In Re Unisys Corp. Retiree Medical Benefits Erisa Litigation

886 F. Supp. 445 | E.D. Pa. | 1995

886 F. Supp. 445 (1995)

In re UNISYS CORPORATION RETIREE MEDICAL BENEFITS ERISA LITIGATION.

MDL No. 969.

United States District Court, E.D. Pennsylvania.

March 20, 1995.

OPINION

CAHN, Chief Judge.

Plaintiffs brought this action alleging that Unisys Corporation ("Unisys") violated the Employee Retirement Income Security Act ("ERISA"), 29 U.S.C. § 1001, et seq., when it unilaterally modified its retiree medical benefits plans. On October 26, 1994, the court approved a partial settlement ("the settlement") as to Sperry and Burroughs incentive retirees ("the settlement class"). Attorneys for plaintiffs now petition the court for an award of fees and expenses.

I. BACKGROUND & PROCEDURAL HISTORY

The court will recite the background and procedural history of this case, and describe the settlement, because of their relevance to the award of attorneys' fees. In September, 1986, Sperry Corporation and Burroughs Corporation merged to form Unisys Corporation. After the merger, Unisys maintained the pre-existing medical benefit plans ("the predecessor plans") for Sperry and Burroughs retirees. In 1989, Unisys created the Post-Retirement and Extended Disability Medical Plan ("the old plan") to cover all employees who retired after April 1, 1989, most of whom were former Sperry and Burroughs employees. At that time, Unisys left the predecessor plans intact. On January 1, 1993, Unisys terminated the predecessor plans and the old plan and replaced them with the new Unisys Post-Retirement and Extended Disability Medical Plan. Under this plan, the retirees no longer receive free medical insurance. Instead, they must pay a portion of the monthly premiums. The plan required the retirees to pay the full cost of premiums beginning January 1, 1995.

This litigation was commenced by a series of lawsuits filed in federal courts in Pennsylvania, New York, Minnesota, and Michigan by various groups of retirees who had worked for Unisys or its Sperry and Burroughs predecessors in those states and elsewhere throughout the country. The New York, Minnesota, and Michigan actions subsequently were transferred to this court by the Judicial Panel on Multi-District Litigation, and this court ultimately consolidated all of the actions. On June 9, 1993, this court certified the case as a class action pursuant to Fed.R.Civ.P. 23(b)(2).

The class action covered all former Sperry, Burroughs, and Unisys employees, and their eligible dependents, who were participating in the former Sperry, Burroughs, and Unisys post-retirement medical plans in November 1992. Among these plaintiffs were "incentive retirees", who retired under a voluntary early retirement incentive program.^[1]

Plaintiffs asserted three general claims. First, they claimed that they were entitled to lifetime benefits as a matter of contract. They based this assertion on the summary plan descriptions ("SPDs") in the predecessor plans, which promised them that their benefits would continue for life.

Second, plaintiffs asserted that even if the SPDs were not enforceable as contracts, Unisys and its predecessors had breached their fiduciary duty to the retirees by affirmatively leading them to believe, by repeated written and oral assurances, that their benefits would continue for life, and that the benefits would not be subject to change after retirement.

Third, plaintiffs asserted that Unisys should be equitably estopped, given these assurances and plaintiffs' reliance on them, from exercising any right to change or terminate the plans.

Additionally, the incentive retirees asserted contract claims separate from the claims they shared with the regular retirees. They claimed that Unisys offered them certain benefits to induce them to retire, that they accepted those offers by retiring earlier than they otherwise would have done, and that the offer and acceptance formed binding contracts, apart from the contracts arising from the SPDs.

Unisys defended each of these claims based on language in all of the SPDs reserving the right to change or end the plans. Unisys argued that these reservation of rights clauses ("RORs") gave it the right to modify the lifetime benefit provisions or terminate the plans at any time.

The court granted partial summary judgment disposing of all claims brought by the regular retirees of Burroughs and Unisys. In re Unisys Corp. Retiree Medical Benefits ERISA Litigation, 837 F. Supp. 670 (E.D.Pa. 1993). As to the regular retirees of Sperry, the court denied Unisys' motion for summary judgment with respect to the breach of contract claims,^[2] but granted summary judgment on the claims for breach of fiduciary duty and estoppel. The partial summary judgment had no effect on the incentive retirees' claims, however, because Unisys moved for summary judgment against the regular retirees only.

The court then held a non-jury trial on all of the claims which remained after the partial grant of summary judgment. Testimony was heard from more than 80 witnesses and approximately 800 exhibits were admitted. After trial, extensive briefing, and oral argument, the court ruled in favor of Unisys on all claims brought by the Unisys incentive retirees. The court also ruled in favor of Unisys on the Sperry regular retirees' contract and estoppel claims. However, the court allowed the Sperry regular retirees to proceed with their breach of fiduciary duty claims.^[3] It remains to be determined whether individual hearings will be necessary to determine which Sperry regular retirees are entitled to a remedy for Unisys' breach of fiduciary duty, and what such remedy should be. Cross-appeals are pending from the summary and post-trial judgments.

II. SETTLEMENT

Following the trial, and four days before closing arguments, the Sperry and Burroughs incentive retirees and Unisys agreed on a proposed settlement. On October 17, 1994, the court held a hearing pursuant to Fed.R.Civ.P. 23(e) on plaintiffs' motion to approve the settlement.^[4] The court subsequently approved the settlement. See In re Unisys Corp. Retiree Medical Benefits ERISA Litigation, No. MDL 969, 1994 WL 702638 (E.D.Pa. Nov. 3, 1994).

The settlement obligates Unisys to pay $111 million — $72.9 million for the Sperry incentive retirees and $38.1 million for the Burroughs incentive retirees — as a "present value" contribution toward the cost of continued medical coverage for the incentive retirees. Since these settlement amounts are "present value" amounts, they will grow each quarter by a fixed rate of interest applied to the outstanding, unused balance of the settlement amounts promised by Unisys.

The settlement amounts will be supplemented by retiree contributions. These monies will be combined and used to pay for a benefits plan designed by the retirees in consultation with Unisys and administered by Unisys.^[5] These benefits will continue until the settlement monies are exhausted, or until the end of the life of the last living incentive retiree, whichever is earlier. After the settlement amounts are exhausted, Unisys will be obligated to permit incentive retirees to participate in any other non-bargaining post-retirement medical plan Unisys then maintains. If Unisys is sold before exhaustion of the funds, its obligations under the settlement will be binding on successor entities.

The Sperry and Burroughs settlement accounts will be supervised by separate Advisory Committees consisting of Sperry and Burroughs incentive retirees. The Committees will be assisted by actuarial consultants and attorneys as needed. The Committees' principal responsibility will be to monitor Unisys' compliance with its settlement obligations, periodically to review the status of the settlement accounts, and to determine any needed changes in the contribution schedules or coverage terms of the plan. Disputes about implementation and compliance with the Settlement Agreement, if not resolved under the negotiated dispute resolution procedures established by the settlement, will be subject to this court's jurisdiction.

III. PETITIONS FOR ATTORNEYS' FEES

Under the terms of the settlement, Unisys will pay to Class Counsel "reasonable attorneys' fees and disbursements relating to the Incentive Retirees' claims and the settlement thereof...." Stipulation and Agreement of Settlement and Dismissal of Claims of Sperry and Burroughs Incentive Retirees at 31. This payment "shall be credited against and reduce the applicable Incentive Retiree Subclass Settlement Amount." Id.

Counsel from the following thirteen law firms now jointly petition the court for an award of attorneys' fees and reimbursement of litigation expenses: Berger & Montague, P.C.; Davis, Miner, Barnhill & Galland; Joseph F. Roda, P.C.; Mansfield & Tanick; Meagher & Geer; Hirsch & Associates; Heller, Kapustin, Gershman & Vogel; Milberg, Weiss, Bershad, Specthrie & Lerach; Rossbacher & Associates; Schiffrin & Craig, Ltd.; Levin, Fishbein, Sedran & Berman; Gottlieb & Goren; and Sommers, Schwartz, Silver & Schwartz, P.C. They seek out-of-pocket costs and expenses of $401,775.30 from the Sperry settlement amount and $153,943.81 from the Burroughs settlement amount. They also seek $7,949,500 in fees — $5,092,000 from the Sperry settlement amount, and $2,857,500 from the Burroughs settlement amount — and have agreed among themselves upon a tentative allocation of this aggregate amount. They argue that this amount is reasonable under either the "lodestar" method of awarding attorneys' fees, as a product of their aggregate lodestar of $3,727,689 times a "blended" multiplier of 2.13,^[6] or as a reasonable percentage of the present value of the settlement funds.

These counsel had anticipated that counsel from the firm of Miller, Faucher, Chertow, Cafferty & Wexler would join their petition. After adding the tentative allocation to that firm, they initially calculated the "reasonable percentage" as 7.5%. However, counsel from Miller, Faucher subsequently filed a separate petition for fees and expenses, recommending that the court use an alternative method for calculating fees. Consequently, the "reasonable percentage" sought in the first petition is 7.5 minus the percentage constituted by the tentative allocation to Miller, Faucher.

Counsel from Miller, Faucher seek reimbursement of expenses in the amount of $21,719.22, and request an award of attorneys' fees in the amount of $286,993.86. They urge that their suggested method for calculating fees be applied to all firms. They also request that the method used by the court with regard to the aforementioned thirteen firms be applied equally to Miller, Faucher.^[7]

The court finds this last request to be reasonable, and will therefore grant it. Moreover, for the sake of simplicity, the court will treat the first petition as if it included Miller, Faucher. Accordingly, the court will consider 7.5% as a possible award under a percentage approach, and will hereinafter refer to counsel from the aforementioned fourteen firms as "class counsel." However, in deciding upon the appropriate method for awarding fees, the court will consider the approach recommended by counsel from Miller, Faucher.

