These appeals require us to revisit the war zone where two groups of plaintiffs' lawyers have struggled over the proposed allocation of roughly $68,000,000 in attorneys' fees. One camp, dissatisfied with the district court's latest formula for distributing the fees, attacks the court's order on three fronts. The disgruntled lawyers contend that the district court (1) violated their due process rights, (2) used an improper method to determine the awards, and (3) divided the available monies in an arbitrary and unreasonable manner. We find appellants' first two plaints to be without merit, but we agree with them that allocating 70% of the fees to the appellees constituted an abuse of the trial court's discretion. And, because we are reluctant to prolong a matter that, like the proverbial cat, seems to have nine lives, we take matters into our own hands and reconfigure the fee awards.

The lay of the land is familiar. We explored much the same terrain in an earlier encounter, see In re Nineteen Appeals Arising Out of San Juan Dupont Plaza Hotel Fire Litig., 982 F.2d 603 (1st Cir.1992), and a plethora of opinions describing the details of the underlying litigation pockmark the pages of the Federal Reports, see, e.g., id. at 605 n. 1 (offering partial listing). Thus, a brief overview of the litigation will suffice.

In 1987, the Judicial Panel on Multidistrict Litigation consolidated over 270 cases arising out of the calamitous conflagration that had ravaged the San Juan Dupont Plaza Hotel on the evening of December 31, 1986. See In re Fire Disaster at Dupont Plaza Hotel, 660 F. Supp. 982 (J.P.M.L.1987) (per curiam). The designated trial judge, Hon. Raymond L. Acosta, handpicked certain attorneys, denominated collectively as the Plaintiffs' Steering Committee (PSC), to act as lead and liaison counsel for the plaintiffs. In Nineteen Appeals, we summarized the roles played by the PSC and the individually retained plaintiffs' attorneys (IRPAs), respectively:

The PSC members looked after the big picture: mapping the overarching discovery, trial, and settlement strategies and coordinating the implementation of those strategies. The IRPAs handled individual client communication and other case-specific tasks such as answering interrogatories addressed to particular plaintiffs, preparing and attending the depositions of their clients, and taking depositions which bore on damages. The IRPAs also worked with Judge Bechtle [the "settlement judge"] on a case-by-case basis in his efforts to identify and/or negotiate appropriate settlement values for individual claims. When Judge Acosta determined that the plaintiffs should try twelve representative claims as a means of facilitating settlement, a collaborative composed of three PSC members and four IRPAs bent their backs to the task.

Nineteen Appeals, 982 F.2d at 605.

The combined efforts of all concerned generated a settlement fund approximating $220,000,000. The district court computed the payments due under the various contingent fee agreements, deducted the total (roughly $68,000,000) from the overall settlement proceeds, and placed that sum in an attorneys' fee fund (the Fund).1 In his initial attempt to disburse the Fund, Judge Acosta used an enhanced lodestar to compute the PSC's fees, and awarded some $36,000,000 (52% of the Fund) to PSC members in their capacity as such, leaving the balance to be distributed among the IRPAs. A group of lawyers (mostly, but not exclusively, "non-PSC" IRPAs)2 succeeded in vacating this award on the ground that the proceedings were procedurally flawed. See id. at 610-16.

The victory proved to be illusory. On remand, the district court abandoned the lodestar approach, adopted the percentage of the fund (POF) method, and recalculated the fees based on what it termed "the relative significance of the labor expended by the IRPAs and PSC members in instituting, advancing, or augmenting the plaintiffs' settlement fund." Using this methodology, the court awarded 70% of the Fund to PSC members in their capacity as such, thereby increasing their share of the fees by some $11,000,000, while simultaneously reducing the IRPAs' share of the Fund by the same amount. These appeals ensued.

In a virtual echo of the claims advanced in Nineteen Appeals, appellants (all of whom are IRPAs) characterize the proceedings by which the district court determined the allocation of the Fund as unfair. Specifically, appellants assert that the revamped procedural framework violated their rights to due process, and that, in all events, the court abused its discretion in erecting the framework. We consider these assertions in sequence.

A. Due Process.

In Nineteen Appeals, 982 F.2d at 610-16, we discussed the due process considerations implicated in the fee-setting aspect of this litigation. We again use the triangular construct of Mathews v. Eldridge, 424 U.S. 319, 96 S. Ct. 893, 47 L. Ed. 2d 18 (1976), to determine whether the district court afforded the IRPAs "the opportunity to be heard 'at a meaningful time and in a meaningful manner.' " Id. at 333, 96 S.Ct. at 902 (quoting Armstrong v. Manzo, 380 U.S. 545, 552, 85 S. Ct. 1187, 1191, 14 L. Ed. 2d 62 (1965)).

The first Mathews factor involves a specification of "the private interest that will be affected by the official action...." Id. at 335, 96 S.Ct. at 903. Rehashing this point would serve no useful purpose. We conclude, for precisely the same reasons articulated in our earlier opinion, that the IRPAs have a salient private interest in the fees due them for services rendered. See Nineteen Appeals, 982 F.2d at 612.

The second Mathews factor requires us to examine the risk of error presented by the district court's procedures. See Mathews, 424 U.S. at 335, 96 S. Ct. at 903. The last time around we determined that the hearing format invited error. See Nineteen Appeals, 982 F.2d at 612-13. Appellants urge us to find that the proceedings on remand represented no real improvement and again presented an intolerable risk of error--this time because the district court refused to hold an evidentiary hearing, to allow free-form discovery, or to permit cross-examination of PSC members. We conclude, for reasons described more fully in Part II(B), infra, that the format revisions cured the infirmities that led us to invalidate the district court's earlier effort.

