This case, which involves an appeal and cross appeal from an order of the district court awarding counsel fees to prevailing parties (two public interest groups) pursuant to the fee shifting provisions of the Clean Water Act, 33 U.S.C. Secs. 1251 et seq., 1365(d) (1982), raises the interesting and difficult question (on which the circuits are sharply divided) of how to calculate the counsel fees for attorneys who operate a for-profit law firm that handles only public interest cases and therefore bills significantly less than a traditional law firm. In litigating this case the firm had no traditional fee arrangement with its clients but pinned all hope of repayment on the fee shifting statute. The firm has a billing rate ($60-$80 per hour) which it applies in certain cases, but did not apply in this one. In cases where the firm has a fee arrangement with its client, it concedes that it receives the $60-$80 rate, but only if it does not prevail and hence is not eligible for a court-awarded fee. Where, as in this case, the firm has no fee arrangement it receives no compensation whatsoever if it loses. In either case, when it does prevail, it receives a fee from the court. The question here is: what should that fee be?

The district court calculated the firm's lodestar by multiplying the number of hours times the prevailing rate for equivalent legal services in Washington, D.C., where the firm's offices are located. It found as fact that conventional firms performing work of equivalent complexity in the city bill at $85-$185 per hour rather than the firm's $60-$80 billing rate. In reviewing the district court's order, we must determine whether the factual findings were clearly erroneous, and also evaluate the legal standard the district court applied.

For the reasons discussed below, we will affirm the district court's decision, 643 F. Supp. 961, as to calculation of the lodestar. We hold that its calculation of the market rate was not clearly erroneous and that it did not apply the wrong legal standard by employing the prevailing "market rate" for equivalent legal services rather than the plaintiff's normal billing rate. We find the determination of the legal standard to be a difficult one with no clear solution. However, after reviewing various approaches employed by other courts of appeals, we reach the conclusion that the method for calculating fees for this unique genre of attorneys, who operate for profit but essentially rely on fee shifting statutes, must depend on local rates for similar work. In coming to this conclusion we rely both on the Supreme Court's holding in Blum v. Stenson, 465 U.S. 886, 104 S. Ct. 1541, 79 L. Ed. 2d 891 (1984) (non-profit legal aid society is entitled to market rates even though it possessed no conventional hourly billing rate and would have received no fee from its clients), and on our assessment that this rule is the fairest, least burdensome way to calculate a reasonable fee.

The appeal and cross appeal raise five additional fee-related questions. First, we must determine whether the district court abused its discretion in failing to award plaintiffs an enhancement for contingency. This inquiry requires us to interpret the Supreme Court's recent opinion in Pennsylvania v. Delaware Valley Citizens' Council for Clean Air, --- U.S. ----, 107 S. Ct. 3078, 97 L. Ed. 2d 585 (1987) (Delaware Valley II ), which was decided after the district court issued its opinion in this case. We are doubtful, in light of Delaware Valley II, that this case could qualify as "risky." Although the firm relied exclusively on fee shifting statutes, which indicates the existence of risk of non-payment, their chances of losing were small. Even if this case could be labelled as risky, we would decline to grant a contingency multiplier. Not only would computing the appropriate contingency multiplier in light of Delaware Valley II be a most burdensome task for the district court and an expensive one for the parties under these circumstances, but we are also convinced that the hourly rate used by the district court in this case will be sufficient to attract competent counsel, thus satisfying the raison d'etre of the fee shifting statute as explicated by Delaware Valley II. Therefore, any further enhancement is unnecessary.

Second, we must determine whether the district court abused its discretion in denying the plaintiffs an enhancement for quality. The district court did not award a quality multiplier because it found that plaintiffs' attorneys' performance, though excellent, was not so extraordinary as to deserve such enhancement. Given that quality multipliers are appropriate only in rare cases, and that we agree with the district court that this is not one of them, we conclude that the district court did not abuse its discretion in denying a quality multiplier.

Third, we must determine whether the district court abused its discretion in failing to award plaintiffs an enhancement for delay. We will affirm the district court's denial of an enhancement for delay because plaintiffs waived the point by failing to develop a record on the issue in the district court.

Fourth, we must decide whether the district court abused its discretion by failing to require plaintiffs' attorneys to allocate some of the time spent litigating the attorneys' fees issues to the firm's other cases raising the same issues. The district court did not make factual findings on this issue, hence we will remand for findings as to whether any of the firm's other cases presented sufficiently similar factual issues and arose at such a juncture that the firm should have allocated part of the costs of litigating the fees issues to other cases.

Finally, we must consider whether the district court abused its discretion in refusing to apply a negative multiplier to the lodestar claimed for the attorneys' fees litigation because plaintiffs' attorneys did not meet with complete success in litigating the attorneys' fees issues. Because the district court did not consider this question at all in its disposition of the fees issue, we will remand it to the district court. Hence we will affirm in part and reverse in part and remand for further proceedings.

Plaintiffs, Student Public Interest Research Group of New Jersey and Friends of the Earth, Inc. (hereinafter referred to collectively as SPIRG), sued defendant AT & T Bell Laboratories (Bell) alleging that Bell violated the Clean Water Act by dumping pollutants into the Whippany River in New Jersey in excess of amounts allowed by law. After the district court granted plaintiffs' motion for summary judgment on the question of liability, the parties settled, agreeing that plaintiffs would drop all their claims in exchange for Bell's paying $75,000 to the United States Treasury.

Plaintiffs' attorneys moved for attorneys' fees pursuant to the Clean Water Act.1 The chief recipient of the fees, the Washington, D.C. law firm of Terris, Edgecombe, Hecker & Wayne (Terris), acted as lead counsel in the original litigation and in the litigation over the fee award. Terris is a private, for-profit law firm which specializes in public interest law. It serves as counsel in employment discrimination, civil rights, consumer, and environmental suits. When it charges clients anything at all, its billing rate varies between $60 and $80 per hour for partners' time and between $60 and $70 per hour for associates' time.2 It is uncontested that in this litigation that the Terris firm had no fee arrangement, and would have received no remuneration from SPIRG. Instead, the firm relied entirely on the attorneys' fees provision of the Clean Water Act.

In calculating its requested lodestar (reasonable hourly rate multiplied by number of hours billed), the Terris firm did not use its own billing rate, but rather calculated what a reasonable market rate would have been for legal services of the caliber Terris provided. The firm apportioned part of its time for its work on the merits to twenty-five other cases that the firm was litigating that raised similar Clean Water Act legal issues. Terris subsequently supplemented its fee applications to include time spent on litigating the attorneys' fees issue. In addition, Terris requested a 50% multiplier to the merits lodestar for quality, the contingent nature of the case, and delay in payment.

Bell objected to the hourly rate employed to calculate the lodestar, arguing that Terris was bound by its own billing rate. In addition, Bell complained that Terris did not apportion its hours among the many cases in which it litigated the issue of attorneys fees as it had done with the merits calculation. Finally, Bell opposed requests for enhancement of the merits lodestar for quality, contingency, and delay. Finally, Bell requested that if the district court did not fully grant Terris such multipliers, it reduce the lodestar for the time spent in seeking attorneys' fees on the ground that Terris achieved only partial success in pursuit of its fees.

