concurring in part and dissenting in part:
Although I agree with my colleagues that a quality enhancement was appropriate in this case, I concur separately to explain my understanding of the limited nature of our holding. I dissent, however, from the majority’s approval of the contingency enhancement because it rests on a misinterpretation of Delaware Valley II that greatly expands its intended scope.
I. Applicable Legal Standards
The majority characterizes the question we must decide as “whether the district court abused its discretion in awarding an enhancement of attorney’s fees_” Majority Op. at 331. An “abuse of discretion” standard, however, applies only to *340appellate review of a trial court’s award of attorneys’ fees in cases in which its legal authority to set them is not disputed. This case, by contrast, requires us to determine whether the district court correctly interpreted legal standards developed by the Supreme Court — questions of law subject to de novo review.
Federal statutory provisions requiring the losing litigant to pay “reasonable” attorneys’ fees to the prevailing party’s counsel promote Congress’ goal of ensuring that private citizens can obtain competent legal help to vindicate their statutory rights. See Delaware Valley I, 478 U.S. 546, 562, 565, 106 S.Ct. 3088, 3096, 3098, 92 L.Ed.2d 439 (1986). The lodestar figure (hours worked multiplied by a reasonable hourly rate) presumptively represents a reasonable fee under the fee-shifting statutes, and a petitioner therefore bears a heavy burden in justifying an enhancement above the lodestar. See id. at 565, 106 S.Ct. at 3098.
The Supreme Court has consistently viewed fee multipliers with considerable suspicion. Delaware Valley I established that enhancements for quality of representation should be granted only in “rare” and “exceptional” cases. See id. Delaware Valley II permitted courts to award premiums for contingency risks, but only in certain limited instances where a petitioner has proven, inter alia, that the prevailing party would have faced “substantial difficulties” finding competent counsel without an enhancement. See 483 U.S. 711, 107 S.Ct. 3078, 3089-91, 97 L.Ed.2d 585 (1987) (O’Connor, J., concurring).
Application of Delaware Valley II compels reversal of the district court’s decision to grant the plaintiffs a contingency enhancement. On the other hand, the trial court’s factual findings support its conclusion that this case presents the exceptional circumstances necessary to justify an award of a quality enhancement under Delaware Valley I.
II. Enhancement for Contingent Fee Cases
Although Delaware Valley II provides courts with somewhat uncertain guidance in determining when contingency enhancements are permissible, I conclude that the district court’s decision rests on an incorrect interpretation of Justice O’Connor’s concurring opinion, which provides the controlling legal standard.
Justice O’Connor ruled that a court may incorporate a contingency risk premium in the award of reasonable attorneys’ fees only if the fee applicant carries the burden of proof on two elements. First, the petitioner must “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.’ ” Id. at 3091 (quoting plurality op. at 3089). See also id. at 3091 (evidence must indicate that enhancement was “necessary to attract competent counsel in the relevant community”). Second, the applicant must show “the degree to which the relevant market compensates for contingency.” Id. at 3090 (emphasis added). Any enhancement must be based on an objective assessment of the “market treatment of contingent fee cases as a class, rather than on ... the ‘riskiness’ of any particular case” (in other words, the “legal” risks that should already be reflected in the lodestar). Id. at 3089.
Our circuit has interpreted Justice O’Connor’s concurrence as setting “limited” and “stringent” standards for contingency multipliers. Thompson v. Kennickell, 836 F.2d 616, 621 (D.C.Cir.1988). See also Weisberg v. Dep’t of Justice, 848 F.2d 1265, 1272 (D.C.Cir.1988) (“contingency enhancements, while typically unavailable, may on occasion be appropriate”). Indeed, this case represents the first time a majority of this court has concluded that fee petitioners have met the strict standards governing contingency enhancements. The two prongs of the Delaware Valley II test will be discussed in turn.
A. “Substantial Difficulties” in Obtaining Counsel
Justice O’Connor joined the plurality in requiring proof that the prevailing party “would have faced substantial difficulties” *341in obtaining competent counsel “in the relevant market,” absent an upward fee adjustment for contingency risks. Delaware Valley II, 107 S.Ct. at 8091. A majority of the Court clearly mandated a judicial determination of the availability of counsel willing to represent the prevailing party without expectation of a contingency premium as of the time he sought representation. My colleagues acknowledge the latter point, holding that the appropriate time frame is the date plaintiffs initiated the litigation (i.e., 1973). Majority Op. at 337. We all agree, then, that the district court erred in reading Delaware Valley II as instructing courts to focus on the current market of available lawyers in determining “substantial difficulty.” See 684 F.Supp. at 1102-03.