The third Mathews factor necessitates an assessment of the public interest, including "the fiscal and administrative burdens" that improved procedural requirements would entail. Mathews, 424 U.S. at 335, 96 S. Ct. at 903. Here, too, past is prologue: we studied this point in the course of the first appeal and remarked the "substantial governmental interest in conserving scarce judicial resources." Nineteen Appeals, 982 F.2d at 614. We also recognized the reasonableness of keeping tight controls on the fee dispute in light of the large number of lawyers involved, the lengthy shelf life of the litigation, and the Supreme Court's admonition that "[a] request for attorney's fees should not result in a second major litigation." Hensley v. Eckerhart, 461 U.S. 424, 437, 103 S. Ct. 1933, 1941, 76 L. Ed. 2d 40 (1983). This important public interest remains intact.

To sum up, the district court reformed its ways, significantly moderating the restrictions originally imposed on the IRPAs. The court levelled the playing field by permitting the IRPAs to present their case in precisely the same manner as their litigation adversaries. Moreover, the court gave both camps adequate notice and a meaningful opportunity to be heard. From a procedural standpoint, then, the adjudicative process employed on remand met the test of fundamental fairness and gave appellants the process that was due.

B. Abuse of Discretion.

Appellants strive to convince us that Judge Acosta abused his discretion in authoring three procedural rulings, namely, (1) denying appellants' entreaty that an evidentiary hearing be held; (2) denying the bulk of their discovery requests; and (3) denying them the privilege of cross-examination. We are not persuaded.

1. Lack of an Evidentiary Hearing. We need not tarry over the supposed error in refusing to hold an evidentiary hearing.3 A district court is not obliged to convene an evidentiary hearing as a means of resolving every attorneys' fee dispute. See Nineteen Appeals, 982 F.2d at 614; Weinberger v. Great N. Nekoosa Corp., 925 F.2d 518, 528 (1st Cir.1991). Because evidentiary hearings in fee disputes are not mandatory, the decision not to convene one is reviewed deferentially, using an abuse-of-discretion standard. See Weinberger, 925 F.2d at 527. In conducting that review, appellate tribunals cannot woodenly apply a preconceived matrix. Rather, flexibility is the watchword. Because a district court has available to it a "wide range of procedures" through which it can "bring a sense of fundamental fairness to the fee-determination hearing while at the same time husbanding the court's resources," Nineteen Appeals, 982 F.2d at 614, flexibility implies substantial discretion. Therefore, when the court chooses among the available options, it can mix and match.

This emphasis on flexibility is heightened when an evidentiary hearing is requested. Even in situations far more inviting than fee disputes, we have been chary about mandating such hearings. See, e.g., Aoude v. Mobil Oil Corp., 862 F.2d 890, 894 (1st Cir.1988) (observing that matters often "can adequately be 'heard' on the papers"). We favor a "pragmatic approach" to the question of whether, in a given situation, an evidentiary hearing is required. Id. at 893. The key determinant is whether, "given the nature and circumstances of the case ... the parties [had] a fair opportunity to present relevant facts and arguments to the court, and to counter the opponents' submissions." Id. at 894. Taking this approach in Aoude, we upheld the issuance of a preliminary injunction without an evidentiary hearing, noting, inter alia, that the judge was "obviously familiar" with the facts and had afforded the parties several opportunities to make written submissions. Id.

The Aoude model can readily be adapted to requests for hearings anent attorneys' fees. Appellants' protest cannot survive the resultant comparison. Judge Acosta knew the case inside and out. He gave the protagonists ample opportunity to present both factual data and legal arguments. He set no page restrictions on written submissions, permitting the IRPAs to proffer thousands of pages of documents both in opposition to the PSC's requisitions and in support of their own fee requests.4 These filings allowed the IRPAs to go into painstaking detail both as to their own contribution to the litigation and as to the reasons why the PSC members deserved a relatively modest slice of the pie for their services in that capacity.

To be sure, this is a high-stakes dispute, but that fact, in and of itself, does not warrant handcuffing the trial court. Matters of great consequence are often decided without live testimony. See, e.g., id. at 893-94 (holding that an evidentiary hearing is not obligatory in respect to an application for preliminary injunction); United States v. DeCologero, 821 F.2d 39, 44 (1st Cir.1987) (same, regarding criminal defendant's motion to reduce his sentence); Amanullah v. Nelson, 811 F.2d 1, 16-17 (1st Cir.1987) (same, regarding habeas review of asylum applicant's detention during exclusion proceedings). In the last analysis, what counts is not the prize at stake, but whether particular parties received "a fundamentally fair chance to present [their] side of the story." Nineteen Appeals, 982 F.2d at 611.