The district court adopted Terris' lodestar calculation using "the market rate for attorneys of comparable experience in the same city" rather than Terris' actual billing rate. J.A. at 23. The district court, acknowledging that an attorney's regular billing rate is usually the best approximation of the market rate, nevertheless found in this case that Terris' billing rate did not adequately reflect market rates. The court cited Blum v. Stenson, 465 U.S. 886, 104 S. Ct. 1541, 79 L. Ed. 2d 891 (1984), for the proposition that actual billing rates do not always reliably indicate market rates. In Blum, the Supreme Court held that a non-profit legal aid organization deserved market rates if it prevailed under a statute that provided fee shifting. The district court reasoned that the principle of Blum should also apply to "for-profit" public interest law firms that give cut rates to poor people and to litigants raising public interest questions, explaining: "The fact that the plaintiff chose a firm which makes financial sacrifices in the public interest should not result in a windfall for the defendant." J.A. at 28.

The district court found as a fact that the market has set no customary hourly rate for Terris' public interest work, and that plaintiffs' attorneys are not in " 'customary private practice.' " J.A. at 26 (quoting Laffey v. Northwest Airlines, Inc., 746 F.2d 4, 18 (D.C.Cir.1984), cert. denied, 472 U.S. 1021, 105 S. Ct. 3488, 87 L. Ed. 2d 622 (1985) quoting Berger, Court Awarded Attorneys' Fees: What is 'Reasonable'?, 126 U.Pa.L.Rev. 281, 292 (1977)). Additionally, it found that the firm's billing rates were substantially lower than the market rate. It therefore refused to use Terris' normal billing rates and instead substituted what it found to be market rates for legal work of equivalent complexity and quality in calculating the lodestar figure.

The district court denied SPIRG's request for enhancement multipliers for quality and contingency. Its order did not discuss delay. The court also did not address Bell's claim for apportionment of the hours spent litigating the fee issues to other similar cases. Finally, the district court did not consider whether the lodestar for litigating the fees issue should be diminished in light of SPIRG's failure to prevail on the enhancement multipliers. The parties have stipulated that pursuant to the district court's opinion the total amount due to plaintiffs for attorneys fees and expenses equals $111,877.31. This figure reflects over $56,000.00 for the merits lodestar. The rest is attributable to the fees and expenses surrounding the fee litigation.

Bell argues that the district court's refusal to apply Terris' normal billing rate derives from clearly erroneous findings of fact and constitutes an abuse of discretion. It argues that there is indeed a market rate for Terris' services: Terris' normal billing rates. Bell contends that Terris' rate is a market rate that reflects the firm's affirmative choice to compete in a less remunerative market. It refers to other for-profit public interest law firms in the Washington, D.C. area to establish that Terris' normal billing rate is about what the market will bear for for-profit public interest lawyers who do Title VII and environmental work. Bell relies on the policy behind the fee shifting statutes, arguing that they are designed for compensation, not reward, and that the best approximation of attorneys' fair compensation lies in their own billing rates.

Additionally, Bell submits that the appellees' standard would engender an administrative nightmare in that it would force the court to search out equivalent markets and determine their value. Bell convincingly argues via its own litigation experience the administrative hassle and the waste of courts' time in determining fee awards. Bell champions its billing rate rule as simpler, as well as fairer.

SPIRG, in advocating the district court's approach, argues that its case represents an exception to the general rule of applying actual billing rates. SPIRG argues that the Terris firm's billing rates do not reflect independent market rates, but can only be interpreted in light of the existence of fee shifting statutes. It contends that Terris' billing rate does not offer a valid shortcut for determining the market value of SPIRG's services because its rate, when it applies at all, constitutes a form of contingency rate whereby the $60-$80 per hour billing rate applies if the attorneys lose, but court awarded fees are provided if they win.

In advocating a market rate derived separately from actual billing rates, SPIRG points to the anomalous results of refusing to pay attorneys the true market worth of their services. In a public interest case litigated in part by corporate lawyers pro bono publico and in part by attorneys of the Terris firm, the same work of equivalent quality would receive very different compensation.

SPIRG emphasizes that Congress intended the fee shifting statutes to promote attorney representation of legitimate claims that would otherwise go unpursued. At oral argument, counsel pointed to evidence in the record that its contingency billing rates of $60-80 could not support its public interest practice and that the firm depended on ultimately receiving market rates in fee awards to remain viable.

Finally, SPIRG argues that calculating the market rate is not as difficult as Bell would make it seem. It argues that the hourly rate commanded by counsel for Bell of $205 per hour provides a good benchmark.

B. The "Lodestar" Principle; Scope of Review

The Supreme Court in Blum v. Stenson, 465 U.S. 886, 104 S. Ct. 1541, 79 L. Ed. 2d 891 (1984), explained that the "initial estimate of a reasonable attorney's fee is properly calculated by multiplying the number of hours reasonably expended on the litigation times a reasonable hourly rate." Id. at 888, 104 S.Ct. at 1544 (citing Hensley v. Eckerhart, 461 U.S. 424, 103 S. Ct. 1933, 76 L. Ed. 2d 40 (1983)). Courts refer to the product of this equation as the lodestar. See Pennsylvania v. Delaware Valley Citizens' Council for Clean Air, 478 U.S. 546, 106 S. Ct. 3088, 3096-99, 92 L. Ed. 2d 439 (1986) (Delaware Valley I ) (explaining the genesis of the lodestar approach and nomenclature). In evaluating the district court's calculation of the hourly rate for Terris' lodestar, we must determine: (1) whether the district court committed clear error in its factual findings, Black Grievance Committee v. Philadelphia Electric Co., 802 F.2d 648, 652 (3d Cir.1986) ("[t]he question of an attorney's marketplace billing rate ... is a factual question which is subject to a clearly erroneous standard of review") (citing In re Fine Paper Antitrust Litigation, 751 F.2d 562, 591 (3d Cir.1984), vacated on other grounds, --- U.S. ----, 107 S. Ct. 3255, 97 L. Ed. 2d 754 (1987)); and (2) whether the district court used the proper standard to calculate the lodestar. See Blum v. Stenson, 465 U.S. 886, 104 S. Ct. 1541, 79 L. Ed. 2d 891 (1984) (without so articulating, reviewing de novo the standard for determining the appropriate hourly rate).3

We are satisfied that the district court's factual determinations that: (1) Terris does not engage in customary private practice; (2) Terris "uniformly charges its clients rates dramatically lower than the market rate," J.A. at 26; and (3) the appropriate market figure for equivalent legal work in Washington, D.C. by attorneys of comparable ability is between $85 and $185 per hour, were not clearly erroneous. The court carefully assembled and weighed evidence that the Terris firm performed excellent work, the value of which significantly exceeded the firm's normal billing rates. The court relied upon affidavits from numerous attorneys in the District of Columbia to determine the community market rate for legal service of equivalent quality and complexity performed by attorneys with similar education, experience, and skill.4 We turn to the appropriate legal standard.