Unfortunately, the majority ignores Delaware Valley II’s requirement that a fee applicant prove that the particular prevailing party would have faced substantial difficulty in finding competent counsel willing to accept his case on a straight lodestar fee basis, even as it acknowledges that the McKenzie plaintiffs did not in fact encounter such difficulty. Majority Op. at 338. It is not enough, as the majority contends, for the fee petitioner simply to establish that the case at issue belonged to a category of cases that lawyers ordinarily would not have taken without the assurance of a contingent fee premium. Id. at 337-38.
Precedent clearly requires an individualized approach. For example, both the plurality and Justice O’Connor repeatedly referred to the burden of proof borne by “the fee applicant,” not to a category of claimants. Furthermore, we implicitly rejected a “class of case” approach in Thompson, 836 F.2d at 621 (finding evidence of “plaintiff class’ difficulty in obtaining counsel” and of “the different court treatment of contingency cases” insufficient to justify enhancement).
Because the record establishes that McKenzie and his fellow plaintiffs had in fact located qualified lawyers willing to represent them in Washington, D.C. in the early 1970’s, I conclude that under Delaware Valley II these fee applicants have failed to prove that the prevailing party “would have faced substantial difficulties” in securing competent attorneys absent the incentive of an enhanced fee. Indeed, the district court commended counsel for “candidly conceding] that they did not explicitly consider the possibility of an enhancement as a distinct component of a reasonable fee_” 684 F.Supp. at 1104.
The trial court determined, however, that the attorneys could not have contemplated such an enhancement “because the terminology and methodology of the lodestar, and multiplier were not even conceived until ... after this suit was commenced.” Id. The court ultimately concluded that a “reasonable” fee should include a premium of fifty percent above the normal hourly rate. Id. at 1105. This argument simply assumes its conclusion. Evidence that class counsel expected a “reasonable” fee does not prove that a reasonable fee included a contingency enhancement above the lodestar figure (or, in 1973 terminology, a premium for risk of nonpayment). Although one affidavit declared that such premiums have always been required in civil rights cases, the district court relied not on this statement, but instead on plaintiffs’ irrelevant evidence that there currently exists a shortage of Washington lawyers willing to accept employment discrimination cases on a contingent fee basis. Id. at 1103.
In sum, I would overrule the district court because it wrongly interpreted Delaware Valley II. The judicial inquiry may not center on whether the prevailing party’s claim placed him in a class whose members would face problems locating a lawyer willing to take their case on a non-contingent fee basis in today’s market. The majority nevertheless rationalizes this departure from Delaware Valley II by asserting that application of its standards “would have perverse effects in practice,” maj. op. at 337, and offers three justifications— none of which withstands scrutiny. First, the majority’s observation that “it would penalize plaintiffs who were lucky enough to stumble across” a non-contingent fee lawyer, id., is irrelevant. The fee shifting statutes are designed to compensate *342prevailing parties for legal fees reasonably incurred. It is the taxpayer who should not be penalized through the award of fees that are higher than the particular attorneys should reasonably expect.
Second, the majority declares that an “actual difficulty” standard “would discourage referral services such as the Lawyers Committee.” Id. Such organizations, however, presumably exist to match needy clients with competent lawyers without reference to the possible implications of the resulting relationships in later fee shifting cases.
Third, the majority asserts that the “substantial difficulty” criterion “would lead to a charade” in which all lawyers automatically assert their disinclination to accept a case except on a contingent fee basis. Majority Op. at 337, 338. We have no authority, however, to disregard a clear Supreme Court directive merely because we believe it might prove to have some unanticipated effects. Even if it were appropriate for us to act upon our speculations about the prospective consequences of the Supreme Court’s decision in Delaware Valley II, such concerns have no bearing on this case, which involves the terms of legal representation arranged many years before Delaware Valley II was issued.
Like it or not, the proper application of the Court majority’s test in Delaware Valley II compels the conclusion that an enhancement was not “necessary to attract competent counsel” in this case, because the evidence demonstrates that these plaintiffs immediately found qualified legal representation.