The controlling legal principle, then, is that parties to a fee dispute do not have the right to an evidentiary hearing on demand. When the written record affords an adequate basis for a reasoned determination of the fee dispute, the court in its discretion may forgo an evidentiary hearing. Here, it is pellucid that the litigants' extensive written submissions comprised an effective substitute for such a hearing--particularly since the judge had lived with the litigation from the start and had an encyclopedic knowledge of it. Under these circumstances, the court did not err in refusing to hold yet another hearing. See, e.g., Norman v. Housing Auth., 836 F.2d 1292, 1303 (11th Cir.1988) (upholding propriety of awarding attorneys' fees without an evidentiary hearing "based solely on affidavits in the record"); Bailey v. Heckler, 777 F.2d 1167, 1171 (6th Cir.1985) (explaining that an evidentiary hearing is not required so long as the record is sufficient to permit meaningful review); National Ass'n of Concerned Veterans v. Secretary of Defense, 675 F.2d 1319, 1330 (D.C.Cir.1982) (holding that district court may in its discretion decline to convene a fee hearing where information generated by "documentation accompanying the fee application and through appropriate discovery ... provides an adequate factual basis for an award"); Konczak v. Tyrrell, 603 F.2d 13, 19 (7th Cir.1979) (indicating that "depth of the briefing" can render a hearing on fees unnecessary), cert. denied, 444 U.S. 1016, 100 S. Ct. 668, 62 L. Ed. 2d 646 (1980); see also DeJesus v. Banco Popular de P.R., 951 F.2d 3, 7 (1st Cir.1991) (finding no error in lack of an evidentiary hearing regarding counsel fees absent some "special issue as to which the court needed the assistance of counsel or witnesses").

2. Restrictions on Discovery. Apart from the refusal to convene a full-scale hearing, appellants also complain that the court demonstrated too great an aversion to discovery initiatives. But unlimited adversarial discovery is not a necessary--or even a usual--concomitant of fee disputes, see National Ass'n of Concerned Veterans, 675 F.2d at 1329 (noting that, in general, fee contests should not involve "the type of searching discovery that is typical where issues on the merits are presented"), and, in the circumstances of this case, we think that the court acted well within the province of its discretion in refusing to allow more elaborate discovery.

The Due Process Clause does not require freewheeling adversarial discovery as standard equipment in fee contests. See Nineteen Appeals, 982 F.2d at 614. This case exemplifies the wisdom of the rule. The district court did not shut off all discovery, and the procedures that the court employed--especially the compelled exchange of documentation--minimized the need for additional discovery by giving the IRPAs the raw material that they needed to sift through the particulars of the PSC's fee application. In other words, the court ensured that the IRPAs had access to all the data reasonably necessary to formulate their objections,5 including all the PSC members' time-and-expense submissions, summaries thereof, detailed accounts of the procedures used by the PSC to gather, review, and audit time records, and the working papers, correspondence, and documentation generated by the PSC's accountants during the compilation process. With this banquet of information spread before them, appellants then partook of the court's liberality in allowing them to formulate extensive written submissions.

Furthermore, the court below also had a right to consider the extent to which appellants' request for discovery threatened to multiply the proceedings and turn the fee dispute into a litigation of mammoth proportions. Judge Acosta characterized the IRPAs' discovery foray-which encompassed, inter alia, production of tax returns for employees of all PSC members' firms and details anent fringe benefits (including vacations, maternity leaves, and the provision of training programs)--as "a discovery scheme of needless and unreasonable proportions." It is surpassingly difficult to fault this characterization.

The sweeping nature of appellants' request, coupled with the fact that the focus of the hearings had shifted away from the lodestar and toward a task-oriented assessment of the lawyers' participation in the litigation, give substance to the district court's fears that granting appellants' supplication would have started the parties on the road to a wasteful and time-consuming "satellite litigation." On this ramified record, appellants can demonstrate neither a high level of need for incremental discovery nor preponderant equities in favor of their request. Hence, we cannot say that the district court's denial of further discovery constituted an abuse of the court's considerable discretion. See, e.g., National Ass'n of Concerned Veterans, 675 F.2d at 1329 (holding that district court "retains substantial discretion based on its view of the submissions as a whole" to limit further discovery).

3. Lack of Cross-Examination. As a subset of their claims regarding the supposed necessity for both an evidentiary hearing and additional discovery, appellants contend that the district court should have allowed them to cross-examine the PSC members concerning the hours that they logged and their contribution to the creation of the Fund. This is merely a backdoor attempt to rekindle an extinguished flame and satisfy appellants' thwarted desire for either an evidentiary hearing or extensive depositions.

In Chongris v. Board of Appeals, 811 F.2d 36 (1st Cir.), cert. denied, 483 U.S. 1021, 107 S. Ct. 3266, 97 L. Ed. 2d 765 (1987), we held that, in the context of an administrative hearing, lack of cross-examination did not work a violation of due process. See id. at 41-42. So it is here. Moreover, because the lower court could reasonably conclude that its liberal policy with regard to written submissions, in conjunction with the IRPAs' access to PSC documentation, obviated the need for further probing via cross-examination, pretermitting cross-questioning did not constitute an abuse of discretion. Cf. Copeland v. Marshall, 641 F.2d 880, 905 n. 57 (D.C.Cir.1980) (en banc) (noting that a live hearing is not necessary if "the adversary papers filed by plaintiff and defendant ... adequately illuminate the factual predicate for a reasonable fee").

Appellants' attempt to anchor their claimed right to cross-question PSC members on language excerpted from our earlier opinion, see, e.g., Nineteen Appeals, 982 F.2d at 615, leaves them adrift. We flatly reject the suggestion, noting that appellants, to their discredit, have pieced the argument together by cutting words loose from their logical and contextual moorings, and ignoring limiting language that contradicts their interpretation.

The bottom line is that the district court did not err in refusing to convene an evidentiary hearing, declining to permit more wide-ranging discovery, and barring cross-examination. Thus, whether the issue is cast in a constitutional mold or considered under an abuse-of-discretion rubric, appellants' challenge fails. Either way, the adjudicative process employed on remand passes muster.

Appellants claim that the district court erred as a matter of law in embracing the POF method, rather than the lodestar method, during the fee-setting pavane. The issue of whether a district court may use a given methodology in structuring an award of attorneys' fees is one of law, and, thus, is subject to de novo review. See Liberty Mut. Ins. Co. v. Commercial Union Ins. Co., 978 F.2d 750, 757 (1st Cir.1992).