Market rates have served as the prime focus of our inquiry in ascertaining reasonable attorneys' fees. See, e.g., Black Grievance Committee, 802 F.2d at 652 (restating inquiry to be the "attorney's marketplace billing rate" (emphasis added)); In re Fine Paper Antitrust Litigation, 751 F.2d 562, 590 (3d Cir.1984) (referring to this court's approach as the "market standards rule"); Hall v. Borough of Roselle, 747 F.2d 838, 841 (3d Cir.1984) ("[a] reasonable attorney's fee is one that compensates a lawyer for the fair market value of his time, experience, and effort").

The most important apposite case on market rates is Blum. In Blum, a unanimous Court rejected the petitioner's argument that non-profit legal organizations with salaried attorneys deserved only their litigation costs because awarding market rates would constitute a windfall for such organizations. The Court held that even though counsel for the Legal Aid Society of New York would not have received a fee from their clients, they were nevertheless entitled to a fair market rate under fee shifting statutes. 465 U.S. at 892-94, 104 S.Ct. at 1545-47. It held that reasonable awards under fee shifting statutes "are to be calculated according to the prevailing market rates in the relevant community, regardless of whether plaintiff is represented by private or non-profit counsel." Id. at 895, 104 S.Ct. at 1547.5

The Court observed that resolution of the question of how to calculate attorneys' fees "begins and ends with an interpretation of the attorney's fee statute." 465 U.S. at 893, 104 S.Ct. at 1546, and discussed at length the Senate Report on Sec. 1988 of the Civil Rights Act that addressed the question of fee shifting. S.Rep. No. 1011, 94th Cong., 2d Sess. 5 U.S.Code Cong. & Admin.News 1976, p. 5908 (1976) [hereinafter referred to as Senate Report]. Quoting both from the Senate Report and from cases cited with approval by the Report, the Court stressed that Congress envisioned employment of market rates in calculating attorneys' fees. It quoted the Senate Report's language that: "It is intended that the amount of fees awarded under [Sec. 1988] be governed by the same standards which prevail in other types of equally complex Federal litigation, such as antitrust cases." 465 U.S. at 893, 104 S. Ct. at 1546 (quoting Senate Report at 6, U.S.Code Cong. & Admin.News 1976, p. 5913.) Furthermore, Blum noted that the Senate Report cited with favor Stanford Daily v. Zurcher, 64 F.R.D. 680, 681 (N.D.Cal.1974), which held that courts "must avoid ... decreasing reasonable fees because the attorneys conducted the litigation more as an act of pro bono publico than as an effort at securing a large monetary return." See 465 U.S. at 895, 104 S. Ct. at 1547. See also Miele v. New York State Teamsters Conference Pension & Retirement Fund, 831 F.2d 407 (2d Cir.1987) (holding that Blum mandated using rates of the private bar in calculating the market rates of public interest attorneys and reversing contrary authority in the Court of Appeals for the Second Circuit).

Given that Blum requires "market rates," our question must be how to apply that concept under the unique facts of this case. We note that other courts have struggled with the very issue we consider here in trying to establish a fair rate governing for-profit public interest law firms.6 Although the caselaw is far from clear, we discern three distinct approaches among the courts of appeals and we have struggled with a fourth. In our view, none of these satisfactorily resolve this difficult question, but they all attest to its complexity. See Webb v. Maldonado, --- U.S. ----, 108 S. Ct. 480, 98 L. Ed. 2d 509 (1987) (opinion sur denial of cert.) (White, J., dissenting) (noting tension within the circuits concerning calculation of reasonable attorneys fees for attorneys with established billing rates). We discuss each of these approaches.

The first approach, which we call the "billing rate rule," forged by the District of Columbia Circuit and followed by the Eighth Circuit, mandates using actual billing rates whenever they exist. The billing rate rule rejects a market rate based on the local market in favor of applying the attorneys' actual billing rates, whenever such rates exist, even where attorneys set their rates artificially low to serve the public interest.

In Shakopee Mdewakanton Sioux Community v. City of Prior Lake, 771 F.2d 1153 (8th Cir.1985), cert. denied, 475 U.S. 1011, 106 S. Ct. 1185, 89 L. Ed. 2d 301 (1986), the court depended exclusively on counsel's billing rate. The court distinguished Blum from the case before it on the ground that the attorneys in the case before it possessed a billing schedule for similar work. The Shakopee court articulated its policy of avoiding "windfalls to attorneys" in awarding fees. Id. at 1160 (quoting Senate Report at 6, U.S.Code Cong. & Admin.News 1976, p. 5913). It reasoned that "the lodestar rate is the opportunity cost of the clients the firm turned away to represent the plaintiffs." Id. at 1160.

In Save Our Cumberland Mountains, Inc. v. Hodel, 826 F.2d 43, rehearing en banc granted, 830 F.2d 1182 (D.C.Cir.1987), Judge Bork addressed the same question:

[I]f an attorney has a customary billing rate, that rate constitutes the presumptively reasonable rate to use in computing a fee award. In general, only if the attorney himself has no customary billing rate may the court base its fee award on a composite average market hourly rate.

826 F.2d at 47-48 (lead opinion) (citing Laffey v. Northwest Airlines, Inc., 746 F.2d 4 (D.C.Cir.1984), cert. denied, 472 U.S. 1021, 105 S. Ct. 3488, 87 L. Ed. 2d 622 (1985)).7 Judge Bork dismissed, as irrelevant, evidence that the fee-seeking attorneys provided discount rates to environmental groups and that the attorneys performed most of their work on a contingent basis.8 Judge Bork opined that the panel's holding did not conflict with the policy announced by the Supreme Court in Blum, but rather, promoted it. Arguing that the existence of hourly fee rates indicated a willingness on the part of the attorneys to undertake this type of litigation, Judge Bork contended that the policy of allowing reasonable fees to ensure representation was satisfied.

The other two members of the panel, however, were considerably troubled by what they perceived to be a contradiction between the jurisprudence of the D.C. Circuit and the Supreme Court's decision in Blum. Judge Ruth Bader Ginsburg concurred in the result only because she felt bound by D.C. Circuit precedent in Laffey, "unless and until overturned by the court en banc or by Higher Authority." 826 F.2d at 54 (R.B. Ginsburg, J., concurring). Chief Judge Wald, in a strong dissent, essentially argued that the D.C. Circuit, by refusing to award market rates, had departed from Congress' intent in providing fee shifting and from the mandate of Blum. Judge Wald wrote that the actual billing rate standard may be appropriate for

standard private law firms that have an established billing practice.... But for the myriad of unconventional hybrid firms where major efforts are devoted to serving non-paying or lower than market rate clients, [the actual billing rate rule] may well be an unmitigated disaster, both for the lawyers involved and for the statutory purposes of the attorneys' fee provisions.