Dale Nathan, Esq., of Nathans and Associates, also petitions the court for attorneys fees and expenses. He seeks $49,835.62 in fees and $15,913.21 in expenses, or two-thirds of the $98,623.25^[8] spent on behalf of settling plaintiffs from the Minnesota Retiree Defense Organization (hereinafter "MRDO").^[9]

Objections to these petitions have also been filed. Mr. Nathan, on behalf of some members of the MRDO, has objected to the award requested by class counsel on a number of grounds, and has submitted a lengthy analysis of the billing records submitted by class counsel. As part of their own fee petition, counsel from Miller, Faucher have objected, as does Mr. Nathan, to the full inclusion by class counsel of hours that were expended on behalf of both the settlement class and non-settling plaintiffs, and have also suggested that a multiplier of 2.0 be used if a lodestar approach is employed.

Defendant Unisys Corporation has objected to class counsel's fee petition. Defendant contends that, in light of the time charges by its own counsel, class counsel's time charges are unreasonable. Defendant also objects on the same ground as Mr. Nathan and Miller, Faucher concerning hours expended on behalf of all plaintiffs.^[10] Defendant also engaged Legalgard, Inc., a legal cost and case management services company, to audit the billing records submitted by class counsel. Additionally, class counsel object to any award by this court to Mr. Nathan. Finally, between 3.3% and 5.4% of the settlement class filed separate objections to the fee petitions.^[11] These objectors cite a lack of a complete recovery by the settlement class, a concern about paying for work done for other classes of plaintiffs, and a conflict of interest on the part of class counsel.

IV. DISCUSSION

A. AWARDS FROM COMMON FUNDS

Under the "American Rule," parties to a lawsuit bear their own expenses, regardless of which party prevails. Alyeska Pipeline Serv. Co. v. Wilderness Soc'y, 421 U.S. 240, 247, 95 S. Ct. 1612, 1616, 44 L. Ed. 2d 141 (1975).^[12] Several exceptions to the American Rule, however, have been created. One well-established exception is known as the "common fund" or "equitable fund" doctrine.^[13] "[A] litigant or a 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."^[14]Van Gemert, 444 U.S. at 478, 100 S.Ct. at 749. This doctrine has been justified by basic equity: in terms of its prevention of unjust enrichment, Trustees v. Greenough, 105 U.S. 527, 532, 26 L. Ed. 1157 (1882); Sprague v. Ticonic Nat'l Bank, 307 U.S. 161, 166, 59 S. Ct. 777, 780, 83 L. Ed. 1184 (1939); Van Gemert, 444 U.S. at 478, 100 S.Ct. at 749, and out of the notion of "quantum meruit", Lindy Brothers Builders, Inc. of Philadelphia v. American Radiator & Standard Sanitary Corp., 487 F.2d 161, 165 (3d Cir.1973) (Lindy I) ("the individual seeking compensation has, by his actions, benefitted another and seeks payment for the value of the service performed"), appeal following remand, 540 F.2d 102 (3d Cir.1976) ("Lindy II").^[15]

The litigation before this court has produced two accounts, with a combined present-day value of $111 million, established by defendant to be shared by a certified class of more than 7000 plaintiffs. The settlement clearly has produced a "common fund." See Federal Judicial Center, Awarding Attorneys' Fees and Managing Fee Litigation 52-59 (1994) (discussing prerequisites of a "common fund").

The fact that this case was brought under ERISA, a fee-shifting statute, does not preclude recovery of attorneys' fees from the common fund that arose from settlement. Fee-shifting statutes should not circumscribe the operation of the common fund doctrine unless that operation conflicts with an intended purpose of the statute. County of Suffolk v. Long Island Lighting, 907 F.2d 1295, 1327 (2d Cir.1990). Here, "[b]ecause the settlement agreement explicitly provides that the money paid by defendants includes an unspecified sum for attorneys' fees, an award of fees from this fund would not be inconsistent with Congress's intention that, in ERISA cases, `the offending party bear the costs of the award rather than ... plan participants.'" Bowen, 760 F.Supp. at 894 (quoting Eaves v. Penn, 587 F.2d 453, 464 (10th Cir.1978)). An award of fees from this fund would also further "the policy, underlying [ERISA], of providing both prospective plaintiffs and their attorneys an economic incentive to bring meritorious ERISA cases." Bowen, 760 F.Supp. at 894.^[16]

Because this is a common fund case, the court is responsible for applying heightened judicial scrutiny to the fee requests. In re Fine, 751 F.2d at 583.^[17] Accordingly, the court will endeavor to determine a reasonable award of attorneys' fees in this case.^[18] This task is two-fold: the court must first select an appropriate method for awarding fees, and then that method must be applied.

B. SELECTING THE METHOD

1. Percentage v. Lodestar

From Central R.R. & Banking Co. v. Pettus, 113 U.S. 116, 5 S. Ct. 387, 28 L. Ed. 915 (1885), until 1973, courts typically based attorneys' fees awards in common fund cases on a reasonable percentage of the fund. Herbert P. Newberg, Attorney Fee Awards, § 2.02 at 31 (1986) [hereinafter Attorney Fee Awards]; Task Force Report, 108 F.R.D. at 242; Arenson v. Board of Trade of City of Chicago, 372 F. Supp. 1349, 1357 n. 14 (N.D.Ill.1974) (listing fifteen cases in which court awarded a fee as a percentage of recovery). Judges employed a variety of factors to set the reasonable percentage^[19]; the most common factor was "the size of the fund or the amount of benefit produced for the class." Task Force Report, 108 F.R.D. at 242.

However, "the percentage-of-recovery system sometimes resulted in strikingly large fee awards." Id. Criticism of excessive awards and unbounded judicial discretion mounted. See, e.g., Illinois v. Harper & Row Publishers, 55 F.R.D. 221, 224 (N.D.Ill.1972). The Court of Appeals for the Third Circuit led the movement away from the percentage approach. Task Force Report, 108 F.R.D. at 242. In 1973, in an opinion by Chief Judge Seitz, the Court of Appeals established an alternative approach to determining fees in order to make meaningful review less difficult and to ensure that fee awards better reflected the reasonable value of the services provided: the "lodestar" approach. Lindy Brothers, Inc., of Philadelphia v. American Radiator & Standard Sanitary Corp., 487 F.2d 161 (3rd Cir.1973) ("Lindy I"), appeal following remand, 540 F.2d 102 (3rd Cir. 1976) ("Lindy II").

Under the lodestar approach to valuing an attorney's services, the court must first determine the "lodestar" — the number of hours reasonably expended multiplied by a reasonable hourly market rate. Lindy I, 487 F.2d at 167. "This calculation provides an objective basis on which to make an initial estimate of the value of a lawyer's services." Hensley v. Eckerhart, 461 U.S. at 433, 103 S.Ct. at 1939.^[20] The court may then adjust the lodestar, by means of a "multiplier" or an "enhancement", to reflect the contingent nature of the litigation and the quality of services rendered by counsel. In re Fine, 751 F.2d at 583-84; Lindy I, 487 F.2d at 167-68.

Other courts of appeals soon followed the Third Circuit's lead, and began emphasizing the time/rate value of counsel's services. Most adopted approaches virtually identical to the Lindy lodestar. Task Force Report, 108 F.R.D. at 244. See, e.g., City of Detroit v. Grinnell Corp., 495 F.2d 448, 470-71 (2d Cir.1974). A few circuits adopted alternative approaches.^[21]

The 1980s witnessed another major change in the determination of attorneys' fees in common fund cases. This change was sparked by two developments. First, the Supreme Court in Blum v. Stenson noted that, unlike in fee-shifting cases, attorneys' fees in common fund cases are "based on a percentage of the fund bestowed on the class." 465 U.S. at 900, n. 16, 104 S.Ct. at 1550, n. 16.^[22]

Second, the Chief Judge of the Court of Appeals for the Third Circuit, Ruggero J. Aldisert, formed a Task Force of judges and lawyers to study the subject of court awarded attorneys' fees. The Task Force, chaired by Judge Sarokin, subsequently recommended a return to the percentage method in common fund cases. Task Force Report, 108 F.R.D. at 255-56.

Echoing criticisms from other quarters^[23], the Task Force Report identified many problems inherent in the lodestar (Lindy) analysis. The most salient of these problems are the following:

1) Lindy increases the workload of an already overtaxed judicial system;

2) The elements of the Lindy process are insufficiently objective and produce results that are far from homogeneous;

3) Despite the apparent simplicity of the Lindy formulation, much confusion and a lack of predictability remain in its administration;^[24]

4) Lindy creates a disincentive for the early settlement of cases; rather, it gives attorneys an incentive to accumulate hours spent on the case, often at the expense of the interests of the plaintiff class.

Task Force Report, 108 F.R.D. at 246-49.

The Task Force noted that "there is a widespread belief that the deficiencies of the current system either offset or exceed [the] benefits" of the lodestar method. Id. at 246. But see Matter of Superior Beverage/Glass Container, 133 F.R.D. 119, 125-28 (N.D.Ill. 1990) (explaining that objections to lodestars are overstated). The Task Force concluded that, while the lodestar approach continued to have merit in some statutory fee-shifting cases, fee awards in common fund cases and in "statutory fee cases that are likely to result in a settlement fund from which adequate counsel fees can be paid" should be based on some percentage of the fund. Task Force Report, 108 F.R.D. at 255. See also Sala v. National R.R. Passenger Corp., 128 F.R.D. 210, 213-14 (E.D.Pa.1989) (discussing the different dynamics in statutory fee cases and common fund cases, and agreeing with the Task Force Report that "applying variant fee recovery methods to these two categories of actions will best achieve the differing policy objectives each was designed to further"); Brown, 838 F.2d at 454 (similar discussion).