A. Historical Perspective.

A few introductory comments may lend a sense of perspective. Traditionally, under what has come to be known as the "American Rule," litigants bear their own counsel fees. See Alyeska Pipeline Serv. Co. v. Wilderness Soc'y, 421 U.S. 240, 245, 95 S. Ct. 1612, 1615, 44 L. Ed. 2d 141 (1975). This rule is not without exceptions. Fee-shifting statutes comprise one category of exceptions. See, e.g., 42 U.S.C. Secs. 1988, 2000e-5(k). So, too, certain equitable doctrines furnish a basis for departing from the American Rule. See Nineteen Appeals, 982 F.2d at 606.

When statutory exceptions pertain, we have directed district courts, for the most part, to compute fees by using the time-and-rate-based lodestar method. See, e.g., United States v. Metropolitan Dist. Comm'n, 847 F.2d 12, 15 (1st Cir.1988); Segal v. Gilbert Color Sys., Inc., 746 F.2d 78, 85-86 (1st Cir.1984); see also City of Burlington v. Dague, --- U.S. ----, ----, 112 S. Ct. 2638, 2641, 120 L. Ed. 2d 449 (1992) (acknowledging, in the statutory fee-shifting context, "a strong presumption that the lodestar represents the reasonable fee") (citation and internal quotation marks omitted). A court arrives at the lodestar by determining the number of hours productively spent on the litigation and multiplying those hours by reasonable hourly rates. See Blum v. Stenson, 465 U.S. 886, 896-902, 104 S. Ct. 1541, 1547-51, 79 L. Ed. 2d 891 (1984); Hensley, 461 U.S. at 433, 103 S.Ct. at 1939; Lipsett v. Blanco, 975 F.2d 934, 937 (1st Cir.1992).

Although the lodestar method is entrenched in the statutory fee-shifting context, a growing number of courts have looked elsewhere in "common fund" cases--a category that encompasses cases in which "a litigant or lawyer who recovers a common fund for the benefit of persons other than himself or his client is entitled to a reasonable attorney's fee from the fund as a whole." Boeing Co. v. Van Gemert, 444 U.S. 472, 478, 100 S. Ct. 745, 749, 62 L. Ed. 2d 676 (1980).6 The POF method represents one such alternative approach to fee-setting. This method functions exactly as the name implies: the court shapes the counsel fee based on what it determines is a reasonable percentage of the fund recovered for those benefitted by the litigation. See, e.g., Camden I Condo. Ass'n, Inc. v. Dunkle, 946 F.2d 768, 771 (11th Cir.1991).

Contrary to popular belief, it is the lodestar method, not the POF method, that breaks from precedent. Traditionally, counsel fees in common fund cases were computed as a percentage of the fund, subject, of course, to considerations of reasonableness. See, e.g., Central R.R. & Banking Co. v. Pettus, 113 U.S. 116, 127-28, 5 S. Ct. 387, 392-93, 28 L. Ed. 915 (1885). It was not until the mid-1970s that judicial infatuation with the lodestar method started to spread. See Swedish Hosp. Corp. v. Shalala, 1 F.3d 1261, 1266 (D.C.Cir.1993) (chronicling history of the debate). Many courts embraced the new approach, and a wall of cases soon arose. See, e.g., Copeland, 641 F.2d at 890-91; Furtado v. Bishop, 635 F.2d 915, 919-20 (1st Cir.1980); City of Detroit v. Grinnell Corp., 560 F.2d 1093, 1098 (2d Cir.1977); Grunin v. International House of Pancakes, 513 F.2d 114, 128 (8th Cir.), cert. denied, 423 U.S. 864, 96 S. Ct. 124, 46 L. Ed. 2d 93 (1975); Lindy Bros. Builders, Inc. v. American Radiator & Standard Sanitary Corp., 487 F.2d 161 (3d Cir.1973).

A crack in the wall appeared in 1984 when the Supreme Court took pains to distinguish the calculation of counsel fees under fee-shifting statutes from the calculation of counsel fees under the common fund doctrine. The court described the latter group as comprising cases in which "a reasonable fee is based on a percentage of the fund bestowed on the class." Blum, 465 U.S. at 900 n. 16, 104 S. Ct. at 1550 n. 16. Since Blum involved the application of the lodestar under a fee-shifting statute, footnote 16 is dictum. Yet, it can hardly be dismissed as a slip of the pen, and considered dictum emanating from the High Court carries great persuasive force.7 See Dedham Water Co. v. Cumberland Farms Dairy, Inc., 972 F.2d 453, 459 (1st Cir.1992) (stating general rule that courts should give "considerable weight" to dictum that appears "considered as opposed to casual"); McCoy v. Massachusetts Inst. of Technology, 950 F.2d 13, 19 (1st Cir.1991) (same), cert. denied, 504 U.S. 910, 112 S. Ct. 1939, 118 L. Ed. 2d 545 (1992).

Hard on the heels of footnote 16, the Third Circuit, which had been in the forefront of the movement toward the lodestar method, see, e.g., Lindy Bros., supra, sounded a note of caution. Its blue-ribbon task force, although recommending continued use of the lodestar technique in statutory fee-shifting cases, concluded that all fee awards in common fund cases should be structured as a percentage of the fund. See Report of the Third Circuit Task Force, Court Awarded Attorney Fees, 108 F.R.D. 237, 255 (1985) (hereinafter "Third Circuit Report").