Id. at 57 (Wald, C.J., concurring in part and dissenting in part). Judge Wald relied on the policy of providing market rates, and argued that billing rates may only be substituted if they serve as a "valid proxy." Id. at 58.

We agree with Judge Wald's criticism of the billing rate rule, for as Judge Wald puts it, where alternative market rate calculations are available, wooden application of the billing rate rule contravenes the mandate of Blum as well as the policies of fee shifting statutes.

We acknowledge that, in most cases, billing rates reflect market rates--they provide an efficient and fair short cut for determining the market rate. See Black Grievance Committee, 802 F.2d at 652 ("[t]he reasonable value of an attorney's time is the price that time normally commands in the marketplace, which is generally reflected in the attorney's normal billing rate" (emphasis added) (citing In re Fine Paper Lindy I at 590-91; at 167)). See In re Fine Paper Antitrust Litigation, 751 F.2d at 590 ("[o]ur premise has been that the reasonable value of an attorney's time is the price that time normally commands in the marketplace for legal services in which those services are offered").

Although we have consistently relied on billing rates in determining market rates, this fact does not bar us from using other methods in a limited group of special situations where billing rates do not approximate a reasonable fee sufficient to attract competent counsel. In cases such as this one, billing rates alone fail to tell the full story.

It is clear to us that the Terris firm's billing rates do not approximate a fair market rate. We rely on the findings of the district court as well as our own observation of the superb advocacy skills of plaintiffs' counsel, Mr. Terris, as evidence that Terris' billing rate for its clients falls far short of what the Terris firm could command in the marketplace.

Although superficially appealing, Bell's argument that the market value of Terris' services is reflected in its billing rate is flawed because it ignores the effect of the fee shifting statutes themselves in the equation, for the existence of fee shifting statutes has itself become part of the dynamic equation that establishes the hourly rate of the Terris firm. See supra slip op. at 1441-42. To concentrate exclusively on Terris' billing rate without acknowledging that this rate for clients is in large part premised on the availability of the fee shifting statutes themselves distorts the true nature of Terris' hourly rate. We are persuaded that Terris' rates constitute a form of contingency billing, whereby the firm recovers a minimal fee if it loses and fair market prices via the fee shifting statutes if it prevails. Given its fee structure and the firm's dependence on fee shifting statutes for payment, it is doubtful that Terris' billing rates alone would be sufficient to attract competent counsel in the kinds of cases it handles.

Furthermore, we are not at all persuaded by the economic approach of the Shakopee and the Cumberland Mountains courts, which see fee shifting statutes as a replacement for opportunity costs. As Chief Judge Wald observed in dissent, this argument proves too much. See Cumberland Mountains, 826 F.2d at 55 (Wald, C.J., concurring in part and dissenting in part). The lost opportunity analysis would have precluded any recovery (beyond the attorneys' minimal salaries) under the facts of Blum, which involved salaried attorneys who charged their clients nothing for their work. Given the Supreme Court's holding in Blum, that cannot be right.

A second approach, adopted by the Courts of Appeals for the Seventh and Eleventh Circuits and, which we call the "micro-market rule," applies market rates but defines the market rather narrowly. In establishing the market for public interest legal work, rather than looking to the rates of attorneys of comparable skill and experience, these courts have looked to what public interest lawyers actually receive. Acknowledging that for-profit public interest law firms often depress their rates to accommodate plaintiffs who otherwise would be unable to sue, these courts nevertheless have looked to those prevailing rates, where available, to determine the market rate. The micro-market rate departs from actual billing rates when there is some indication that the billing rate is out of sync with the micro-market--in this case what the court believes the going rate for public interest work to be. See, e.g., Lightfoot v. Walker, 826 F.2d 516, 524 (7th Cir.1987) (rejecting an argument that the district court abused its discretion by failing to apply an attorney's private practice rate of $70 per hour as a "rate ceiling" for fee recovery, but nevertheless applying a tailored market rate for Title VII litigation); Mayson v. Pierce, 806 F.2d 1556 (11th Cir.1987) (per curiam) (reducing an attorney's normal billing rate to conform with market rate for Title VII work, over strong dissent by Judge Clark); Tomazzoli v. Sheedy, 804 F.2d 93, 98-99 (7th Cir.1986) (indicating a strong preference for relying on actual billing rates but holding the relevant market rate for counsel in a Sec. 1988 suit would be the going rate in the community for "civil rights matters"); see also Coulter v. Tennessee, 805 F.2d 146, 148-50 (6th Cir.1986) (awarding community rates for Title VII actions), cert. denied, --- U.S. ----, 107 S. Ct. 3186, 96 L. Ed. 2d 674 (1987).

At first blush, the micro-market approach seems sound. It seems to follow the mandate of Blum, relying on billing rates where they appear to reflect the market, and substituting a tailored market rate where billing rates are unavailable or are unreliable proxies for the rate attorneys could command in litigating public interest cases. Despite its obvious appeal, however, the micro market-rule is premised on a number of theoretical fallacies--the assumptions that: (1) an independent public interest market exists; (2) this market generates reasonable and fair fees; and (3) courts can rely upon this market in granting fees pursuant to fee shifting statutes. Unfortunately, we cannot accept these propositions.

As we see it, in the context of fee shifting, no such independent market exists; rather, the micro-market for public interest work is largely rooted in court-generated fee shifting awards. Cf. Norman v. Housing Authority, 836 F.2d 1292, 1300 (11th Cir.1988) ("The court recognizes that few practioners who regularly defend the poor and disadvantaged have the opportunity to bill and collect on an hourly basis. Accordingly, it may be virtually impossible to establish a prevailing market rate for such services."). Courts that try to establish public interest market rates by looking to the going rate for public interest work therefore do not examine an independently operating market governed by supply and demand, but rather recast fee awards made by previous courts into "market" rates. Courts adopting this micro-market approach, therefore, engage in a tautological, self-referential enterprise. They perpetuate a court-established rate as a "market" when that rate in fact bears no necessary relationship to the underlying purpose of relying on the marketplace: to calculate a reasonable fee sufficient to attract competent counsel.9

A third approach, which we have never seen discussed but which we find the most conceptually sound of all those advanced, involves a two-step process, using a modified billing rate. The first step would apply the firm's billing rate, even though it does not appear to reflect fully the true value of the firm's services. Under the second step, the lodestar derived from the firm's billing rate would be enhanced for contingency, to reflect the inherent risk of public interest cases that rely only on fee shifting statutes. See Pennsylvania v. Delaware Valley Citizens' Council for Clean Air, --- U.S. ----, 107 S. Ct. 3078, 97 L. Ed. 2d 585 (1987) (Delaware Valley II ) (discussing contingency multipliers). We find this solution so attractive because it seems to account realistically for the operation of a practice like Terris'. Under this theory, the contingency enhancement serves to compensate for the difference between Terris' artificially low billing rates and what would be required to attract competent counsel to such cases, that consideration being the rationale of Delaware Valley II, see infra at 1453. As we have explained above, Terris' billing rates, which it sometimes charges clients when it loses, are essentially contingency rates that are lower than it requires to be attracted to such cases, and premised upon the notion that when it wins, it will receive a higher rate from the courts. See supra at 1441-42. Conceptually, therefore, the contingency multiplier appears to be the ideal vehicle for determining the difference between Terris' putative billing rate and the rate required to attract such counsel.