B. Relevant Market
Justice O’Connor’s second criterion requires that a fee petitioner must prove “the degree to which the relevant market compensates for contingency,” Delaware Valley II, 107 S.Ct. at 3090, based on the particular legal community’s differing “treatment of contingent fee cases as a class.” Id. at 3089. These passages might be construed as dividing the legal market into just two categories — non-contingent fee and contingent fee cases. Other portions of the concurrence, however, suggest that Justice O’Connor did not intend to include, in the “contingent fee” category, private sector cases in which fees are awarded as a percentage of damages (e.g., personal injury suits). For example, she declares that “[i]n most fee-shifting cases, ... the private model of contingency compensation will provide very little guidance.” Id. at 3090. I find the majority’s formulation (“all contingency claims in the District of Columbia”) too broad. Majority Op. at 334. Delaware Valley II requires that we target the relevant market more precisely.
C. Public Interest Law Firms
The plurality in Delaware Valley II noted amici’s arguments against the award of contingency enhancements to nonprofit firms, 107 S.Ct. at 3087, but left the question open. Id. at 3088 n. 10. The dissent, in contrast, declared that nonprofit legal organizations should receive the same fees —and enhancements — as private firms. Id. at 3090. Justice O’Connor, however, did not address the matter.
In awarding contingency enhancements to the Institute for Public Interest Representation and the Washington Lawyers Committee, the district court and the majority rely on cases establishing that pro bono lawyers must be awarded the same fees as private practitioners. See McKenzie, 684 F.Supp. at 1104-05, citing Blum v. Stenson, 465 U.S. 886, 894, 104 S.Ct. 1541, 1546, 79 L.Ed.2d 891 (1984), and Copeland v. Marshall, 641 F.2d 880, 889 (D.C.Cir.1980) (en banc). See also maj. op. at 333-34, citing, e.g., Save Our Cumberland Mountains v. Hodel, 857 F.2d 1516, 1521-24 (D.C.Cir.1988) (en banc). See also id. at 338.
Admittedly, if we were to conclude that a public interest organization was entitled to a risk enhancement, then the cases cited by the majority would require that the amount of the enhancement match that found appropriate for private law firms. But such an analysis begs the threshold question under Delaware Valley II: whether, in a particular case, plaintiffs would have had “substantial difficulties” finding competent *343pro bono attorneys without an enhancement. The eases cited by the majority do not require a finding that if a private firm proves its entitlement to compensation on a contingency basis under Delaware Valley II, then public interest lawyers must also be awarded a contingency premium even if it is established that pro bono firms will represent Title VII clients on a lodestar fee basis.
Neither the district court nor the majority cite anything in the record to suggest that McKenzie or similarly situated parties would not have been able to secure competent public interest representation absent this incentive. Although a few affiants speculated that certain public interest organizations may become increasingly reluctant to accept Title VII class actions against the government without the guarantee of risk-enhanced compensation, 684 F.Supp. at 1105 and n. 28, the evidence overwhelmingly indicates that public interest law firms in Washington neither demand nor expect a contingent fee premium as a condition for accepting Title VII cases. Accordingly, I would reverse the contingency enhancement granted to counsel for the Lawyers Committee and the Institute.
D. Policy Considerations
Delaware Valley II provides a narrow exception to the Supreme Court’s general rule against enhancements by allowing courts to award contingency fee premiums in certain limited circumstances. My colleagues nevertheless appear to justify their departure from its strict requirements by focusing on what they perceive to be the Court’s overriding objective in attorneys’ fees cases — establishing fixed rules that will discourage lengthy secondary litigation. Majority Op. at 335, 338. Specifically, the majority abandons consideration of the test of whether a particular party would have had difficulty in securing representation on a non-contingent fee basis. The burden on fee applicants is thereby reduced to showing merely that the prevailing party’s case belongs to a class of cases in which parties experience significant difficulty in securing competent representation except on a contingency enhancement basis. While this approach may be sensible and may better reflect the reality of the legal marketplace than Delaware Valley II, we are not at liberty to discard criteria with which we disagree. Because the majority’s holding departs from Justice O’Con-nor’s two-part test, I cannot join it.