The percentage approach has subsequently been employed with increasing frequency in common fund cases. Two circuits require the percentage approach. See Swedish Hospital Corp. v. Shalala, 1 F.3d 1261, 1271 (D.C.Cir.1993); Camden I Condominium Ass'n v. Dunkle, 946 F.2d 768, 774 (11th Cir.1991). Additionally, at least five other circuits allow the use of either the lodestar approach or the percentage approach. In re Washington Public Power Supply System Securities Litigation., 19 F.3d 1291, 1296 (9th Cir.1994); Rawlings v. Prudential-Bache Properties, 9 F.3d 513, 515-17 (6th Cir.1993); Weinberger v. Great Northern Nekoosa Corp., 925 F.2d 518, 526 n. 10 (1st Cir.1991); Harman v. Lyphomed, Inc., 945 F.2d 969, 974-75 (7th Cir.1991); Brown, 838 F.2d at 454.^[25]

The Court of Appeals for the Third Circuit has not considered whether the Lindy decisions precludes use of the percentage approach.^[26] District courts within the Third Circuit, however, have endorsed the use of the percentage method on many occasions. See, e.g., In re U.S. Bioscience Securities Litigation, 155 F.R.D. 116, 118 (E.D.Pa.1994) (following the rationale of the Task Force Report and also noting that by adopting a percentage approach, the number of lawyers for the class (34) becomes irrelevant); Sala, 128 F.R.D. at 214 (court should employ percentage approach in common fund cases so long as there "are no circumstances suggesting its application would be unjust"); In re SmithKline Beckman Corp. Securities Litigation, 751 F. Supp. 525, 533 (E.D.Pa.1990); In re First Fidelity Bancorporation Securities Litigation, 750 F. Supp. 160 (D.N.J.1990); In re TSO Financial Litigation, No. CIV.A. 87-7903 et al., 1989 WL 80316, at *6 (E.D.Pa. July 17, 1989); In re GNC Shareholder Litigation, 668 F. Supp. 450, 451-52 (W.D.Pa. 1987). But see In re Centocor, Inc. Securities Litigation, No. 92-CV-1071, 1993 WL 189937, at *5 (E.D.Pa. June 2, 1993) (using lodestar approach); Waldner v. Shulman, No. CIV.A. 86-7381, 1989 WL 100184, at *6 (E.D.Pa. Aug. 25, 1989) (although Task Force Report was favorably received and discussed at 1985 Judicial Conference, percentage approach impermissible until Third Circuit formally approves or adopts the Report and Recommendation).

Class counsel urge the court to follow this trend and calculate fees based on a percentage of the settlement fund.^[27] Eschewing the lodestar approach for the percentage approach offers many advantages. "A number of salutary effects can be achieved ... including removing the inducement to unnecessarily increase hours, prompting early settlement, reducing burdensome paperwork for counsel and court[,] and providing a degree of predictability to fee awards." In re Activision Securities Litigation, 723 F. Supp. 1373, 1376 (N.D.Cal. 1989). Moreover, "[t]he percentage method is widely used in the legal marketplace in contingent fee agreements and better reflects what a client, at the outset of the litigation, is willing to pay." Steiner v. Hercules Inc., 835 F. Supp. 771, 792 (D.Del.1993) (citation omitted). Additionally, the sharper focus on the common fund mandated by a percentage approach is appropriate because "a common fund is itself the measure of success." Herbert P. Newberg, Newberg on Class Actions, § 14.03 at 186 (1985) [hereinafter Newberg on Class Actions]. Finally, the savings in time borne by use of a percentage approach would "reduce the delay period between the settlement of a common fund case and the award of fees to counsel." Waldner, 1989 WL 100184, at *5-6.

As with the lodestar approach, a percentage approach might also create some perverse incentives. For example, counsel might push for an early settlement simply to ensure compensation or to gain a larger fee, Chesny v. Marek, 720 F.2d 474, 477 (7th Cir.1983). Moreover, percentage awards, if not aligned to the work actually done by counsel, run the risk of being excessive. Nevertheless, these concerns are outweighed by the need to avoid the tremendous amount of time and energy required by the lodestar approach, to cultivate efficiency by the attorneys, and generally to simplify and make more predictable the process of awarding attorneys' fees. Given these advantages and the trend in this circuit and many others, the court is indeed inclined to employ the percentage approach.

2. Percentage v. Percentage

It might be more accurate to say that the court is inclined to employ a percentage approach. This is because there is no consensus on how to determine a reasonable percentage. See Brown, 838 F.2d at 454.^[28]

The Task Force recommended that a percentage fee arrangement be negotiated between plaintiffs and their attorneys as early as practicable in the litigation. Task Force Report, 108 F.R.D. at 255. See also Shalala, 1 F.3d at 1269 (recommending the approach of the Task Force). The Task Force made further recommendations to facilitate negotiations and ensure fairness. For example, it recommended that the court appoint a negotiator "who will vigorously represent the interests of the [class]," and who will negotiate "in an open and appropriately arm's length manner" a reasonable compensation plan; the court should then thoroughly review this plan. Task Force Report, 108 F.R.D. at 256-57.^[29]

Like contingent fee agreements generally, a negotiated agreement would infuse simplicity, certainty, and fairness into the attorneys' fees process. Regrettably, no such agreement was negotiated in this case.^[30] Nor was there a negotiated agreement in In re Bioscience, 155 F.R.D. at 118-19. In that case, Judge Dalzell appointed Judge Adams as a special master to determine what contingent fee arrangement would have been negotiated "in this District for legal services in business litigation similar to this Consolidated Action." Id. at 119. Judge Adams subsequently concluded that a 30% fee award was appropriate under the circumstances of that case. Id. (citation omitted).

The court is not convinced that, as suggested by class counsel, a negotiated percentage between the settlement class and counsel would have been at least 7.5%. While it is likely that the parties would have agreed upon a rate in the range of 7.5%-12.5%, a different rate could reasonably have resulted from such negotiations. The negotiated percentage could have been as high as 15%. However, a negotiated agreement might have provided that a recovery of $111 million — roughly 65% of what the settlement class would have received if they had won a full judgment — warranted only 5% attorneys' fees.

Moreover, this post hoc approach, while still simpler than reverting to a lodestar analysis,^[31] defeats a major purpose of the negotiated percentage. The negotiated percentage draws its strength from the fact that it is negotiated beforehand, while the results of the litigation are unknown. This helps ensure fairness to both counsel and plaintiffs. A post-settlement determination of what agreement would have been negotiated, even if done carefully, jeopardizes this fairness. See First Fidelity, 750 F.Supp. at 163 ("For counsel to seek a contingent fee after the matter has been resolved is akin to placing a wager after the outcome of event is known or playing poker with everyone's cards face up.") See also Task Force Report. 108 F.R.D. at 258 ("[T]he advantage of the negotiated fee procedure will be entirely undermined if, at the end of the litigation, counsel have the right to renegotiate depending upon ... factors relating to the case.") Despite admiring Judge Dalzell's pragmatism in forging a solution, the court prefers not to engage in this form of hindsight.^[32]

Other courts employ different approaches in determining a reasonable percentage. Some judges note a range within which attorneys' fees typically fall. See, e.g., GNC Shareholder, 668 F.Supp. at 452 (noting that no general rule exists for awarding a reasonable percentage, court considers fact that "the normal range of common fund securities and antitrust suits is between 20 and 30 percent"); SmithKline Beckman, 751 F.Supp. at 533 (fee awards under percentage approach typically range from 19% to 45% of settlement fund); In re TSO, 1989 WL 80316 at *7 (awards in other common fund cases range between 19% and 45% of fund); Sala, 128 F.R.D. at 214-15 (listing ranges of percentages found by other courts and commentators); Paul, Johnson, Alston & Hunt v. Graulty, 886 F.2d 268, 271 (9th Cir.1989) ("25 percent has been a proper benchmark"); Weseley v. Spear, Leeds & Kellogg, 711 F. Supp. 713, 718 (E.D.N.Y.1989) ("[i]n this circuit, fees typically range from 15% to 30% of the recovery"); In re Warner Communications Securities Litigation, 618 F. Supp. 735, 749 (S.D.N.Y.1985) ("Traditionally, courts in this circuit and elsewhere have awarded fees in the 20% to 50% range in class actions"), aff'd, 798 F.2d 35 (2d Cir. 1986); In re Activision, 723 F.Supp. at 1377 ("absent extraordinary circumstances ... the rate should be set at 30%"). See also Attorney Fee Awards, supra, § 2.07 at 51 (the majority of common fund fee awards fall between 20%-30% of the fund); Rescuing the Private Attorney General, supra, 42 Md. L.Rev. at 239 (typical award falls between 20% and 35% of fund).

Class counsel argue that a percentage award of 7.5% is "well below the typical range of awards of 20-45% in complex class actions cases." Memorandum in Support of Joint Petition of Counsel for an Award of Attorneys' Fees and Reimbursement of Litigation Expenses at 16. They list in an appendix 37 cases in which the award of fees equaled or exceeded 33% of the fund. Id. at Appendix 1.

