profit method of fixing a reasonable fee, first enunciated in our panel opinion for the court, pondered and refined in the light of comment and discussion during the almost two years since our first opinion issued. We are convinced that the cost-plus formula is likely better to achieve the statutory goal then any other method yet proposed, and no case is apt better to illustrate this than the case at bar.

Returning to the specific facts of this case, putting the matter in perspective, the labors of the two young attorneys (plus some hours of partners' time) resulted in a promotion to a higher GS level and back pay of $4,169.80 for the plaintiff Copeland. In addition, following the negotiation of a settlement, remaining members of the class secured several promotions and back pay awards totaling $27,175.71. The law firm "suggested" a fee of $206,000, plus $12,602.59 in costs. District Judge Gesell awarded a flat $160,000 in fees and $11,567.11 in costs.

A. The Need for Additional Guidelines for the District Court

As witness this case, the application of the existing standards led to a claim of $206,000 by a responsible law firm and an intelligent, experienced judge fixing a fee of $160,000 on a $31,345 monetary benefit to all members of the class. To state the matter this baldly is to make out at least a prima facie case that something is wrong with the previously constructed standards-when applied to fix a fee award in a Title VII case against the Government.

Previous Title VII attorney's fees cases before this court have involved private party defendants. While the statute provides that "the United States shall be liable for costs the same as a private person," it is rather likely that awards of attorney's fees against a private company, found guilty of race or sex discrimination, have contained a certain amount of a punitive element and have been scrutinized less sharply than awards against the Government (all taxpayers) should be. Hence, the panel of this court believed that it should give careful scrutiny not only to this particular award but to the standards which were to be followed subsequently in attorney's fees cases against the Government.

In this opinion, Part I, we examined at some length the unique factors affecting attorney's fees in Title VII litigation against the Government, an analysis which is an elaboration and development of thoughts set forth in our first panel opinion. At that time it was readily apparent that something was needed as a substitute for the commercial fee basis; hence, the court's original opinion suggested that actual cost plus a reasonable and controllable profit be substituted instead of the market value rate as the initial starting point for the district court's calculation.

We pointed out the rather surprising fact that when law firms assert the value of their work to their clients-itself a rather nebulous concept, as witness the claim here of $206,000 plus expenses versus the $31,345 in back pay plus promotions awarded-the firm "never reveals the value of the attorneys' work to the firm, i. e., the value of the gross income brought to the firm by the attorneys in the ordinary course of business, as compared with the sums paid out to those attorney's as personal income and to defray overhead costs attributable to the maintenance of the attorneys in the firm."44 It seemed only reasonable to us that where a firm is ostensibly performing a pro bono publico service, its reimbursement for that service should bear a direct relationship to the actual costs incurred by the firm and, in fact, that this was the best possible reference point, at least in the initial calculation.

We therefore stated: "Thus the trial court should give consideration to abandoning the traditional(ly) claimed hourly-fee starting point for its calculations in favor of a principle of reimbursement to a firm for its costs, plus a reasonable and controllable margin for profit. Such a principle can be applied through separation of the several hidden components of the usual attorney's fee."45

In cases such as this, in which the Government is the alleged offending employer, the Government does not undertake the prosecution of the suit for the benefit of the aggrieved private individuals, as it can where there is a private employer involved. The individual Government employees must turn to private firms for help, and thus the private firms are in effect acting as "private attorneys general," and, in fact, they are the only "attorneys general" who can assist the individual employees. The private law firm, in other words, is performing the same function that Government attorneys frequently perform in cases against private employers. This is a pro bono publico service, or so it is claimed to be, with the objective of bringing about fair and equitable treatment of its employees by the Government itself.

Where such a pro bono legal service is involved, what could be a better and more fair measure to the law firm than its actual "cost, plus a reasonable and controllable margin of profit"? This is the same cost-plus formula which has been applied for innumerable years in innumerable government service contracts. It is recognized as fair and equitable; the usual criticism is that it results in a greater award to the private contractor than if he had been forced to made a competitive bid and be stuck with it. Such a competitive bidding system, of course, is inapplicable for legal services and no one suggests that it should be adopted. Contrary to the rather surprising and unnecessary language of the law firm's petition-"Lawyers and law firms are not public utilities, and cost-plus information is irrelevant to determining the reasonable value of their services in the marketplace"46-this cost-plus formula is a well recognized and equitable one for many various type Government service contracts.