III. Enhancement for Quality of Representation
Delaware Valley I, 478 U.S. 546, 106 S.Ct. 3088, 92 L.Ed.2d 439 (1986), held that the lower courts had erred in increasing attorneys’ fees based on “superior quality” of performance. The Court noted the “strong presumption” that the lodestar figure already reflects the factors constituting a “reasonable” attorney’s fee, such as quality of representation, counsel’s special skill and experience, the novelty and complexity of the issues, and results obtained. Id. at 564-65, 106 S.Ct. at 3096-98, citing Blum v. Stenson, 465 U.S. 886, 898-900, 104 S.Ct. 1541, 1548-49, 79 L.Ed.2d 891 (1984). The Court therefore ruled that fees should not be awarded above the lodestar amount except in certain “rare” and “exceptional” cases where the enhancement is supported by a fee petitioner’s “specific evidence” on the record and by a trial court’s “detailed findings.” 478 U.S. at 565, 106 S.Ct. at 3098, citing Blum, 465 U.S. at 898-901, 104 S.Ct. at 1548-50. In the case before it, the Court found a “quality” enhancement to be unnecessary to serve the statutory goal, which was to enable private parties to obtain competent legal assistance — not to replicate the fee an attorney might earn in a private case. Id. at 565, 106 S.Ct. at 3098.
Applying Delaware Valley I, we recently reversed the district court’s award of a twenty-five percent fee multiplier for “exceptional results obtained” on the ground that neither the fee applicant’s evidence nor the court’s findings contained the specific justification required to overcome the presumption that the lodestar figure provided full compensation. Thompson v. Kennickell, 836 F.2d 616, 621-23 (D.C.Cir.1988).
*344The instant case is similar to Thompson in that the district court granted a twenty-five percent enhancement for quality of representation based, inter alia, on the attorneys’ “remarkable success” in obtaining “exceptional results.” 684 F.Supp. at 1105-06. The government argues, in this case, that Delaware Valley I, 478 U.S. at 565-66, 106 S.Ct. at 3098-99, barred the trial court from examining litigation results because they are already reflected in the lodestar.
I believe, however, that a careful reading of Delaware Valley I supports the district court’s (684 F.Supp. at 1105-06) and the majority’s (Majority Op. at 339) treatment of outstanding results as one element to be considered in determining whether a quality enhancement should be awarded. See, e.g., 478 U.S. at 567-68, 106 S.Ct. at 3099-3100 (requiring evidence showing why results were so outstanding, and why lodestar figure fell below awards made in similar cases where court found equally superior quality of performance). See also Thompson, 836 F.2d at 622 (results are one aspect of quality of representation).
Moreover, the district court did not rely solely on results in awarding the enhancement, see Majority Op. at 338-339, but found rather that the fee applicants had presented “specific evidence” demonstrating that their superior quality of legal service was not reflected in the lodestar. The court identified two unusual factors. First, certain attorneys had billed at junior associate hourly rates even though their performance warranted the higher fees commanded by more experienced lawyers. 684 F.Supp. at 1107. A trial judge’s finding that the younger attorneys’ performance exceeded the expected level of quality is entitled to deference. Second, the court noted that the continuous service provided by two key lawyers promoted unusually efficient representation. The trial judge was also able to draw on his own extensive experience in order to make the additional finding that the plaintiffs’ legal representation was of a significantly higher quality than that he normally encountered in comparable litigation. Id.
I concur with the majority because I conclude that these findings suffice to demonstrate the “rare” and “exceptional” circumstances necessary to justify an enhancement. I am troubled, however, by certain of the district court’s justifications — for example, its reliance on marginally relevant evidence such as counsel’s self-serving declarations that they displayed exceptional skill. What is more important, the trial court was plainly wrong in considering unbilled hours spent in negotiations as forming a basis for a quality enhancement, see id. at 1106-07; the time devoted to such discussions should have been billed and reflected in the lodestar. Because quality enhancements are appropriate only in rare cases, courts must be very careful in specifying the findings that support such awards.
In sum, the instant case presents what is clearly one of those exceptional instances where fee petitioners have met their heavy burden of rebutting the presumption that the lodestar adequately reflected the quality of their representation. As I understand that the majority does not endorse the routine granting of quality enhancements, I concur.
IV. Conclusion
For the reasons discussed above, I agree with the majority that the district court’s enhancement for quality of representation award was appropriate. As the district court’s fifty percent contingency multiplier was based on a misapplication of Delaware Valley II, however, I dissent from the majority’s affirmance on this issue.