Awarding 20 to 45 percent of the $111 million fund in this case would be manifestly unjust:

Fixed percentages will drastically overcompensate lawyers in some cases and drastically undercompensate them in others. Twenty-five or thirty percent might be an appropriate award on a recovery of a million dollars. It is likely, however, to result in a windfall where the recovery totals many millions of dollars.

Superior Beverage, 133 F.R.D. at 124 (N.D.Ill.1990). See also Task Force Report, 108 F.R.D. at 256 (the percentage should "decrease as the size of the fund increases"). Moreover, justifying an award of 7.5% simply because it is significantly less than 20% would be unsatisfactory. Instead, the court would need to determine an appropriate range. The number of cases involving a common fund in the neighborhood of $111 million is relatively small. Initial research reveals that in cases where the fund is between $100 and $200 million, fees typically range from 4%-10%.^[33] Even if an appropriate range were determined, however, the court would also need to decide what percentage within the range to award.

Some courts have considered factors other than the results in other cases when determining a percentage.^[34] For example, in GNC Shareholder, the court found 25% to be a reasonable percentage based in part on the following factors: counsel's experience in class action securities litigation; counsel's competency; counsel acted in best interest of class; and a significant settlement fund was obtained at an early stage of litigation. 668 F.Supp. at 452. In Edmonds, the court considered, inter alia, the novelty of the issues and the vigorous opposition of the government. 658 F.Supp. at 1141-42. In Mashburn, the court took into account the lack of substantial objections by the class and the non-monetary benefits conferred upon the class by the settlement. 684 F.Supp. at 692-93. Other courts have considered these and/or other factors. See, e.g., Camden I, 946 F.2d at 775; First Fidelity, 750 F.Supp. at 163. In re TSO, 1989 WL 80316, at *6-7. "In most instances, there will ... [arise] factors unique to a particular case which will be relevant to the district court's consideration." Camden I, 946 F.2d at 775.

Equitable factors such as these should play a role in many cases where reasonable attorneys' fees are awarded under a percentage approach. But they do not on their own suggest an appropriate percentage. Rather, they must be applied to a percentage or range of percentages that is an appropriate starting point or benchmark.

One option which readily suggests itself is to apply these equitable factors to a range based upon the results in other cases. See, e.g., GNC Shareholder, 668 F.Supp. at 452 (range of 20% to 30% apparently used as a starting point); Sala, 128 F.R.D. at 215-16 (starting with range of 25%-30% and awarding roughly 32% of fund after noting the skill and efficiency of counsel); SmithKline Beckman, 751 F.Supp. at 533-34 (1990) (starting with range of 20%-27% and awarding 25% of fund after noting the skill and efficiency of counsel); In re TSO, 1989 WL 80316, at *7 (noting range in other cases of 19%-45% and then finding request for 23% of fund "wholly justified" in light of the results of litigation and efficiency of counsel).

However, even if a range appropriate for a $111 million settlement were used as a starting point for applying other factors, this approach would be unsatisfactory. Such an approach might still prove unfair in that it could greatly undercompensate or overcompensate the attorneys for the amount of work they have actually done on the case. While the court would be willing to risk foregoing the close alignment of an attorneys' efforts to the award of fees for the sake of fairness and simplicity in the case of a negotiated percentage, the court will not do so in the instant litigation by simply granting a percentage that is "in the ballpark" of previously awarded fees.^[35]

In the absence of a previously negotiated percentage, the court will determine a reasonable percentage by incorporating a modified lodestar analysis, less exacting and time-consuming than a full-scale lodestar analysis. Under this approach, the court will calculate a range of fees that would result from a full-scale lodestar analysis. The court will then determine a reasonable percentage based upon the lodestar range as well as other, equitable considerations.

3. Percentage + Lodestar

Based upon an extensive review of cases in which attorneys' fees were awarded out of a common fund, the court concludes that combining the percentage and lodestar approaches is consistent with the way in which courts calculate reasonable attorneys' fees. To begin with, the outcomes of these fee determinations are instructive. Courts that employ the percentage approach appear to be motivated in part by a lodestar dynamic. Because courts are reluctant to give fee awards totally incommensurate with the efforts of the attorneys, percentage awards generally decrease as the amount of recovery increases. See, e.g., SmithKline Beckman, 751 F.Supp. at 534 (sliding scale implicit in many decisions in which percentage approach is applied); Sala, 128 F.R.D. at 215 (recognizing and approving of this "inverse relationship between fund and fee"); In re TSO, 1989 WL 80316, at appendix (only cases with total awards below $12 million showed fee awards of 25% or above); In re Domestic Air Transportation Antitrust Litigation, 148 F.R.D. 297, 352 (N.D.Ga.1993) (though the benchmark in common fund cases is 20%-30%, fee awards usually fall in the 13%-20% range for funds of $51-$75 million, and in the 6-10% range for funds of $75-$200 million; fee award of 5.25% of roughly $300 million "adequately accounts for the economies of scale in class actions of this size"); First Fidelity, 750 F.Supp. at 163 (sliding scale used to determine fee award: 30% of the first $10 million, 20% of the next $10 million, and 10% of all monies over $20 million). As Judge Sarokin noted in First Fidelity, "[t]here is considerable merit to reducing the percentage as the size of the fund increases. In many instances the increase is merely a factor of the size of the class and has no direct relationship to the efforts of counsel." Id. at 164 n. 1. Accordingly,

the fee percentage would be significantly more modest as the common fund recovery begins to reach recoveries approaching or exceeding $100 million, based on the notion that the effort necessary to achieve recovery dollars at the high end was less onerous, on a sliding scale, than the effort expended for recovering the threshold sums by judgment or settlement.

Newberg on Class Actions, supra, § 14.03 at 190.^[36]

A similar phenomenon is evident in cases where the lodestar approach is used. The outcomes in many of these cases suggest a concern with how much of the common fund will remain after attorneys' fees are deducted. As the proportion of the lodestar to the common fund increases, courts are less willing to employ large multipliers and transfer money from the class to the attorneys.^[37] Conversely, when this proportion is very small, courts often grant large multipliers. Thus, the state of attorneys' fees is such that one judge can heap lavish praise on counsel in three separate cases, yet use multipliers of 8.4, Muchnick v. First Federal Sav. & Loan Ass'n of Philadelphia, No. CIV.A. 86-1104, 1986 WL 10791, at *1 (E.D.Pa. Sept. 30, 1986) ($250,000 award of fees — between 3.7% and 6.25% of fund — justified under lodestar method even though lodestar is only $29,732.50); 1.78, Sherin v. Smith, [1987-88 Transfer Binder] Fed.Sec.L.Rep. (CCH) ¶ 93,582, at 97,608, 1987 WL 18851 (E.D.Pa. 1987) (award is 28% of settlement fund); and 0.87, In re Fiddler's Woods Bondholders Litigation, [1987-88 Transfer Binder] Fed.Sec. L.Rep. (CCH) ¶ 93,537, at 97,409-10, 1987 WL 19239 (E.D.Pa.1987) (where lodestar method is used and lodestar was $2,439,321, reasonable fees award is $2,116,500, or 32.7% of the settlement fund). In short, "lodestar analyses in practice tend to result in a bottom line that approximates a fair percentage." Howes v. Atkins, 668 F. Supp. 1021, 1027 (E.D.Ky.1987) (citing Understanding the Plaintiff's Attorney, supra, 86 Colum.L.Rev. at 678-79 n. 26).

Not only do the results suggest a connection between the two approaches, but the very methodology of the approaches often provides for both lodestar and percentage considerations as well. Many courts determine an award using one method, convert that award into terms of the other method, and then eyeball the second figure to ensure that the award would be appropriate under either method. See, e.g., Eltman v. Grandma Lee's, Inc., [1986-87 Transfer Binder] Fed.Sec.L.Rep. (CCH) ¶ 92,798, at 93,907, 1986 WL 53400 (E.D.N.Y.1986) (after calculating lodestar with a multiplier of 2, court finds that resulting award of slightly less than 30% of fund fell reasonably within range of 20% to 50% awarded in similar cases); Golden v. Shulman, [1988-89 Transfer Binder] Fed.Sec.L.Rep. (CCH) ¶ 94,060, at 90,953, 1988 WL 144718 (E.D.N.Y.1988) (calculating lodestar, applying multiplier, and finding that 30% is within appropriate range for percentage awards and thus justified); Brewer v. Southern Union Co., 607 F. Supp. 1511, 1536 (D.Colo.1984) (fee award calculated under lodestar approach amounts to 13.97% of fund, a percentage "well within the range of awards in similar cases"); Steiner, 835 F.Supp. at 791 (fee award calculated under lodestar amounts to 24.5% of the fund, a percentage consistent with other awards where the funds were of similar size); Meshel v. Nutri/System, Inc., 102 F.R.D. 135, 140 (E.D.Pa.1984) (fee calculated under lodestar approach amounts to 25% of fund, which is within the range of fee recoveries in class actions); Muchnick, 1986 WL 10791, at *4 ("low percentage of the fund which this [lodestar] fee represents confirms my conclusion that the fee is a reasonable one"); Superior Beverage, 133 F.R.D. at 133; Sala, 128 F.R.D. at 216 ("in order to avoid allowing the kind of windfall fee awards which depressed support for the percentage of recovery method in the first place," court finds that a similar result would also have resulted under lodestar approach); In re Crazy Eddie Securities Litigation, 824 F. Supp. 320, 327 (E.D.N.Y.1993); First Fidelity, 750 F.Supp. at 164 (court admits to "manipulat[ing] the [multiplier] so as to arrive at the pre-ordained fee," i.e. the previously calculated percentage award, in determining that a multiplier of 2.5 would be appropriate).^[38]

Moreover, some courts calculate an award under one approach, and then undertake a second calculation using the other approach to demonstrate the reasonableness of the award. See, e.g., Edmonds, 658 F.Supp. at 1143 (out of "an abundance of caution," court will also calculate fees using lodestar approach); Mashburn, 684 F.Supp. at 698 (same); In re Public Service Co. of New Mexico, [1992 Transfer Binder] Fed.Sec. L.Rep. (CCH) ¶ 96,988, at 94,290, 1992 WL 278452 (S.D.Cal.1992) ("in light of the magnitude of [the] settlement fund," court utilizes each approach); Domestic Air, 148 F.R.D. at 356-57 (undertakes lodestar analysis to help determine reasonableness of percentage award).