The market rate is usually stated as so many dollars per hour, but we all known that in fixing fees in private practice this is only the beginning. Unfortunately, in fixing fees by a court it is too often both the beginning and the end, and in many cases-the instant one for example-the result is indefensible on a commercial basis. In fixing fees in private practice the hourly rate for all the hours worked is a starting point, plus a consideration of the benefit to the client, the client's ability to pay, the previous business with that client or business hoped to be gained in the future, the relationship of the billing lawyer with the lawyers and the clients on the other side, the alternative work on which the lawyers for whom the fee is billed could have been engaged-all these factors are taken into consideration. Not many of them are readily applicable to Title VII suits against the Government. If the market value hourly rate is not to be modified by the usual factors brought into play in a private commercial case, then it is an unsafe and unrealistic starting point. Since pro bono government legal work is involved, actual cost (salary plus overhead) should provide a more accurate starting basis, to which should be added a reasonable profit.

To the extent that the cost-plus formula and the other factors mentioned in the opinion might result in a lower than commercial fee, government legal services-and here the private law firm is acting as a private attorney general in a pro bono publico suit-have always been paid less than those in the private sector. For example, Justice Department lawyers do not start at the beginning salary paid at large law firms in Washington, D.C. The very top Justice Department lawyers do not even approach the earnings of partners in large firms such as that involved here. Government legal work has never been expected to pay the same as the top private legal work. What the law firm here and other large law firms, or public interest law firms, are doing is government legal work -on the other side of the issue in these cases from the government employer itself.47

Our colleagues are insistent on using "market value" as the correct basic concept in fixing fair and reasonable legal fees. Very well. Our colleagues should recognize the market in which we are dealing in a Title VII case against the Government: we are dealing in a "market" created by the Government,48 we are paying lawyers to defend the Government and to sue the Government, we are dealing in the "market" for government legal services; "a reasonable attorney's fee" must bear some relationship to the usual compensation for government legal services.

What our "cost plus a reasonable and controllable profit" guideline does is to return the actual out-of-pocket cost to the firm for its attorneys' legal services and all overhead, plus a reasonable and controlled profit. How much more should a private law firm receive for its legal services? How much more than cost plus a reasonable profit can it be entitled to? How much more than cost plus a reasonable profit would this court be justified in awarding against the Government and its taxpayers?

The petitioning law firm conjures up enormous difficulties of application by the court and of prying into confidential firm matters. We see absolutely none of this, if the proposed cost plus reasonable profit formula is applied on any sensible basis. If the trial court understands what it is doing, there should be no substantial additional evidentiary burden on the trial court. The basic figures are simple and simple to arrive at, as discussed below.

The three general components we identified in our opinion were salary, overhead, and profit.

1. Salary -The starting salary for young lawyers with the large firms in Washington, New York, and many other large cities is almost a matter of public record. The "going rate" for hiring is strictly competitive and well known to both the law firms and the young lawyers coming in. Furthermore, even the raises for the first few years are standardized.

And, of course, while our opinion talked in terms of the salary actually paid the individual lawyers involved in the case, it would be entirely satisfactory if the law firm merely furnished information on the average salary paid young lawyers in that firm with the same number of years' experience. (In the instant case, both associates were assigned this matter in their first year of employment at the firm.)With regard to partners' "salary," our opinion deliberately put this on the basis of an extrapolation from the highest associate's salary. We recognize that the income of partners in the same firm may vary widely, even more widely between firms, and a partner's share may very well be thought of as confidential. Typically, in Title VII cases, the share of the fee claimed for partners' labors is usually very small, as in the instant case.

2. Overhead -The firm knows this at least annually, or it is not computing its income tax correctly. If the firm is well managed, it should know its overhead factor quarterly or even monthly. Every firm surely makes a calculation as to the average overhead factor for its individual fee-producing lawyers.

One of the preposterous arguments, but strongly made, is that most firms now make no accounting of costs, overhead, and profit on an attorney-by-attorney basis, and that a mountain of details will be called for. This rests totally on a misinterpretation of our opinion. It is the average overhead cost per attorney to which the opinion refers.49

It may be that partners in fact develop more overhead cost for the firm than do associates (bigger offices, more luxurious furnishings, etc.), but this could be taken care of by an average overhead figure for associates and a different figure for partners.

In any given city, there is probably no great difference in the overhead per lawyer of similar sized firms in similar practice; hence, there are no great secrets to be revealed. If there are great differences in firm overhead, the firm management claiming a higher overhead factor is out of line and should know about it.

The claim is made that since the cost-plus formula allows all overhead costs to be shifted to the losing party, this cuts down the incentive to keep overhead costs low. Unless the firm is engaged virtually exclusively in Title VII work, this would not be so at all. The overhead cost average would reflect all of the work done by all attorneys in the firm on all type cases.