Appealing as this modified billing rate seems, however, we ultimately reject it for two reasons. First, although the Court in Delaware Valley II did not address the question of a contingency multiplier in the context of a for-profit public interest law firm, it is clear that five justices are very wary of any enhancement beyond the lodestar. See infra at 1453-54. Even those justices who allowed contingency multipliers assumed that the lodestar generally already incorporates and accounts for the skill of the attorneys. See, e.g., Delaware Valley II 107 S. Ct. at 3091 (O'Connor, J., concurring) (observing that the lodestar rate ordinarily reflects the " 'special skill and experience of counsel' " (quoting Blum v. Stenson, 465 U.S. at 898, 104 S.Ct. at 1549)). Therefore, we hesitate to adopt a formula that relies so heavily on such enhancement.

Second, and more important, our unwillingness to allow a contingency enhancement in this case is prompted by our practical assessment that contingency calculations are cumbersome. Because the only notion that all of the justices agree upon is that contingency should be awarded when necessary to attract competent counsel, an approach based on a hybrid of billing rate and contingency would have to offer guidance on how a district court could make such a determination. As discussed infra at 1454-55, we believe that the questions of when a contingency multiplier is available and how much is deserved are very complicated and difficult to resolve. Given the general perception that such fee calculations already squander too much precious judicial attention, we determine that we could not adopt this modified billing rate approach absent some simple and workable means for ascertaining the correct contingency multiplier.

The last approach, which we call the "community market rate" rule, reads the Supreme Court's mandate in Blum as requiring courts to assess the experience and skill of the attorneys and compare their rates to those of comparable lawyers in the private business sphere. For example, in White v. City of Richmond, 713 F.2d 458 (9th Cir.1983), the court held that application of market rates without regard to public interest attorneys' regular billing rates was not an abuse of discretion. The court explained:

[W]e take judicial notice of the fact that many civil rights practitioners do not bill their clients at an hourly commercial rate. While evidence of counsel's customary hourly rate may be considered by the District Court, it is not an abuse of discretion in this type of case to use the reasonable community standard that was employed here.

Id. at 461. The Court of Appeals for the Ninth Circuit has recently reaffirmed its adherence to community market rates in Maldonado v. Lehman, 811 F.2d 1341, 1342 (9th Cir.1987), cert. denied, --- U.S. ----, 108 S. Ct. 480, 98 L. Ed. 2d 509 (1987).10

Although the district court in the case sub judice did not articulate its approach as such, it clearly adopted the community market rate rule. As discussed above, the district court relied on affidavits from partners in Washington, D.C. law firms who engage in complex federal litigation.

This approach has one significant drawback: in some cases, it may end up compensating public interest attorneys beyond the amount necessary to attract them to the case. The purpose of the fee shifting statutes is to provide reasonable fees, which at one point the Senate Report defined as "adequate to attract competent counsel, but which do not produce windfalls to attorneys." Senate Report at 6, U.S.Code Cong. & Admin.News 1976, p. 5913. See also Delaware Valley II, 107 S. Ct. at 3090-91 (O'Connor, J., concurring) (emphasizing that the essential focus of attorneys' fees awards must be on attracting adequate counsel); Delaware Valley I, 106 S. Ct. at 3098 ("[t]hese [fee shifting] statutes were not designed as a form of economic relief to improve the financial lot of attorneys, nor were they intended to replicate exactly the fee an attorney could earn through a private fee arrangement with his client").

In awarding local market rates for the community, the community market rate rule does not account for the possibility that such high rates may not always be necessary to attract competent counsel and that other factors such as the psychological benefit derived by attorneys who work for public interest firms may attract counsel to public interest litigation. Using the term "windfall" in its non-pejorative sense to connote receipt of a sum greater than expected, we acknowledge that for those attracted to public interest practice, the community market rate rule may, in some cases, constitute a windfall.

E. Our Choice: The Community Market Rate Rule

Although we are mindful that we confront a difficult question about which the circuits disagree, we nonetheless hold that the district court applied the correct standard in employing the community market rate rather than the actual billing rate of the Terris firm. We adopt the community market rule, after wrestling with the various approaches taken by the circuits, for three reasons. First, we believe that Blum requires market rates and that to a large extent Blum discounted concerns about potential windfalls. Second, we believe that the community market rule represents the best compromise among the conflicting policies behind the fee shifting statutes. Finally, and most important, the community market rule is the simplest, most workable rule (aside from the billing rate rule, which we believe contravenes the mandate of Blum ). We take these reasons up in turn.

1. Blum Counsels in Favor of Market Rates.

In Blum, the Court provided guidance as to how its rule mandating prevailing market rates should be applied. It noted that attorneys should provide courts with affidavits demonstrating "that the requested rates are in line with those prevailing in the community for similar services by lawyers of reasonably comparable skill, experience and reputation." Id. 465 U.S. at 896 n. 11, 104 S. Ct. at 1547 n. 11. Furthermore, the underlying facts of Blum itself illuminate what the Court meant by "market rates." In Blum the rate for work performed in 1978 by attorneys with no more than one and one half years experience in practice was $95-$105 per hour. As the Supreme Court noted, the petitioner in Blum did not contest the amount of the market rate but rather whether cost or market rate should apply. Apparently all sides agreed that market rate meant the rate for comparable services on work of similar complexity in the community.11

We must interpret the reach of Blum under the unusual facts of this case, which involves a for-profit public interest law firm. The Terris firm had no fee arrangement whatsoever with its clients in this case. Terris operates for profit and possesses a billing schedule, but, like the attorneys in Blum, the firm did not collect a fee from its clients for their service, and would have received no compensation at all if it had lost. Therefore, we are struck by the great similarity between the facts of this case and the facts of Blum.12

Moreover, even if the Terris firm had contracted for an artificially low fee with its clients in this case, the logic of Blum indicates that it could not be bound by that artificially low billing rate in securing attorneys fees through fee shifting statutes if it prevailed. Although we could base our holding on the facts of this particular arrangement wherein Terris had no billing arrangement whatsoever with SPIRG (as opposed to other cases where it does), we think that such an overly-narrow reading of Blum draws the wrong lesson. Blum focused on the fairness of awarding appropriate fees to competent counsel and the policy of attracting such counsel in the future. We therefore believe that the rule of Blum mandating the primacy of market rates may apply equally well where a for-profit public interest firm possesses a fee arrangement with its client.