Even when courts focus on only one of the two approaches, the approach chosen is seldom unrelated to the other. The lodestar approach typically entails the classical percentage consideration. Courts in this circuit consider percentages in assessing a contingency enhancement. For example, the Lindy court cautioned that the lodestar might be "so large a proportion of the total recovery that [a contingency enhancement] would be minimal." Lindy I, 487 F.2d at 168. As one court in this district explained in finding that a contingency multiplier was inappropriate under the circumstances of the case, "[a]n upward adjustment would make the Settlement Fund a fund for counsel for the class rather than a fund for the class." Fickinger v. C.I. Planning Corp., 646 F. Supp. 622, 635 (E.D.Pa.1986). See also Baughman v. Wilson Freight Forwarding Co., 583 F.2d 1208, 1218-19 (3d Cir.1978) (where lodestar had been multiplied by 2 to account for contingency and the resulting fees were more than two-thirds of the settlement fund, Court of Appeals for the Third Circuit instructed the trial court to consider the size of plaintiff's award in comparison to the initial lodestar calculation before awarding an enhancement for contingency); Zeffiro v. First Pennsylvania Bank, N.A., 574 F. Supp. 443, 450 (E.D.Pa.1983) ("[I]t would be an abuse of this Court's discretion to adjust the lodestar for contingency given that the lodestar ... is a significant amount in comparison to the total amount of the settlement fund.").

Other courts have made similar percentage considerations in determining an adjustment to the lodestar. See, e.g., Warner Communications, 618 F.Supp. at 747 (a proper factor is "the requested fee in relation to the settlement")^[39]; Van Gemert v. Boeing Co., 516 F. Supp. 412, 420 (S.D.N.Y.1981) ("`[recognizing] that the amount of fees must bear a reasonable relationship to actual recovery ...'" the court awards a multiplier of 1.7.); In re AIA Industries, Inc. Securities Litigation, No. CIV.A. 84-2276, 1988 WL 33883, at *4 (E.D.Pa. March 31, 1988) (in determining multiplier, "a court should take into account what percentage of the settlement the lodestar figure represents"; no enhancement warranted since combined expenses and fees constitute over 40% of fund); Bebchick v. Washington Metropolitan Area Transit Com'n, 805 F.2d 396, 406-07 (D.C.Cir.1986) (first factor supporting enhancement of lodestar was fact that this was a common fund case, for which a percentage approach is permitted, and that the requested enhancement of 60% would result in an award which was 25% of the fund, a reasonable percentage); State of Florida v. Dunne, 915 F.2d 542, 546 (9th Cir.1990) (in applying the lodestar approach, district court erred in not considering the effect of other pending fee applications, namely that 72% of the common fund could be distributed to attorneys). In light of this phenomenon, it may be concluded that "even judges who use [lodestar] analyses always blend lodestars and percentages." Superior Beverage, 133 F.R.D. at 126.

Similarly, the percentage approach currently employed by courts often entails the lodestar calculation. Some courts use the traditional lodestar factors — including the labor expended — to determine a reasonable percentage. See, e.g., Brown, 838 F.2d at 454-55 (the Johnson factors, including time and labor involved and customary fee, are appropriate in determining percentage fee awards in common fund cases)^[40]; Camden I, 946 F.2d at 775 (same); Basile v. Merrill Lynch, Pierce, Fenner, & Smith, Inc., 640 F. Supp. 697, 700 (S.D.Ohio 1986) (in Sixth Circuit, factors to be considered in determining reasonable percentage include the value of the services on a hourly basis); Crazy Eddie, 824 F.Supp. at 326-27 (E.D.N.Y.1993) (one of factors justifying 33.8% award is that the award is "a reasonable multiple of the hours spent").

Clearly, then, courts may use the lodestar, or an adjusted lodestar, as a starting point in determining a final percentage. In this regard, the court finds particularly relevant the case of Diamond v. Fogelman, [1992 Transfer Binder] Fed.Sec.L.Rep. (CCH) ¶ 96,980, 1992 WL 203779 (E.D.N.Y. 1992).

In December 1991, the Board of Judges of the Eastern District of New York adopted the Civil Justice Expense and Delay Reduction Plan ("the Plan"). According to the Plan:

[i]n cases that settle after significant attorney time has been expended, [the] fee award shall be based on a percentage of recovery; but the attorneys shall be required to submit time records, as is required under the lodestar approach, which shall serve as a guideline for the court in setting the percentage recovery.

Plan, V(A). In Diamond, Judge Sifton used this approach. [1992 Transfer Binder] Fed. Sec.L.Rep. (CCH) ¶ 96,980, at 94,243-44. The court found that approximately 75% of the hours listed in the petitions by counsel were reasonably spent on the litigation; found that a multiplier of 1.9 applied to the lodestar would result in the fee, constituting 25% of the fund, requested by counsel; discussed how the lodestar adjustment factors applied to counsel; found that fees in the Second Circuit range from 15% to 30%; and awarded counsel 20% of the fund.^[41]Id. at 94,244.

By utilizing a modified lodestar analysis as a first step towards determining a reasonable percentage, this court will be employing an approach similar to Diamond and to other cases where the lodestar played a prominent role in the percentage calculation. The specific methodology might be novel: the court is not aware of any other case in which lodestar principles were employed to calculate a lodestar range as a benchmark for a percentage determination. However, as demonstrated, this court will be far from the first to utilize both the lodestar and percentage approaches in a single determination of an award of attorneys' fees.

More importantly, the approach to be employed by this court incorporates some of the valuable features of the traditional percentage and lodestar approaches. As one court concluded: "[R]igid application of any method, whether lodestar or percentage, is a mistake ... [rather,] some form of time/rate multiplier analysis modified ... to reflect a reasonable percentage of the recovery is the best way to calculate reasonable and fair fees." Superior Beverage, 133 F.R.D. at 124. This court's approach, consistent with the emerging preference of courts to return to determining a reasonable percentage as an award, reflects the percentage approach's relative simplicity when compared to a full-scale lodestar approach,^[42] and can well account for the recovery by the settlement class achieved by counsel. At the same time, through its modified lodestar analysis, this approach ensures fairness by reflecting the actual work done by counsel in producing the common fund.^[43] Finally, by considering the equities both in the multiplier segment of the lodestar analysis and in its final calculation of a reasonable percentage, the court is able to make an award based on all the circumstances involved in this case. See GNC Shareholder, 668 F.Supp. at 451 ("Determination of what is a fair and reasonable fee award from a common fund ... calls for the informed exercise of judicial discretion under all the circumstances.") (citation omitted).^[44] In sum, "the composite use of percentage and loadstar ... allow[s] the judge great[] insight into the reasonableness of the fee requested." Diamond v. Fogelman, [1992 Transfer Binder] Fed.Sec.L.Rep. (CCH) ¶ 96,873, at 93,562 (E.D.N.Y.1992).^[45]

It remains for the court to undertake the task of applying the method it has chosen. This task is far from an exact science — such is the nature of awarding attorneys' fees^[46] — but is one the court believes will produce the just result.

C. APPLYING THE METHOD

1. Hours Expended

The court begins its calculation of a lodestar range for class counsel^[47] by determining the number of hours for which the attorneys should be compensated from the fund. The total number of hours for which compensation is requested is 20,470.^[48]

The burden of proving reasonable attorneys' fees is on the party seeking them. Rode v. Dellarciprete, 892 F.2d at 1183. Class counsel from each firm have submitted billing summaries which they have compiled from contemporaneous time records maintained by each attorney and paralegal who worked on this case. See Hensley, 461 U.S. at 437, 103 S.Ct. at 1941 (party seeking fees should support the petition with contemporaneously prepared records of the time spent on each task).

Some of these summaries were more vague or confusing than others. It is most helpful to the court if a firm lists each activity separately and describes it with sufficient detail to inform the court of its nature: to what did the "phone call" pertain? What was the subject of the "legal research"? As the court in Lindy I explained, "without some fairly definite information as to the hours devoted to various general activities ... and [to] the various classes of attorneys, ... the court cannot know the nature of services for which compensation is sought." 487 F.2d at 167. The court recommends that the clear and precise format suggested by Legalgard — replicated by some of the petitioning firms — be employed in future cases to better enable courts to assess the petitions. See also Federal Judicial Center, supra, at 127 (example of billing statement from U.S. Trustee Guidelines). Generally, however, the records are satisfactory for the "petite lodestar" task at hand. See Pawlak v. Greenawalt, 713 F.2d 972, 978 (3d Cir.1983) ("The court's focus in assessing the adequacy of submitted documentation ... is whether the documentation permits the court to determine if the claimed fees are reasonable.")

Unisys initially complains that class counsel spent too much time on this case. As evidence of this, defendant points out that class counsel's time charges roughly triple its own counsel's time charges. Defendant also finds relevant the fact that 80 attorneys and paralegals were used by class counsel in this case.