Together, footnote 16 and the Third Circuit Report led to a thoroughgoing reexamination of the suitability of using the lodestar method in common fund cases. This reexamination, in turn, led to more frequent application of the POF method in such cases. See Federal Judicial Center, Awarding Attorneys' Fees and Managing Fee Litigation 63-64 (1994) (hereinafter "FJC Report") (canvassing case law). Today, the D.C. Circuit and the Eleventh Circuit require the use of the POF method in common fund cases, see Swedish Hosp., 1 F.3d at 1271; Camden I, 946 F.2d at 774, and four other circuits confer discretion upon the district court to choose between the lodestar and POF methods in common fund cases, see In re Washington Pub. Power Supply Sys. Sec. Litig., 19 F.3d 1291, 1295 (9th Cir.1994); Rawlings v. Prudential-Bache Props., Inc., 9 F.3d 513, 516 (6th Cir.1993); Harman v. Lyphomed, Inc., 945 F.2d 969, 975 (7th Cir.1991); Brown v. Phillips Petroleum Co., 838 F.2d 451, 454 (10th Cir.), cert. denied, 488 U.S. 822, 109 S. Ct. 66, 102 L. Ed. 2d 43 (1988). We have yet to pass upon the legitimacy of the POF method in common fund cases.8

B. Computing Fees in Common Fund Cases.

We have previously classified this as a common fund case.9 Appellants do not dispute this taxonomy, but, rather, they insist that Judge Acosta erred in using the POF method because the lodestar technique should hold sway in all attorneys' fee determinations.10 Though appellants concede that this court has not yet decided what method(s) of fee allocation appropriately may be invoked in common fund cases, they assert that the lodestar is a far better alternative and that its use should be mandated in this circuit.

We think that a more malleable approach is indicated. Thus, we hold that in a common fund case the district court, in the exercise of its informed discretion, may calculate counsel fees either on a percentage of the fund basis or by fashioning a lodestar. Our decision is driven both by our recognition that use of the POF method in common fund cases is the prevailing praxis and by the distinct advantages that the POF method can bring to bear in such cases.

In complex litigation--and common fund cases, by and large, tend to be complex--the POF approach is often less burdensome to administer than the lodestar method. See Swedish Hosp., 1 F.3d at 1269 (finding POF approach "less demanding of scarce judicial resources"). Rather than forcing the judge to review the time records of a multitude of attorneys in order to determine the necessity and reasonableness of every hour expended, the POF method permits the judge to focus on "a showing that the fund conferring a benefit on the class resulted from" the lawyers' efforts. Camden I, 946 F.2d at 774. While the time logged is still relevant to the court's inquiry--even under the POF method, time records tend to illuminate the attorneys' role in the creation of the fund, and, thus, inform the court's inquiry into the reasonableness of a particular percentage11--the shift in focus lessens the possibility of collateral disputes that might transform the fee proceeding into a second major litigation.

For another thing, using the POF method in a common fund case enhances efficiency, or, put in the reverse, using the lodestar method in such a case encourages inefficiency. Under the latter approach, attorneys not only have a monetary incentive to spend as many hours as possible (and bill for them) but also face a strong disincentive to early settlement. See Third Circuit Report, 108 F.R.D. at 247-48 (finding that, in common fund cases, the lodestar method "encourag[es] lawyers to expend excessive hours" and "creates a disincentive for the early settlement of cases"); see also FJC Report, supra, at 310. If the POF method is utilized, a lawyer is still free to be inefficient or to drag her feet in pursuing settlement options--but, rather than being rewarded for this unproductive behavior, she will likely reduce her own return on hours expended.

Another point is worth making: because the POF technique is result-oriented rather than process-oriented, it better approximates the workings of the marketplace. We think that Judge Posner captured the essence of this point when he wrote that "the market in fact pays not for the individual hours but for the ensemble of services rendered in a case of this character." In re Continental Ill. Sec. Litig., 962 F.2d 566, 572 (7th Cir.1992). In fine, the market pays for the result achieved.

Let us be perfectly clear. We do not pretend that the POF approach is foolproof, or that it suffers from no disadvantages. For example, it may result in the overcompensation of lawyers in situations where actions are resolved before counsel has invested significant time or resources. See Six Mexican Workers v. Arizona Citrus Growers, 904 F.2d 1301, 1311 (9th Cir.1990) (counselling use of the lodestar method rather than the POF method when "the percentage recovery would be either too small or too large in light of the hours devoted to the case or other relevant factors"); see also Third Circuit Report, 108 F.R.D. at 242 (noting "criticism from within the profession" that fees under the POF method sometimes are "disproportionate to actual efforts expended by the attorneys"). The converse is also true; law firms may be less willing to commit needed resources to common fund cases, even those for the public benefit, if the likely recovery is relatively small. It can also be argued that the percentage method may lend itself to arbitrary fee awards by some courts. See generally Washington Pub. Power, 19 F.3d at 1294 n. 2 (counselling that, to avoid arbitrary fee awards, neither the POF nor the lodestar method "should be applied in a formulaic or mechanical fashion"); cf. Laffey v. Northwest Airlines, Inc., 746 F.2d 4, 12-13 (D.C.Cir.1984) (attributing widespread adoption of lodestar method to desire to reduce "arbitrariness characteristic of court awards of attorneys fees" under other methods), cert. denied, 472 U.S. 1021, 105 S. Ct. 3488, 87 L. Ed. 2d 622 (1985). Given the peculiarities of common fund cases and the fact that each method, in its own way, offers particular advantages, we believe the approach of choice is to accord the district court discretion to use whichever method, POF or lodestar, best fits the individual case. We so hold, recognizing that the discretion we have described may, at times, involve using a combination of both methods when appropriate. Cf. Metropolitan Dist. Comm'n, 847 F.2d at 15 (advocating flexible approach to determining fee awards because an overly mechanical rule "sacrifice[s] substance on the altar of form").