Finally, to the extent that the "windfall" objection to our rule is valid, we note that the Supreme Court in Blum was not particularly receptive to that very argument. Comparing the "windfall" reaped by the non-profit salaried attorneys in Blum with the "windfall" received by the Terris firm, we observe that such excess above expected return was much higher in Blum than under the facts of this case. In Blum the Legal Aid Society of New York would have pursued the litigation anyway, and fee shifting, at least in the short run, was not strictly necessary to attract the attorneys to the case. In this case, by contrast, Terris is a for-profit endeavor that must cover its costs plus provide reasonable compensation to its partners and associates. Furthermore, as SPIRG asserts, the $60-$80 billing rate anticipates the possibility of higher rates under fee shifting statutes when the firm prevails.

2. The Underlying Policies of Fee Shifting Statutes

Although we acknowledge above that in some cases attorneys may receive more than the absolute minimum necessary to attract them, we believe that, on balance, the community market rule vindicates the underlying policies of the fee shifting statutes.

The community market rate rule will further the congressional policy of attracting competent counsel to public interest litigation and thereby "enable private parties to obtain legal help in seeking redress for injuries resulting from actual or threatened violation of specific federal laws." Delaware Valley I, 106 S. Ct. at 3098. Congress provided fee shifting to enhance enforcement of important civil rights, consumer-protection, and environmental policies. By providing competitive rates we assure that attorneys will take such cases, and hence increase the likelihood that the congressional policy of redressing public interest claims will be vindicated. See Senate Report at 6, U.S.Code Cong. & Admin.News 1976, p. 5913 ("If our civil rights laws are not to become mere hollow pronouncements which the average citizen cannot enforce, we must maintain the traditionally effective remedy of fee shifting in these cases.").13

Finally, such fee-shifting statutes will better deter illegal behavior by defendants if such defendants know that counsel with rates and skills comparable to their own attorneys' will challenge them for their misdeeds. See Berger, Court Awarded Attorneys' Fees: What is "Reasonable"?, 126 U.Pa.L.Rev. 281, 309 (1977) (discussing deterrence value of attorneys' fees awards).

Finally, our last and perhaps ultimately most persuasive reason for adopting the community market rate rule reflects our concern with how the rule we ultimately choose will affect the district courts. We are mindful of the Supreme Court's admonition in Hensley that we must avoid allowing litigation over fees to become "a second major litigation." 461 U.S. at 437, 103 S.Ct. at 1941. See also Shadis v. Beal, 692 F.2d 924, 928 n. 5 (3d Cir.1982) (noting the "danger that [fees] controversy will overshadow the underlying civil rights claims" and that courts should therefore seek to apply a "procedure that helps to place fee claims in their proper perspective"); Third Circuit Task Force on Court Awarded Attorney Fees, 108 F.R.D. 237, 246 (expressing concern that the lodestar approach to fees "increases the workload of an already overtaxed judicial system ") (emphasis in original)). Of the various approaches we consider, the easiest rules to apply are the billing rate rule and the community market rate rule. We have rejected the billing rate rule because it undervalues the time and the skills of attorneys who litigate public interest cases. The community market rate rule is almost as simple and inexpensive to apply. It requires only a number of representative affidavits from attorneys in the community who possess comparable qualifications and skill.14

We hold that under the facts of this case, involving a for-profit public interest law firm that has an artificially low billing rate, the community billing rate charged by attorneys of equivalent skill and experience performing work of similar complexity, rather than the firm's billing rate, is the appropriate hourly rate for computing the lodestar. We adopt this community market rule for three reasons. First, we believe that Blum counsels in favor of such a rule, particularly in this case where no fee arrangement existed at all. Second, the community market rule advances congressional policy of attracting competent counsel. Third, although we acknowledge the possibility that it may provide windfalls in certain cases, the community market rule provides the best compromise between achieving a reasonable fee and administering such fee grants easily.

We reject the billing rate rule as too rigid and at odds with the Supreme Court's holding in Blum. In addition we find that we cannot apply the micro-market rule because it does not account for the artificiality of the "market" in fee shifting cases. Finally, despite our attraction to the proposed modified billing rate rule, which uses the actual billing rate plus a contingency multiplier, we find that applying a contingency multiplier as part of attaining a reasonable fee in such cases would be at odds with the tenor of Delaware Valley II and would present major administrative problems.III. CONTINGENCY ENHANCEMENT

The purpose of the contingency multiplier is to compensate counsel for the riskiness of undertaking the litigation. In Delaware Valley II, the Supreme Court, in a 4-1-4 decision, upheld the availability of contingency multipliers15 but refused to apply one in the case before it. Four justices would have held that contingency multipliers are improper under any circumstances, stating that "in the absence of further legislative guidance, we conclude that multipliers or other enhancement of a reasonable lodestar fee to compensate for assuming the risk of loss is impermissible under the usual fee-shifting statutes." See 107 S. Ct. at 3087 (White, J., joined by Rehnquist, C.J., Powell, J., and Scalia, J.). Four justices would have permitted them. See 107 S. Ct. at 3091 (Blackmun, J., joined by Brennan, J., Marshall, J., and Stevens, J., dissenting). Justice O'Connor concurred with the White plurality on narrower grounds, recognizing, along with the dissenters, the possibility that contingency enhancement might be appropriate in some cases but denying it under the facts of Delaware Valley II. Because the four dissenters would allow contingency multipliers in all cases in which Justice O'Connor would allow them, her position commands a majority of the court. See Blum v. Witco Chemical Corp. (Witco), 829 F.2d 367, 379-82 (3d Cir.1987) (interpreting Justice O'Connor's opinion and offering the district court guidance on how to apply it).16

As we interpret the mandate of Delaware Valley II, one central theme elicits the agreement of the otherwise divided court. All of the justices seem to agree that the ultimate inquiry in establishing a reasonable fee must be the rate necessary to attract competent counsel, and that contingency multipliers should be awarded only when necessary to attract such counsel. See Delaware Valley II, 107 S. Ct. at 3079 (White, J. plurality opinion); (a multiplier is only available where "the applicant can establish that without an adjustment for risk the prevailing party 'would have faced substantial difficulties in finding counsel in the local or other relevant market.' " (O'Connor, J., concurring) (quoting Justice White's plurality opinion)); id. at 3091, 3092-93, 3094 (Blackmun, J., dissenting). See Catlett v. Missouri Highway and Transportation Commission, 828 F.2d 1260, 1271 (8th Cir.1987) (interpreting Delaware Valley II and explaining that contingency multipliers are only available when the lodestar does not reflect the success achieved through litigation and enhancement is necessary to attract competent local counsel).

Five Justices (Justice O'Connor and the dissenters) agreed that contingency may be available where there is a risk of non-payment. Although the point is not free from doubt, we believe that because Justice O'Connor opposed delving into the individualized risk of the case, she would agree with Justice Blackmun that "it is the fact of contingency, not the likelihood of success in any particular case, that mandates an increase in an attorney's fee...." Id. 107 S.Ct. at 3098 (Blackmun, J., dissenting); see Norman v. Housing Authority, 836 F.2d 1292, 1302 (11th Cir.1988) ("[f]ive justices in Delaware Valley Citizens' Council II indicated that in the rare case enhancement may be appropriate if there is a risk of nonrecovery of a fee in the case").