3. Profit -Our opinion made no suggestion to the trial court as to what is a reasonable profit. This is a calculation involving many factors, including the attorneys' usual profit return to the firm, the social benefits, the direct gain to the litigants, the skill demonstrated in the particular case, and the degree of contingency involved.

D. Substantive Inequities Feared from the New Method

1. Special Problems with Small Firms and Solo Practitioners

a. The argument is made that small law practices will not be able to prove expenses as high as those of better established practices or larger firms, and thus will not be able to secure equivalent fees for the same work.

This may or may not be true (small firms frequently have higher per lawyer overhead), but the question of course arises: Is there any alternate work in which such firms would have been engaged in which they would get greater fees? The answer is that these firms, perhaps composed of younger lawyers, will get just as good fees as any alternative work they could possibly do, and probably better at this stage of their careers.

A more fundamental answer is that if both small and large firms obtain a full return of their actual expenses, plus a reasonable profit, there can be no inequity in the treatment of large and small firms.

b. It is alleged that the formula is inapplicable to a solo practitioner, because there are no guideposts as to his "salary" from his associate's salary or his partner's income.

Perhaps the solo practitioner is a case in which a fairer award can be made without using the cost-plus formula, and there is nothing in our opinion which mandates the cost-plus formula in every case. However, it would be possible to take the going salary rate of several firms for lawyers of the same years of experience, extrapolate if need be, and award the solo practitioner a fee based on that plus his overhead and a reasonable profit.

2. Similar Problems for the Public Interest and Civil Rights Bar

A similar claim is made for these specialists, alleging that the public interest and civil rights firms operate under far lower salary, overhead, and profit margins than others in private practice. The argument is made that somehow cost plus reasonable profit penalizes these firms.

The answer is the same as to a small firm or solo practitioner. They will be guaranteed a reasonable profit above their actual costs. We are not aware that public interest and civil rights firms usually receive more than this. However, the trial court could-and, in our view, should -evaluate the special skills which a public interest or civil rights firm may bring to bear among the other quality factors which the court applies to the actual cost plus reasonable profit figures. If the attorneys are specialists in Title VII cases, and thus may have been able to do the work at a high standard with a minimum of hours expended, then their "profit" in the total compensation awarded should indeed be higher than it otherwise would.

It is also argued that most Title VII cases are on a contingent fee basis and thus a successful Title VII case should produce a substantial reward in order to allow their practice to continue. We agree. Contingency is a factor which should be evaluated at the time the trial judge is pondering the "reasonable profit" part of the fee to be allowed. A "reasonable profit" in a contingency fee case should be higher than where the fee is relatively certain and the only question decided by the outcome of the case is who pays it.50

The argument is made that somehow a fully remunerative contingency fee recovery from the Government will not be had, and therefore class action claims cannot be sustained. The basis of this is obscure, for it seems readily apparent that the formula of actual cost plus a reasonable profit is in every way "fully remunerative," especially when it is remembered that other factors are to be taken into account, as set forth above.

5. Counsel Representing Poor Clients Generally

It is claimed that counsel in this type practice usually have a very low salary overhead and profit margin. If this is true, the cost plus reasonable profit will guarantee them a return that will be at least equal to that received in their usual practice, and probably better.

It is argued that those who habitually work long hours will be disadvantaged by any formula based on the average hourly rate of return, since their average hourly return will necessarily be low. If that is true, then the return in these Title VII cases will in no way differ from the alternative work which they might be doing.

The theory implicit in several of these hypothetical problems raised above is that somehow compensation for the Title VII work should be a bonanza to lawyers in these particular cases, in order that they may continue their practice in other worthwhile but relatively unremunerative type cases. We cannot see that this is a valid argument at all for fixing a fee against the Government-in fact, it virtually admits that some lawyers have been relying on Title VII work to gain relative bonanzas.

7. Requirement of Public Disclosure of Financial Information

This is the bugaboo that good firms will simply be unwilling to disclose comprehensive financial information.51

As pointed out above, the required information is not nearly so "comprehensive" as the parties make it out to be. Unless a firm is drastically out of line in its salary scale or its overhead costs, no detailed data need be submitted to the court. An affidavit of a partner that, based on its accounting records, the average overhead per lawyer was so many dollars an hour and the average salary paid to, for example, lawyers with two years of experience, was a certain amount, should be sufficient, unless this appeared drastically in error when compared to firms similarly situated.

E. Fear of Deterring Representation in Title VII Cases

The law firm argues that for three principal reasons the cost-plus formula will deter lawyers from taking on Title VII cases: (1) lower rates of remuneration; (2) difficulty of compiling relevant data; (3) disclosure requirements.52

As to prospective lower rates of remuneration, the question immediately arises, what kind of remuneration have these lawyers been receiving in Title VII litigation? Have they been receiving their customary salary, all overhead, plus an un reasonable profit? If the latter, it is high time the courts and the taxpayers knew about it.