Neither of these facts move the court. First of all, it is unclear whether the actual hours spent by class counsel, as opposed to the time charges, really triple those spent by Unisys's counsel. Class counsel correctly point out that Unisys did not submit its counsel's billable hours or rates, and also failed to include the time spent by in-house counsel to assist with the case.^[49]

Even assuming, however, twice or three times as many hours were spent by class counsel, this is not dispositive of the necessity of those hours. Class counsel, unlike Unisys, had to start virtually from scratch to understand the facts and history relevant to the incentive retirees' claims. This entailed sifting through thousands of documents and acquiring information from hundreds of retirees regarding approximately 100 different retirement plans. Moreover, the claims which class counsel needed to establish were very fact-intensive, entailing the composition of evidence from this huge number of sources, whereas Unisys' defense was primarily legal in nature. Additionally, class counsel dealt with thousands of clients, as opposed to defense counsel's one client, and a substantial amount of time was expended explaining to them the ramifications of the litigation.

Class counsel's use of over 80 attorneys and paralegals is also justified. "[I]n cases with a number of collaborating lawyers, the district court assesses claims of overstaffing and overresearching in light of the `scope and complexity of the particular litigation.'" Planned Parenthood of Southeastern Pennsylvania v. Casey, No. CIV.A. 88-3228, 1994 WL 675260, at *5 (E.D.Pa. Nov. 21, 1994) (quoting New York State Ass'n for Retarded Children v. Carey, 711 F.2d 1136, 1146 (2d Cir.1983)). Here, over 80% of the total hours were billed by only nine attorneys. Many attorneys and paralegals spent small amounts of time on the case preparing this case for trial on behalf of 32,000 plaintiffs in just eleven months. The use of so many people was integral to the quick resolution of the litigation for the settlement class.^[50] The use of these additional employees also allowed simpler tasks to be delegated to lawyers and paralegals charging lower rates than the lead attorneys.

The number of hours and attorneys is thus not unreasonable when compared to the deployment of defense counsel. However, 20,000 hours is, to use the vernacular, "a lot" of hours, and it remains for the court to determine whether some of these hours should be eliminated from the compensable total for other reasons.

The number of hours for which counsel may be compensated "excludes hours that are excessive, redundant, inadequately documented, or otherwise unnecessary." Lugo v. Williams, No. CIV.A., 1994 WL 45100, at *2 (E.D.Pa. Feb. 15, 1994) (citing Hensley, 461 U.S. at 433-34, 103 S.Ct. at 1939-40). Based upon findings by Legalgard and extensive spot-checking, the court finds that 150-300 hours must be deducted from the lodestar for vagueness, irregularity, or excessiveness.^[51] These deductions pertain, for example, to accidental entries of 27-hour work days, duplicate entries, entries where more time was expended than was reasonably necessary to perform the task at hand,^[52] and many entries that do not adequately inform the court about the task being performed.

Some time must be deducted as unnecessary because it was not expended on behalf of the settlement class. All time spent in this case falls into one of three categories:

Category One: hours expended exclusively on behalf of the settlement class;

Category Two: hours expended exclusively on behalf of plaintiffs other than the settlement class;

Category Three: hours expended on behalf of both the settlement class and other plaintiffs.

Category One hours, if reasonable and necessary, should be assessed against the settlement fund. This time consists primarily of (1) time after February, 1994 spent negotiating a settlement agreement which would include only the settlement class; (2) time leading up to and including the trial spent on factual work unique to the settlement class; (3) time in the late summer and early fall of 1993 researching the law supporting a possible summary judgment motion on the settlement class's behalf; (4) time spent writing the portions of the post-trial briefs dedicated to making the special arguments on behalf of the settlement class; and (5) time spent communicating with members of the settlement class about the status of the case.

Category Two hours should not be assessed against the settlement fund. Each of the petitioning firms followed guidelines developed by lead counsel in eliminating Category Two hours. These guidelines required exclusion from the petitions of time spent on work that advanced claims of Unisys class members and Sperry and Burroughs regular retirees ("the non-settling class") and that would not have been done if only settlement class claims were being litigated. See Planned Parenthood, 1994 WL 675260, at *7 (in calculating an appropriate reduction for limited success, "the Court considers the fact that much of the legal services would have been necessary even if the [one fully successful claim] were the only claim litigated"). Time eliminated from the petitions by class counsel under this rubric consists primarily of (1) all work on claims of the non-settling class after the tentative settlement was reached on April 14, 1994; (2) work on Unisys class claims^[53]; (3) meetings and phone calls with members of the non-settling class unrelated to facts also pertinent to the settlement class; (4) pleadings and interrogatory answers that would not have been drafted if only settlement class claims were being litigated; (5) all work relating to recordkeeping and the fee petition.^[54]

This process resulted in the elimination from most petitions of 7%-18% of the total hours expended.^[55] The court finds that the firms did a reasonable job in this regard. A few firms, however, failed to eliminate some of these hours. For example, small amounts of time spent on the fee petitions and on Unisys class claims were not eliminated. Accordingly, the court will deduct 55-70 hours.

However, an additional reduction of Category Two hours is warranted. Many of the firms have included in their fee petitions time spent on the summary judgment motion.^[56] Class counsel contend that the legal issues addressed in the summary judgment motion would have come before the court at some stage — a motion for summary judgment or post-trial motion — had the settlement class alone brought this action. They add that the summary judgment motion also provided a legal framework for the remainder of the case. They further contend that fact investigation, preparation of witness affidavits, and drafting of proposed undisputed facts laid a factual foundation for the settlement class at trial. They argue that the summary judgment time thus jointly benefitted the settlement class and the non-settling class, i.e. that these hours are Category Three hours.

In determining whether these hours are Category Three hours, however, the court must distinguish between (1) time expended on behalf of the settlement class that would have been expended even had the settlement class proceeded alone, and (2) time expended on behalf of groups other than the settlement class that would have been expended even had the settlement class proceeded alone. At a minimum, time to be assessed against the settlement fund must have actually been expended on behalf of the settlement class. Thus, the court must determine not only whether the time spent on summary judgment would have been expended on behalf of the settlement class if the case had been brought only by that class, but also whether this time is otherwise properly considered as time "expended on behalf of" the settlement class.

The court finds that time spent researching the legal issues, drafting the response, and arguing the motion should largely be excluded. Although there is never certainty as to what would have happened, in all likelihood Unisys would have made the same arguments that it made in the summary judgment motion if the case had been brought only by the settlement class.^[57] But the fact that class counsel under a different scenario would have responded to a motion directly against the settlement class does not change the fact that this scenario did not occur. The court must thus address whether these hours can nevertheless be properly considered as hours expended on behalf of the settlement class.

It is true that class counsels' efforts in responding to the motion did provide some benefit to the settlement class. The partial success responding to the motion undoubtedly helped prevent Unisys from successfully bringing a similar motion against the settlement class. Furthermore, the response did help to create a legal framework for the trial and post-trial briefs, and presumably clarified the issues and affected both sides' bargaining power at settlement negotiations.

However, these effects were incidental to the actual focus of those hours, namely to keep the case alive for the non-settling class. It would not be completely accurate to say that the time spent formulating the legal arguments and preparing and arguing the opposition to summary judgment was expended on behalf of the settlement class. It thus would be unfair to require the settlement class to pay in full or in large part for class counsel's efforts to preserve the case for the non-settling class. For these reasons, the court deems 75% of the time spent researching the legal issues and preparing and arguing the actual response to the summary judgment as Category Two hours.

The benefits to the settlement class from the factual work done during summary judgment hours — investigation, affidavit preparation, and drafting of undisputed facts — were more direct. First, although Unisys ultimately did not agree to the establishment of undisputed facts at this stage, such facts were eventually offered by class counsel in pre-trial or post-trial proposed findings of fact. The drafting of undisputed facts at the summary judgment stage thus provided much of the legwork for these later findings.

Second, many of the facts dealt with at this stage were established at trial and would have been used even had the settlement class proceeded alone.^[58] The factual work done for summary judgment and then presented at trial pertained to regular retirees. However, because the incentive retirees had SPDs identical or similar to those of the regular retirees, the settlement class claimed that they were entitled to their pre-October, 1992 post-retirement medical benefits under the same promises that applied to the regular retirees. Indeed, a victory for the regular retirees on their contract, breach of fiduciary duty or estoppel claims ultimately would have meant victory for the incentive retirees as well. Moreover, Unisys argued from the outset of the case that, because of the RORs, the incentive retirees were entitled to the same benefits as the regular retirees, no more.

Thus, class counsel needed to present evidence demonstrating the pervasive practices by Unisys toward regular retirees as part of the case for the incentive retirees. A large percentage of the factual work done for summary judgment would have been done to enable the incentive retirees to make this presentation if they had proceeded alone. Therefore, a large percentage of this factual work may properly be considered as time expended on behalf of both the non-settling class and the settling class. Therefore, a large percentage of this factual work may properly be considered as time expended on behalf of both the non-settling class and the settling class. Therefore, the court deems 75% of summary judgment factual work as Category Three time. Accordingly, 25% of the summary judgment "fact" hours are Category Two hours not compensable out of the settlement fund.

To make an appropriate reduction of lodestar hours, the court must determine how many hours were spent by class counsel on the legal research, drafting and oral argument pertaining to the response to the summary judgment motion. The court must also determine how many hours were spent on the factual workup for this response. Under a full-scale lodestar analysis these determinations would require careful scrutiny of each of the entries in the petition, and further cataloging by firms of hours spent on each activity.^[59] For the purposes of estimating a lodestar, however, the court performed a preliminary analysis of the petitions and also received a submission by class counsel estimating this breakdown. The court finds that 900-1000 hours included in the petitions were expended on legal research, drafting of briefs, and oral argument for the summary judgment response, and that 800-900 hours were expended on factual investigation for this response. Accordingly, excluding 75% of the former and 25% of the latter, the court concludes that 875-975 hours must be deducted.