We note however, that Justice O'Connor emphasized that she would allow such enhancement only in rare cases. Courts interpreting Delaware Valley II have uniformly found that Justice O'Connor's opinion indicates that contingency multipliers should be granted only rarely. See Coup v. Heckler, 834 F.2d 313, 324 (3d Cir.1987) (observing in dictum that Delaware Valley II "severely limits the occasions in which contingency enhancement is appropriate in fee-shifting cases"); see also Thompson v. Kennickell, 836 F.2d 616, 621 (D.C.Cir.1988) (describing Justice O'Connor's standard as "stringent"); Witco, 829 F.2d at 380 ("[a]t least five justices have indicated through the tenor of their opinions that the plaintiff has a significant burden to carry to justify a contingency multiplier").

Although SPIRG's case might appear, at first blush, to qualify as a type of case eligible for contingency multiplier because Terris did face the risk of non-payment, we nevertheless conclude that it is not among those rare cases that deserves such enhancement. Under any general notion of assessing risk, as well as the previous standard,17 SPIRG was not a risky case. As the district court found, the risk of losing this particular Clean Water Act claim against Bell was not particularly great. J.A. at 29. SPIRG investigated the files of the Environmental Protection Agency, discovered the names of industries in violation of the Clean Water Act and thereupon developed its case. See Jenkins v. Missouri, 838 F.2d 260, 263 (8th Cir.1988) (affirming district court rejection of enhancement because of "likelihood of success" and the "small risk of non-payment was fully reflected in the reasonable attorneys' fee"); Jacobs v. Mancuso, 825 F.2d 559, 561 (1st Cir.1987) (post-Delaware Valley II case in which contingency was rejected because the case was inherently not risky).

Finally, we note that determining and applying the contingency multiplier would present vast administrative problems. To apply Justice O'Connor's standard for a contingency fee, we would have to remand this case to the district court for inquiry into whether, given the lodestar calculation, a contingency multiplier is necessary to attract competent counsel to this class of cases. This determination might be difficult to make absent extensive market information. Furthermore, Justice O'Connor opined that for a contingency enhancement to be available, it must account for the "difference in market treatment of contingent fee cases as a class." Delaware Valley II, 107 S. Ct. at 3089. By this, we believe that Justice O'Connor meant that courts should calculate how the relevant community compensates for contingency and then apply that multiplier in every case.18 Therefore, the district court would have to calculate how the market accounts for contingency cases as a class. Finally, the district court would have to determine the appropriate factor by which to multiply Terris' billing rate in order to reflect the risk. Such extensive inquiry, however, would undermine our community market approach toward the lodestar which emphasized, as one of its most notable strong points, its simplicity.

In the final analysis, however, we rely on the one clear message that we can derive from Delaware Valley II: contingency multipliers are only available to the extent they are necessary to attract competent counsel. Because we feel that in granting the community market rate we have assured that the fee is sufficient to attract competent counsel, we will not disturb the district court's holding that no contingency enhancement was necessary in this case.

In its cross appeal, SPIRG argues that the district court abused its discretion in refusing to increase the lodestar to reflect the quality of Terris' representation. Although it concedes that the district court did not err in stating the legal standard, SPIRG argues that the district court abused its discretion by failing "properly [to] identify the criteria used for such determination." Silberman v. Bogle, 683 F.2d at 65. In Delaware Valley I, the Supreme Court addressed the question of a quality multiplier and held that such an enhancement is available only in very rare circumstances where the attorney's work is so superior and outstanding that it far exceeds the expectations of clients and normal levels of competence. 106 S. Ct. at 3099. Indeed, three Justices observed that the Court's standards "heighten[ed] the showing required to the point where it may be virtually impossible for a plaintiff to meet." Id. at 3100 (Blackmun, J., concurring in part and dissenting in part).

The record clearly indicates that the district court carefully considered the quality of Terris' work. In refusing to grant an enhancement for quality the district court stated:

The work done by both side's [sic] attorneys in this case was excellent. The Supreme Court has, however, recently reiterated that upward adjustments based on the quality of representation are rarely justified because of the strong presumption that the lodestar is a reasonable fee.

J.A. at 29 (citing Delaware Valley I and Blum ). The court explained that despite its excellent quality, the representation did not approach the high threshold necessary to meet the standard of Delaware Valley I. We agree. Because the Supreme Court has indicated that a quality multiplier should be granted only in rare and exceptional circumstances, we cannot say the district court abused its discretion in opting for brevity in its denial of this extraordinary alteration of the community market rate lodestar.

SPIRG notes that this litigation commenced in March 1984 and that at least three years will have passed before Terris receives any fee. It contends that the district court abused its discretion in denying Terris any increase for delay of payment. SPIRG submits that because the district court did not discuss the question of delay at all, it abused its discretion by failing "properly [to] identify the criteria used for such determination." Silberman v. Bogle, 683 F.2d at 65.

In Delaware Valley II, 107 S. Ct. at 3081-82, five justices held that compensation for delay constitutes a separate inquiry from the question of contingency. (White, J. plurality opinion; O'Connor J., concurring in that part of the decision.) In Library of Congress v. Shaw, 478 U.S. 310, 106 S. Ct. 2957, 92 L. Ed. 2d 250 (1986), the Court explained that a fee enhancement to compensate for delay is essentially equivalent to an award of interest. Id. 106 S.Ct. at 2965-66.

This court has recognized that delay constitutes an appropriate reason for enhancing the lodestar. E.g., Black Grievance Committee, 802 F.2d at 655-56; In re Fine Paper, 751 F.2d at 588. We have explained the policy behind compensating plaintiffs for delay: " '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'...." Institutionalized Juveniles v. Secretary of Public Welfare, 758 F.2d 897, 922 (3d Cir.1985) (quoting Copeland v. Marshall, 641 F.2d 880, 893 (D.C.Cir.1980) (en banc)). Delay compensation does not depend on the type of case or the quality of representation but rather on economic factors encompassing the cost to the plaintiff's law firm of waiting for its fee. See Black Grievance Committee, 802 F.2d at 656.

In announcing the important policy of recompensing plaintiffs for delay we have also, however, indicated the crucial importance of making an adequate showing in the district court. In Institutionalized Juveniles, we emphasized the need for "carefully developed evidence of the costs to plaintiffs of receiving the delayed payment for services." 758 F.2d at 923. We placed "the burden on plaintiffs to document the need for a [delay] multiplier." Id. at 923-24. We itemized the type of showing necessary including prevailing interest rates and market rates for attorneys' services. Finally, we rejected plaintiffs arguments that the multiplier for delay that it received was insufficient because the "plaintiffs presented to the district court only very primitive evidence of the costs they incurred because of the delay in payment." Id. at 923.

Bell claims, and SPIRG does not contest, that SPIRG's entire argument to the district court on delay consisted of the following:

As to the third factor, there has been a considerable delay in the receipt of payment for services rendered. If plaintiffs were fee-paying clients, counsel would have received payment for their services every month. However, since plaintiffs are not paying clients, counsel have foregone payment for their services and even reimbursement for their out-of-pocket expenses for approximately three years.