If these Title VII litigators have not been receiving fees amounting to all their customary costs plus an un reasonable profit, their fears of the cost plus reasonable profit formula are groundless.

One source of these firms' apprehension may be that they are not sure how a "reasonable profit" will be calculated. Bear in mind that, as explained above, among other factors bearing on profit a trial court should take into account: any contingent nature of the fee, and the extent to which the particular firm depends upon contingent fees; the expertise in Title VII matters which the lawyers bring to the case, which would favor public interest or civil rights firms who could validly claim a larger profit on their specialized line of practice; and the social benefits arising from a class action of this type. Conversely, another consideration, which the trial court should NOT consider as increasing the profit portion of the fee, is a large overhead cost per lawyer hour. This may reflect, and certainly might encourage, inefficiency. While both the large commercial firm and the small firm or solo practitioner devoting itself largely to poor clients should be recompensed in full for actual overhead costs, the lawyer with the larger overhead should not receive a larger profit component of the total fee because of the overhead.

Indeed, it is in the trial court's evaluation of the "reasonable profit" component that the factors we have discussed herein and in our previous opinion come into play to modify, one way or another, the data as to the profit derived from the services of the lawyers concerned.

As to the alleged (2) difficulty of compiling relevant data and (3) disclosure of confidential data, these have been discussed above.

On analysis, the fears expressed by the appellee law firm and some amici as to any deterrence in doing pro bono work are wildly exaggerated. The Attorney General, who after all bears some responsibility for the public interest, while not receiving all he sought in the opinion of the panel, had a much more realistic appraisal in his Memorandum submitted to the court:

What was of principal concern to the Court, and what is of concern to the Attorney General, is recognition of the strong public interest that the computation of reasonable Title VII attorneys' fees "not mechanistically set as its cornerstone the 'customary' and unadjusted fees charged by private attorneys in their unrelated and most highly paid lines of work." In this regard, what the panel stated merits repetition here:

It is not our intent to establish a subclass of fees payable for Title VII litigation or to relegate, in any way, the vindication of Title VII rights to second-class status. To do so, obviously, would run counter to the very purpose for which Title VII was enacted .... But it is our intent to resist the imposition on the public sector of the highest standards of legal remuneration adopted in the private sector-standards which are out of line with the ethic of public service with which attorneys are encouraged to engage in public interest litigation, and which may even risk turning the public against the very provisions for the award of attorneys' fees on which appellees rely in this case. (594 F.2d at 257 n.75.)

The Attorney General, who is committed to the concept of reasonable attorneys' fees for prevailing Title VII plaintiffs, endorses this statement of the panel.53

We also agree with the Attorney General's interpretation that "(t)he panel soundly found that the evidentiary record in this court was completely inadequate for the determination of attorney's fees by any method."54 We find totally inexplicable the majority's refusal to remand this case to the trial court on any theory-the majority's own new formula, previous precedent, or our cost plus reasonable profit method. Not only was the evidentiary record inadequate, but there was a complete failure by the trial court to articulate a rationale on any theory. Is the award of $160,000 really defensible on any method of calculating "a reasonable attorney's fee"?

(T)he issue of attorney fee compensation for Title VII plaintiffs claiming unlawful discrimination by the federal government is new, difficult, and complex. The traditional customary approaches, as demonstrated by what has occurred in this case, may not be adequate. More refined analysis and consideration of alternatives are required. When representing a private client, attorneys must exercise billing judgment; they must consider the labor expended in view of the result to be achieved. This economic judgment is absent when the federal treasury is footing the bill. Other solutions must be explored. This is what the panel has wisely suggested.55

We in dissent are convinced that the traditional customary commercial fee approach to billing the Government for attorneys' fees in a Title VII case is not adequate and is likely to lead to grossly excessive fees. A new approach is necessary. We have suggested that the trial court apply an actual cost plus reasonable profit formula in this case. Our analysis, in our original opinion and herein above, indicates that this formula would provide a much more precise and equitable basis, both to the Government and the private attorneys engaged in Title VII practice, for the award of attorneys' fees. We have not ruled out any other innovative methods which may commend themselves to the trial court. Since the trial court has yet to hold a hearing in this case, we think it should do so, and that in adducing evidence it should do so along the lines necessary to provide a foundation for the application of the actual cost plus reasonable profit formula. Then the trial court will have the opportunity to apply the actual cost plus reasonable profit formula to the established facts of a comparatively complex attorneys' fees case. The courts and all litigants engaged in Title VII litigation will benefit.