In arriving at this decision, we reject appellants' suggestion that Dague, a case in which the Court barred the use of contingency enhancements in respect to fee-shifting statutes, compels a different conclusion. Although the Dague Court stated that "[t]he 'lodestar' figure has, as its name suggests, become the guiding light of our fee-shifting jurisprudence," --- U.S. at ----, 112 S. Ct. at 2641, and remarked that it had "generally" abjured "the contingent-fee model--which would make the fee award a percentage of the value of the relief awarded in the primary action--[in favor of] the lodestar model," id. at ----, 112 S.Ct. at 2643 (citations, footnotes, and internal quotations omitted), these statements were made in the course of a discussion of statutory fee-shifting cases. The Court's reasoning reflected this environment; the opinion stressed the limiting effects of statutory language in fee-shifting cases, see id., and set out "a number of reasons for concluding that no contingency enhancement is compatible with the fee shifting statutes at issue," id. This case, unlike Dague, involves a common fund rather than a statutory fee-shifting scheme. Since Dague, fairly read, does not require abandonment of the POF method typically used in common fund cases, it is not controlling here. Accord Swedish Hosp., 1 F.3d at 1267-70 (concluding that Dague does not bar use of the POF method in common fund cases).

C. Applying the Rule.

Having placed our imprimatur on a decisional model that maximizes flexibility, we move from the general to the specific and turn next to the order under review. In this connection, we rule that the court below did not err in purposing to allocate fees based on the POF method, emphasizing the attorneys' "relative contribution" to the creation of the Fund. In the first place, Judge Acosta had originally stated an intent to compensate the PSC members under a percentage approach. See supra note 10. In "justifiable reliance" on this statement, see Nineteen Appeals, 982 F.2d at 614 n. 19, the majority of the IRPAs did not maintain time records. The difficulties inherent in implementing the lodestar under these circumstances militate in favor of sticking to the POF method. In the second place, as we have explained above, the POF approach offers significant structural advantages in common fund cases, including ease of administration, efficiency, and a close approximation of the marketplace. Finally, a further case-specific factor counsels against using the lodestar here. Unlike the prototypical common fund case, this case involves a subdivision of a fee fund amassed by the operation of sundry contractually determined percentages. Thus, using the POF method to effectuate the subdivision of the Fund brings a sort of elemental symmetry to the fee-setting process. Relatedly, because this case calls for a subdivision of a fee fund, rather than a unitary award of fees, "a trier who attempted punctiliously to follow the classic lodestar formula, to the exclusion of all else, could theoretically wind up awarding the entire fee pool to the PSC, leaving nothing for the IRPAs." Id. at 614 n. 20. Use of the POF method negates any possibility of this totally indefensible result.

In allocating counsel fees, the district court assigned 70% of the Fund to the PSC, leaving 30% to be split among the IRPAs.12 Appellants object. We review this allocation for abuse of discretion, see, e.g., Foley v. City of Lowell, 948 F.2d 10, 18 (1st Cir.1991), mindful that, in respect to fee awards, the trial court's latitude is "extremely broad," Lipsett, 975 F.2d at 937. After scrutinizing the Brobdingnagian record in this case, we are convinced that the court below erred in weighing and synthesizing the factors relevant to a division of the fees, and in settling upon so lopsided a split.

A. Cutting the Pie.

In the proceedings on remand, Judge Acosta lavished praise on all the plaintiffs' lawyers, lauding the "high caliber legal representation" provided by both the PSC and the IRPAs. He then summarized the tasks undertaken by the two sets of attorneys in the course of the litigation. In the judge's view, the PSC's most significant accomplishments included (1) performing a comprehensive on-site investigation of the accident scene, (2) "identif[ying] the manufacturers and suppliers of many ... products and services ... and develop[ing] theories of liability against each opponent," (3) drafting plethoric pleadings, including the master complaint, weekly agendas, and several pretrial memoranda, (4) filing "literally hundreds of motions ... on numerous topics, including many novel and creative issues," (5) orchestrating extensive pretrial discovery, (6) conducting the nine-week Phase I trial and the fifteen-month Phase II trial (in the course of which the PSC called 313 witnesses and offered 1,455 exhibits), and (7) "aggressively pursu[ing] settlement negotiations." The court visualized the IRPAs' main accomplishments as comprising (1) maintaining direct client communication, counselling clients, and keeping them abreast of developments in the litigation, (2) carrying out the factual investigation incident to individual cases, with especial emphasis on issues pertaining to damages, (3) retaining experts, including physicians, economists, and actuaries, and, once the experts had been located, collaborating with them to establish damages, (4) researching client-specific legal issues (e.g., standing, assumption of risk), (5) representing individual plaintiffs in connection with ancillary matters, including probate, inheritance, insurance, and domestic relations matters, (6) meeting with Judge Bechtle "as part of the settlement scheme to negotiate settlement values for [individual] cases," and (7) assisting clients in reaching informed decisions (including decisions about whether to accept or reject proffered settlements). Moreover, certain selected plaintiffs were used as exemplars for purposes of the Phase II trial, and the IRPAs who represented those plaintiffs actually presented the evidence pertaining to their clients' damages.