Reply Brief for Appellant/Cross-Appellee (Bell) at 27 n. *.

SPIRG has failed to meet its burden of demonstrating that its showing in the district court was sufficient to meet the test of Institutionalized Juveniles. Although recognizing that the district court did not discuss delay, we find that this failure was not an abuse of discretion where the plaintiff's showing was so minimal. Contrary to Institutionalized Juveniles, which was filed a full year before this case was litigated, no specific showings were made. We therefore hold that SPIRG waived its ability to pursue delay damages. See Jones v. Central Soya Co., 748 F.2d 586 (11th Cir.1984) (rejecting claim for delay damages where plaintiff did not adequately raise the question before the district court); cf. Delaware Valley I, 106 S. Ct. at 3092 n. 2 (noting with approval that district court eliminated hours that were not documented with sufficient detail); Blum v. Stenson, 465 U.S. at 892 n. 5, 104 S. Ct. at 1545 n. 5 (petitioner waived right to challenge reasonableness of prevailing market rate where it failed to submit any evidence challenging the calculation in the district court).

VI. FAILURE TO ALLOCATE HOURS TO THE OTHER TERRIS CASES

Bell submits that SPIRG neglected to distribute the costs of litigating the fees issues to the twenty-five other cases it litigated contemporaneously that raised identical questions concerning fees. The district court did not make factual findings on this issue. Bell argues that the district court's failure to require such a distribution constituted an abuse of discretion.

SPIRG makes two arguments in response. First, it claims that Bell never raised the issue in the district court and hence is barred from raising it for the first time here. We are convinced that Bell, albeit obscurely, raised this issue in the district court and therefore we must address it. Second, SPIRG represents that this case was the first of twenty-six similar Clean Water Act claims and asserts that, although fees for work on the underlying Clean Water Act claims could be distributed, the time spent on the fee issues could not be distributed because Terris could not rightfully assess fee shifting research to cases in which it had not yet prevailed. SPIRG argues that to require Terris to redistribute hours researched solely on SPIRG's behalf to other accounts of the Terris firm runs contrary to ordinary attorney practice:

While it is true that the brief may become useful in future cases if similar issues are involved, it is as a practical matter impossible to anticipate in which cases which issues might arise.... We know of no lawyers who ... go back to the first client and refund some of its money because the expertise was later useful to him in some other case.

Brief for Appellees/Cross-Appellants (SPIRG) at 31-32.

In Prandini v. National Tea Co., 557 F.2d 1015 (3d Cir.1977), questioned on other grounds, Ashley v. Atlantic Richfield Co., 794 F.2d 128 (3d Cir.1986), we held that where plaintiff sued two separate defendants for "nearly identical" claims, "double payment for the same effort should be avoided by some apportionment of the fee between the two cases." 557 F.2d at 1019 n. 3. We agree with SPIRG that it need not apportion its claims between a first case, not litigated contemporaneously with other potentially similar cases, and those later cases where the issues in the first case are not guaranteed to arise. Even outside the fee shifting area, subsequent clients often reap the benefits of the firm's experience. We must nonetheless remand this question to the district court for it to make factual findings concerning: (1) the timing of the other twenty-five cases; (2) the overlap between those cases and this case; and (3) whether Terris in litigating the fees issue had already prevailed in those other actions and therefore should have anticipated employing the fee litigation research of this case in litigating other fees cases.

VII. REDUCTION IN FEES LODESTAR FOR FAILURE TO ACHIEVE

The Supreme Court in Hensley v. Eckerhart, 461 U.S. 424, 435, 103 S. Ct. 1933, 1940, 76 L. Ed. 2d 40 (1983), held that "[w]here a plaintiff has obtained excellent results, his attorney should recover a fully compensatory fee," and that "the fee award should not be reduced simply because the plaintiff failed to prevail on every contention raised in the lawsuit." See City of Riverside v. Rivera, 477 U.S. 561, 106 S. Ct. 2686, 2692, 91 L. Ed. 2d 466 (1986) (quoting Hensley ); Cooper v. Dyke, 814 F.2d 941, 950 (4th Cir.1987) ("[f]ull compensation for hours expended does not require plaintiff to prevail on every claim ... where all claims are closely related because based on common facts or related legal theories, a court is not obligated to determine which hours were spent on which claims"). On the other hand, "where the plaintiff achieved only limited success, the district court should award only that amount of fees that is reasonable in relation to the results obtained." Hensley, 461 U.S. at 440, 103 S.Ct. at 1943. The Court noted that there is no "precise rule or formula for making these determinations," but that the district court must exercise its discretion in reducing the fee downward. Id. at 436-37, 103 S.Ct. at 1941. In Institutionalized Juveniles, we explained that the Hensley decision applies not only to litigation on the merits but also to the degree of success achieved in litigation of the fee award. See 758 F.2d at 918-21. We held that where fee applicants do not fully succeed in recovering their fees, the fee award must be reduced to reflect incomplete success on the fee award.

Bell contends that the district court abused its discretion because it failed to reduce SPIRG's fee award to reflect that it had not succeeded in all of its claims for fees. More specifically, Bell argues that because the district court rejected SPIRG's multiplier requests, it was duty bound to reduce the lodestar for the fee litigation.

SPIRG rejoins with two arguments, the second of which has some merit. First, SPIRG argues that only a small amount of its total hours spent on the fee litigation concerned the multiplier issues. The district court, made no findings on the apportionment question, however, and we would be exceeding our role if we made our own judgments without the benefit of such factual findings.

Second, SPIRG argues that it was essentially successful on the fee issues, and that its arguments on multipliers were alternative arguments in case this court decided to reduce the market rate employed by the district court in determining the lodestar. Although this argument has some force, we nevertheless note that it is not clear that SPIRG advanced the multiplier issues before this court or before the district court as alternative arguments. Arguably, SPIRG wanted to eat its cake (the community market rate) plus have it (the multipliers). Nor is delay an alternative argument. Because the district court made no findings on Hensley restriction, we believe the wisest disposition is to let it determine the question. On remand the district court will consider the relatedness of the lodestar and multiplier questions and, if it deems appropriate, it will make findings about the number of hours to be reduced.

We conclude that the district court committed no clear error in its assessment of market rate for attorneys of comparable experience in the same city for equivalent legal services. Furthermore we hold that the district court applied the correct legal standard in calculating the lodestar for the fee shifting statute of the Clean Water Act based on market rates for comparable quality private-interest law firms rather than the actual billing rates of the for-profit public interest law firm representing the plaintiff.

Additionally, we hold that the district court did not abuse its discretion by failing to award plaintiffs an enhancement to the lodestar for contingency, quality, and delay. We must remand for further factual inquiries on the questions whether the fee award should be diminished for (1) failure to allocate costs to other cases raising similar issues argued by the same firm, and (2) failure to achieve complete success on the fees issues.

We therefore will affirm in part, reverse in part, and remand for further factual proceedings consistent with this opinion.