A further reduction of Category Two hours is necessary regarding discovery, the trial, and settlement negotiations. Most of the time spent by class counsel during these stages benefitted the settlement class, either by its sole focus on the incentive retirees, or by its contribution to the regular retirees' claims.

However, the court finds that some of this time, though benefitting the settlement class, may not be properly considered as time expended on behalf of the settlement class. This is because not all of this time would actually have been spent by class counsel had the case been brought solely by the settlement class. Although a full-scale lodestar analysis would warrant further debate by the parties involved, the court has examined these issues and finds that a case brought only by the settlement class, consisting of two sub-classes, would not have demanded as much negotiation as that done in the instant case for four additional subclasses, or as much discovery or trial work as that done for two additional subclasses.^[60] For example, settlement negotiations would have focused earlier on the special claims of the incentive retirees, fewer documents would have been collected and reviewed, and fewer witnesses and other evidence would have been produced at trial.

Class counsel dispute this. For example, regarding their trial work, they argue that because the settlement class needed to demonstrate that their benefits were systematically promised to them "for life," "even if the trial had concerned only the incentive retirees' claims, class counsel still would have had to present the same pattern and practice evidence (from the testimony of incentive and regular retirees, policy makers and counselors, and thousands of pages of documents concerning both the incentive programs and regular plans)." Reply of Class Counsel to Memorandum of Dale Nathan at 9.

While the settlement class proceeding on their own would indeed have needed to establish something akin to a "pattern and practice" of guarantees by Unisys, the court is of the opinion that not all of the evidence pertaining to the "regular retirees" that was actually used would have been developed before trial and used at trial to make this showing. First of all, a significant portion of the "regular retiree" evidence could have been provided — in fact, some of it was provided at the actual trial — by the incentive retirees, who also had the same type of SPDs. Second, the factual development and presentation regarding the SPDs would have been less robust for the settlement class, which had additional and separate claims available to it, than it was for the actual regular retirees, for whom the SPDs and related evidence was the basis for all of their claims.^[61]

Accordingly, the court deems 15%-20% of discovery hours, 15%-20% of settlement hours, and 10%-15% of trial hours as Category Two hours.^[62] Using the same methodology to determine the total hours spent at each stage as that used regarding summary judgment, the court finds that 7000-7250 hours were spent on discovery, 2000-2500 hours were spent on settlement negotiations, and 6750-7000 hours were spent on trial activities. Accordingly, 1050-1450 discovery hours, 300-500 settlement hours, and 675-1050 trial hours are Category Two hours not compensable from the settlement fund.

The court must next determine whether any Category Three hours should be eliminated from the petitions. Class counsel argue that none of this time benefiting both the settlement and non-settling classes should be eliminated since it would have been necessary even if the case had been brought by the settlement class alone. As previously mentioned, these hours would generally have been necessary because part of the incentive retirees' claims — the claims based on the SPDs — were the very claims made by the regular retirees.

Mr. Nathan, Unisys, and counsel from Miller, Faucher object that class counsel have not prorated these hours benefitting both the settling and non-settling classes based on the percentage of the overall classes represented by the settlement class. Mr. Nathan would reduce Category Three hours by 75% because the settlement class comprises 25% of the plaintiff class. Counsel from Miller, Faucher espouses a similar rule; in their petition concerning work only for the Sperry class, they would reduce the Category Three hours by 65% because the Sperry incentive retirees constitute 35% of the Sperry class. Unisys would also reduce Category Three hours, but does not suggest what reduction would be appropriate.^[63]

It is clear that counsel may not be compensated for work expended solely on behalf of unsuccessful plaintiffs. Category Three hours, however, were expended on behalf of both successful and unsuccessful plaintiffs. Class counsel point to three cases from other circuits, all of which support their contention that such time should be compensated in full. The court is not aware of any opinion reaching a different conclusion.

In Wooldridge v. Marlene Industries Corp., 898 F.2d 1169 (6th Cir.1990), the defendant was held liable for maintaining a discriminatory maternity leave policy, but only 38 out of 131 backpay claimants were awarded backpay. The Court of Appeals for the Sixth Circuit held that time spent by counsel on the liability portion of the trial was fully compensable because "the time spent litigating issues common to class members benefitted those who received backpay awards as well as those who did not receive backpay." Id. at 1175. The court made a similar ruling concerning time spent on damages claims of non-prevailing class members:

The District Court may ultimately discover that much of the time spent by the attorneys in the damages phase was spent on issues common to both the successful and unsuccessful plaintiffs. To the extent that such time aided or was reasonably calculated to aid the claims of the successful plaintiffs it is compensable — even though it may also have been performed to aid those plaintiffs who did not receive backpay. The guiding principle is that the work must in some way have been expended in pursuit of a successful claim for backpay.

Id.

In Duke v. Uniroyal, 928 F.2d 1413 (4th Cir.1991), cert. denied, 502 U.S. 963, 112 S. Ct. 429, 116 L. Ed. 2d 449 (1991), two of four plaintiffs, Fox and Duke, prevailed in an ADEA suit. The Court of Appeals for the Fourth Circuit affirmed the district court's 25% reduction of time spent on post-trial motions. Id. at 1425. Part of this 25% was spent on motions for reconsideration on behalf of the two non-prevailing plaintiffs, and the district court found that "time spent on these motions was not related to the success of Fox and Duke in their ADEA claims and is not compensable." Id. (citation omitted). Although only two of the four plaintiffs prevailed, the district court found that an overall reduction of 50% would not be appropriate because "it would not be accurate to say that the plaintiffs could have prosecuted the claims of Duke and Fox with only half of the labor that they expended in prosecuting the claims of [all four plaintiffs]." Id. (citation omitted). With these statements, the district court implied that time expended on behalf of unsuccessful plaintiffs may still be compensable if it also was expended on behalf of successful plaintiffs.

The Court of Appeals for the Fifth Circuit approved the same approach in Norris v. Hartmarx Specialty Stores, Inc., 913 F.2d 253 (5th Cir.1990), where only one of three plaintiffs prevailed on her Title VII claims at trial. The court observed that although counsel were not entitled to fees for legal work performed solely to develop the discrete claims of the unsuccessful plaintiffs, they were entitled to fees for time spent both on the claim of the successful plaintiff and on "other ... claims if that time contributed to her ultimate success in the case." Id. at 257.^[64]

Thus, precedent supports a decision by this court to give class counsel full value for Category Three hours. However, the court has a reservation about doing so. The court's concern is not the concern expressed by Miller, Faucher, which is the possibility of subsequent victories for the non-settling class. As class counsel contend, "[t]he value to the Incentive Retirees of counsel's work in securing the settlement is not diminished by the fact that others may ultimately benefit from the same work in the future." Moreover, an equitable reimbursement of money from the subsequent fund to the present settlement amounts could be made at that time.

Nor is the court greatly concerned, as are Miller, Faucher, Dale Nathan and Unisys, that in the event that the present settlement turns out to be the only victory, the settlement class is paying for unsuccessful time. This same work fully, albeit not solely, benefitted the settlement class.

Rather, the court is not positive that the attorneys deserve full value for time not wholly successful. After all, one of the attributes of contingent fee agreements is that the attorneys will not receive payment unless they are victorious. Had the claims been brought only on behalf of the non-settling class, the attorneys might well receive no payment in this case. It is far from clear that these attorneys should receive full payment just because work on an unsuccessful claim happened also to be work towards a successful one.

On the other hand, attorneys who help bring these important class actions play a crucial role in vindicating the rights of plaintiff classes; they should not be deterred from representing a class simply because one or more subgroups within the class might ultimately lose. Furthermore, appeals regarding the remaining claims in this case are still pending, and might ultimately be victorious.

In light of these considerations, as well as the existing case law, the court resolves this issue as follows. In calculating the lodestar range, it will give full value to Category Three hours. However, the court will make a very modest reduction when determining an overall percentage award to account for the possibility that no other group of plaintiffs is ultimately successful. Should such a group later succeed, the court may then consider this modest reduction when determining a subsequent fee award to class counsel.

One final area of concern for the court is whether there was any unnecessary duplication of efforts by class counsel.^[65]See In re Fine, 751 F.2d at 594-95 (endorsing reductions for duplicative work). Class counsel took steps to avoid such duplication. Responsibilities initially were assigned according to each firm's existing contacts with or proximity to potential witnesses and retiree organizations. These functional assignments continued throughout the case, including preparation for and conduct of the trial. To avoid duplication of efforts, other tasks that could not be assigned in this way were usually assigned to only one firm, with review and oversight of that work also limited to a few firms.^[66]

Legalgard scrutinized every hour of each petition, and detailed numerous instances where two or more firms billed time for the same general task, e.g. summary judgment, document review, etc. However, most of these findings are inconclusive, since they don't inform the court whether there was actual duplication of the precise work that was performed. For example, Legalgard does not specify whether the same document was reviewed twice, or whether two firms performed the same pre-trial or trial activity. Moreover, not every duplication of a specific activity is unnecessary per se. Nevertheless, the Legalgard analysis did help the court to locate areas of possible duplication. Moreover, its findings regarding duplication of depositions was witness-specific; its identification of conference calls and meetings was also on point.