Having made these ledger entries, the district court then tabulated the columns. It concluded that "reasonable compensation for the work undertaken requires recognition of the massive undertaking of the PSC in terms of the organizational and financial requirements, the overwhelming amount of work performed, the significant time constraints, and the numerous complex and novel issues addressed during the proceedings...." Contrasting this workload "with the IRPAs' efforts in client communication and counseling, client preparation for settlement, and handling of the damages issues," the court awarded the PSC 70% of the fee due under each individual contingency agreement, thus permitting each IRPA to retain only 30% of the fee promised by the client.

B. Evaluating the Court's Handiwork.

We are uneasy with the way in which the lower court cut the fee pie, and with the size and shape of the resultant wedges, for several reasons.

First, we are troubled by the implications of a scheme in which the trial judge selects a chosen few from many lawyers who volunteer, assigns legal tasks to those few (thereby dictating, albeit indirectly, the scope of the work remaining to be done by the many), and then, in awarding fees, heavily penalizes the very lawyers to whom he has relegated the "lesser" duties. Courts must recognize that while such an arrangement may be a necessary concomitant to skillful case management of mass tort suits, it nevertheless significantly interferes with an attorney's expectations regarding the fees that his or her client has agreed to pay. Conversely, lead counsel are typically volunteers, as in this case, and, as such, they have no right to harbor any expectation beyond a fair day's pay for a fair day's work if a fee fund develops. Cf. Matthew 20:1-16 (recounting parable of the laborers in the vineyard). We believe that trial courts should take these differing expectations into account in allocating fees. Here, the judge's rescript does not suggest that he factored these expectations into the decisional calculus.

Courts must also be sensitive to a second facet of economic reality: the power to appoint lead counsel gives the trial judge an unusual degree of control over the livelihood of the lawyers who practice before the court. Though such appointments are often an administrative necessity in complex litigation, and disproportionate fees are at times an unavoidable consequence of the classic common fund "free rider" problem, see generally Mancur Olson, Jr., The Logic of Collective Action (1971), the judge must attempt to avoid any perception of favoritism. This need is especially acute in mass tort litigation where, as this case illustrates, free rider concerns are minimized by the important nature of the work to be done by claimants' individually retained attorneys. In this case, moreover, free rider concerns are also lessened by the fact that most of the IRPAs applied for appointment to the PSC, see Nineteen Appeals, 982 F.2d at 605 (noting that over 40 of the 56 IRPAs volunteered to serve on the PSC), thus signifying their willingness to pay full fare. The record does not contain any clue intimating that Judge Acosta considered these factors in ordering that 70% of the fees be paid to the PSC.

Third, and relatedly, this case required the IRPAs not merely to go along for a free ride but to earn their keep. They exhibited great versatility, counseling clients, researching medical histories, arranging for specialists to evaluate injuries, preparing the damages aspect of each case (including extensive work with physicians, psychologists, actuaries, vocational specialists, and other witnesses), obtaining evidence needed to prove losses of earnings and earning capacity, responding to client-specific discovery, preparing for and attending clients' depositions, negotiating settlement values before Judge Bechtle, assisting clients with probate, insurance, and tax matters, and handling a bewildering array of idiosyncratic problems as they developed. This is a far cry from the paradigmatic common fund case--say, a securities class action--in which class counsel do virtually all the work, and other counsel piggyback on their efforts. See, e.g., In re Ivan F. Boesky Sec. Litig., 948 F.2d 1358, 1364-65 (2d Cir.1991); see also Randall S. Thomas & Robert G. Hansen, Auctioning Class Action and Derivative Lawsuits: A Critical Analysis, 87 Nw.U.L.Rev. 423, 429 (1993) (explaining that, in general, lead counsel in class actions have "substantial authority to conduct the litigation, even to the exclusion of other counsel"); Jonathan R. Macey & Geoffrey P. Miller, The Plaintiffs' Attorney's Role in Class Action and Derivative Litigation: Economic Analysis and Recommendations for Reform, 58 U.Chi.L.Rev. 1, 3 (1991) (observing that plaintiffs' class action attorneys have "nearly plenary control over all important decisions in the lawsuit" because of the absence of monitoring by clients). We see no sign that the district court gave significant weight to this reality.

This leads directly to a fourth point. We have carefully considered the IRPAs' compendious submissions and are of the view that Judge Acosta undervalued the worth of the client contact/counseling aspect of this litigation. Such services are labor-intensive and frequently low in visibility--at least in visibility from the bench. Thus, they are susceptible to being overlooked, leading to an overemphasis on the relative value of the court-related work. Despite their lack of visibility, however, the mundane chores incident to client representation are particularly critical in a mass tort common fund case. We explain briefly.

In a securities class action many of the victims do not participate in the lawsuit, and are aware of their loss dimly, if at all. See, e.g., Macey & Miller, supra, at 30 (noting that "[i]n the large-scale, small-claim class action context ... [plaintiffs] are typically unaware that they even have a claim against the defendant"). The mass tort context supplies a stunning contrast. In a mass tort action, the victims' losses (whether of life, limb, or loved ones) are almost always keenly felt, and are usually not amenable to computation by a simple arithmetic formula. As a result, the individual plaintiffs typically require a multitude of services, many of which cannot be satisfied by an impersonal steering committee. In such circumstances, the attention of the individually retained attorneys becomes crucial to the success of the overall enterprise.13 That important contribution demands appropriate recognition.

Fifth, although we do not dispute the district court's assessment of the quality of the PSC's work, this factor cancels itself out to some extent. After all, the district court repeatedly commented upon "the excellence of the work performed by all attorneys" (emphasis supplied), and left no doubt but that both sets of plaintiffs' lawyers had rendered exemplary service. Given these widespread plaudits, it seems manifestly unfair to reward excellence on the part of one group and not the other.

Sixth, the district court failed to advance any reasoned explanation as to why it boosted the PSC's share of the Fund from 52% in the initial go-round to 70% on remand.14 Though we have great confidence in Judge Acosta, his silence on this subject leaves the award open to a perception that appellants have been penalized for successfully prosecuting their previous appeals. Cf. North Carolina v. Pearce, 395 U.S. 711, 89 S. Ct. 2072, 23 L. Ed. 2d 656 (1969) (discussing importance of dispelling any appearance of vindictiveness when a judge imposes a more severe sentence upon a criminal defendant after the defendant wins a new trial).

Seventh, the district court erred in failing to compensate the representative trial counsel--those IRPAs who, though not members of the PSC, prepared and/or tried the so-called "representative" cases--for their work in that capacity. Just as the PSC members deserved compensation for their endeavors on behalf of the whole, the IRPAs who labored as representative counsel conferred a common benefit, and must be compensated accordingly.

Last--but far from least--we are persuaded, on whole-record review, that it is simply unreasonable to award 70% of the aggregate fees to the attorneys who managed the litigation, leaving only 30% of the Fund to those who brought in the clients and worked hand-in-hand with them throughout the pendency of this long safari of a case. Because mass tort cases are a breed apart, it is difficult to envision situations in which, if fees are divided between lead counsel and individually retained counsel under a POF formula, the latter will not be entitled to at least half the fees.15 We do not think that this litigation, though unique, so far overshoots all other cases as to warrant a substantially larger differential. See, e.g., Vincent v. Hughes Air W., Inc., 557 F.2d 759 (9th Cir.1977) (upholding district court's allocation of 5% of gross recovery, or approximately 20% of the fee fund, to lead counsel in mass tort action).

Concluding, as we do, that the fee allocation reflects a serious error of judgment, and therefore an abuse of discretion, we vacate the award.

Ordinarily, "an improper calculation of attorneys' fees necessitates remand for reconfiguration of the award." Lipsett, 975 F.2d at 943. But this rule admits of exceptions, so long as "the record is sufficiently developed that we can apply the law to the facts before us and calculate a fair and reasonable fee without resorting to remand." Id. Here, that qualification is satisfied; the record is voluminous and this court is painfully familiar with the particulars of this fee imbroglio. Nonetheless, an appellate court must think long and hard before usurping the district court's usual prerogatives, and, therefore, we doubt that this case would fall within the narrow confines of the exception under ordinary circumstances. But the circumstances here are extraordinary, and common sense commands that we not turn a blind eye to the reality of events.

This litigation has passed the point of diminishing returns. The holocaust that underlies the plaintiffs' claims occurred almost a decade ago. The meat-and-potatoes litigation is over; with one small exception, see supra note 1, only a side dish--attorneys' fees--remains on the table. The amount of time, energy and money already devoted to this peripheral item has careened virtually out of control. Remanding would invite an even greater investment in the side dish--and we are reluctant to sanction the squandering of additional resources for this purpose. We have, at times, with considerably less provocation, simply grasped the bull by the horns and fixed the fees ourselves. See, e.g., Jacobs v. Mancuso, 825 F.2d 559, 562 (1st Cir.1987); Grendel's Den v. Larkin, 749 F.2d 945, 951 (1st Cir.1984).

We realize that dividing the Fund among groups of attorneys in accordance with the POF method cannot be accomplished with surgical precision. We--or a district court, for that matter--must necessarily traffic in estimates. Taking into account all the facts and circumstances, we conclude that we should subdivide the Fund ourselves, rather than remand yet again. We also conclude that, on balance, assigning 50% of the Fund to the PSC and 50% to the IRPAs comprises a fair and reasonable allocation.

This division reflects the district court's determination that the PSC contributed handsomely to the creation of the Fund--it is, after all, at the high end of what a court should usually award16--while at the same time correcting for the district court's undervaluation of the IRPAs' contributions. This division also strikes a sensible balance between the equity-based common fund doctrine, which guards against the unjust enrichment of free riders, and the need to avoid adding insult to injury in a situation in which the court selects lead counsel from amidst a group of willing volunteers and thereafter invades the contingency agreements of the rejected lawyers to compensate the select few. Moreover, this division is not incommensurate with the time records of the PSC. Even if, as an uncritical reading of the record suggests, the PSC spent as many as 166,000 hours on the litigation,17 a 50% allocation (roughly $34,000,000) pays the members well. Although we have no tabulation of IRPA hours to compare with this total, the PSC's time records are still a valid measure of the vast resources its members expended in the course of the litigation.

One loose end remains. It involves appropriate compensation for the IRPAs who tried the "representative" cases. As we stated earlier, see supra p. 311, their participation in the Phase II trial inured to the benefit of all plaintiffs. Thus, in presenting the representative claims, the lawyers were acting as de facto PSC members. It is only logical, therefore, that their compensation for those services be drawn from the PSC's share of the fee Fund. Since the record is inadequate to permit us to place a dollar value on these services, we leave it to the district court to determine the amount of compensation due to the non-PSC members who served as representative trial counsel during the Phase II trial for their services in that capacity, and then to order that sum paid out of the PSC's share of the Fund.

We need go no further. For the reasons we have expressed, we vacate the order allocating attorneys' fees; direct that the fee Fund be divided equally among the PSC, on the one hand, and the IRPAS, on the other hand; and remand for the entry of a suitable decree and for further proceedings consistent with this opinion. Costs shall be taxed in favor of the appellants.

It is so ordered.