The court asked class counsel to elucidate its position that no avoidable duplication occurred within any of the categories of work. Guided by this response and the Legalgard findings, the court performed a preliminary analysis.

In a case requiring over 20,000 hours of work distributed among thirteen firms, it is almost inevitable that some unnecessary duplication will occur. Yet, the amount of unnecessary duplication in this case was extremely small. Class counsel did a praiseworthy job of streamlining their efforts in a way that enabled the case to go to trial in less than one year and minimized duplicative work. Most of the duplication that did occur — preliminary review and drafting of the complaint, conferences, a few depositions with more than one attorney present, negotiations with Unisys and discussions with plaintiffs regarding settlement, class certification activities, and some document review and legal research — was necessary either because it occurred before consolidation or because it enabled the firms to knowledgeably carry out their different responsibilities and to act as a coordinated entity.^[67] Based on the analysis, the court finds that only 50-100 hours should be deducted as unnecessarily duplicative.

Based upon the various foregoing deductions, the court determines that a total of 3,155-4,445 hours are not compensable from the fund under a lodestar analysis.^[68] Accordingly, the range of compensable hours is 16,025-17,315.

2. Hourly Rate

The court next determines the rates to be applied to the compensable hours. Because the court is unable definitively to say precisely which hours would be eliminated under a full-scale lodestar analysis, the court will apply a uniform rate to the remaining hours. This uniform rate, however, will reflect the individual rates of all attorneys and paralegals involved in this case.^[69]

To determine the uniform rate, the court begins with the actual charges listed for this case by the individual attorneys and paralegals.^[70] Across all of the firms, billing rates were, with rare exceptions,^[71] in the following ranges:

Paralegals[72]: $40-$125
Associates[73]: $100-$300
Partners[74]  : $175-$425

The average rate — the quotient of the aggregate time charges divided by the aggregate hours — is $194.^[75]

The individual rates which make up this average must be reasonable. "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." Cunningham v. City of McKeesport, 753 F.2d 262, 267 (3d Cir.1985). Accordingly, the court must ensure that the rates charged are not out of line with each attorney's (and paralegal's) normal rates or with the rates charged in each attorney's community.

The court first focuses on the attorneys' customary billing rates.^[76] Many of the attorneys in this case charged their normal hourly rate. However, attorneys from several of the petitioning firms charged rates either higher than those charged in other cases, or did not employ a discounted rate that they have used in other cases. But this is not an uncommon practice in cases of complex litigation, which requires considerable time and expertise, or in cases taken on behalf of single-litigation clients, who cannot guarantee a certain volume of work.

The court must further satisfy itself that the rates charged are reasonable compared to the rates charged in the relevant community for legal services. Some of the rates in this case are high-end in the relevant communities. Contingent-fee rates often are high. Furthermore, attorneys with more years of experience or at more prestigious firms typically command a higher rate. Similarly, the complex and difficult nature of this class action ERISA case demands a quality of service for which relatively expensive representation is to be expected. See Steiner, 835 F.Supp. at 787 (relevant community is "the national legal community that practices securities litigation"). In short, the requested rates are standard in complex class actions taken on a contingency basis, with fluctuation due to the reputation and location of each firm, and the experience and expertise of each attorney.^[77] The court thus find that the rates, albeit expensive, are reasonable.^[78]

The average rate must, however, be reduced slightly. A large percentage of the hours to be deducted — pertaining to the settlement, summary judgment, and the trial — were hours requiring expertise. Accordingly, they were spent by attorneys, as opposed to paralegals. Moreover, a substantial number of these hours were spent by those attorneys charging the higher rates in this case. Therefore, the average rate for the remaining hours will be less than the average rate for all hours. The court calculates that using an average rate of $190 will adequately account for this disparity.

To determine the aggregate unadjusted lodestar for class counsel, the court multiplies the total number of compensable hours, 16,025-17,315, by the average rate, $190. Accordingly, the aggregate lodestar before adjustments is $3,044,750-$3,289,850.

3. Multiplier

The court must next determine what multiplier would be applied to the aggregate lodestar under a full-scale analysis. "[T]he party proposing any adjustment to the lodestar bears the burden of justification." Blum, 465 U.S. at 898, 104 S.Ct. at 1548-49. Class counsel contend that, under a lodestar approach, 2.13 is an appropriate multiplier. They argue that this number is reasonable when compared to multipliers granted in other cases and, following the Lindy format, in light of counsel's efforts in this litigation. The court will discuss how the various Lindy factors play out in this case, and then assess what multiplier should be applied.

An enhancement of the lodestar might be called for in light of "the contingent nature of success." Lindy II, 540 F.2d at 117. Under this rubric, the court must examine the plaintiffs' burden at the time of filing suit, risks assumed in developing the case, and delay in receipt of payment. Id.

The likelihood that the settlement class would be unsuccessful was substantial. There is no doubt that this case presented complex legal and factual issues. Moreover, this an area of law where plaintiffs do not often succeed. Before Bixler v. Central Pa. Teamsters Health & Welfare Fund, 12 F.3d 1292 (3d Cir.1993), revived the prospects of success for plaintiffs' breach of fiduciary duty claims, the settlement class faced a strong possibility that they would not prevail. Plaintiffs also faced the likelihood that even if they were ultimately successful in this court, Unisys would keep the litigation alive through years of post-trial motions and appeals.

Class counsel also assumed substantial risks in representing the settlement class. Confronted with shaky odds, they expended more than 20,000 hours of labor on a contingency basis, without guarantee of remuneration.^[79] They also advanced more than half a million dollars in expenses.

Class counsel have not asked for an adjustment based on delay of payment. Instead, they have used their current, rather than historical rates, in their fee petitions. See Institutionalized Juveniles v. Secretary of Public Welfare, 758 F.2d 897, 923 & n. 41 (3d Cir.1985) ("using current market rates to calculate the lodestar figure may counterbalance the delay in payment as well as simplify the task for the district court" (quoting Murray v. Weinberger, 741 F.2d 1423, 1433 (D.C.Cir.1984))).

An enhancement of the lodestar might also be appropriate because of "the quality of [the attorneys'] work." Lindy II, 540 F.2d at 117-18. Under this rubric, the court may consider the results obtained, both in terms of a comparison with the extent of possible recovery and in consideration of how much time was invested. Id. at 118; Prandini v. National Tea Co., 557 F.2d 1015, 1019-20 (3d Cir.1977) (quality includes efficiency). The court may also evaluate "the professional methods utilized in processing the case." Lindy II, 540 F.2d at 118.

This settlement represents a good, albeit not the best, result for the settlement class. The Sperry incentive retirees have recovered 66-71% of what it would have cost Unisys to maintain the former Sperry plans for the incentive retirees' lifetimes. The Burroughs settlement amount comprises between 58-60% of what it would have cost Unisys to maintain the former Burroughs plans.^[80] In light of the risks of no recovery at all, the benefits provided by settlement are substantial. See Meshel, 102 F.R.D. at 139 (where the total recovery possible for 1,608 claimants is roughly $9.4 million, a settlement of $4 million is excellent under the circumstances).

Moreover, class counsel should be commended for obtaining the settlement — after a full trial — in less than two years. While counsel did not skimp on hours^[81], the speedy resolution preserved benefits for the settlement class, who faced a substantial cost increase for medical coverage in January, 1995 under the existing plans.

Finally, the court was impressed with the professionalism and skill with which class counsel have proceeded. Class counsel brought with them considerable experience in class action and ERISA litigation. They efficiently brought the case for the settlement class to resolution against a defendant energetically and ably represented by Morgan, Lewis & Bockius, a prominent firm with much experience in labor and employment litigation.^[82] They have conducted themselves with the utmost professionalism, combining zealous advocacy with respect for the court. Their legal work, from briefs to trial presentation, was of a very high caliber. Overall, the quality of their work was excellent.

In determining a multiplier for class counsel, the court must consider whether contingency enhancements are permissible in common fund cases. Though traditionally contingency enhancements have been used in common fund cases, a recent Supreme Court decision has cast some doubt on this state of affairs. In City of Burlington v. Dague, ___ U.S. ___, 112 S. Ct. 2638, 120 L. Ed. 2d 449 (1992), the Court disavowed contingency enhancements under fee-shifting statutes, but did not address whether they survive in common fund cases.^[83]

Although several district courts have held to the contrary, see, e.g., Nensel v. Peoples Heritage Financial Group, 815 F. Supp. 26, 29-30 (D.Me.1993); Bolar Pharmaceutical v. Gackenbach, 800 F. Supp. 1091, 1096 n. 7 (E.D.N.Y.1992), the two courts of appeal which have ruled on the issue held that Dague does not apply in common fund cases and that district courts retain discretion to award risk enhancements in such cases.^[84] In Florin v. Nationsbank of Georgia, 34 F.3d 560 (7th Cir.1994), where settlement produced a common fund worth more than $15 million, the Court of Appeals for the Seventh Circuit vacated the district court's refusal to enhance counsel's lodestar and held that the fee should have been considered under common fund principles. The court stated that "a risk multiplier is not merely available in a common fund case but mandated, if the court finds that counsel had no sure source of compensation for their services." Id. at 565 (citation omitted).

The Florin court explained that "the policy considerations in Dague that militate for forbidding risk multipliers in statutory fee-shifting cases have little force in common fund cases." Id. at 564. The Dague Court was concerned that awarding a multiplier to plaintiffs' counsel would unjustly cause the defendant to, in effect, "pay for the attorney's time ... in cases where [her] client does not prevail." Dague, ___ U.S. at ___, 112 S.Ct. at 2643. This concern does not exist in common fund cases, where defendants' liability is fixed and plaintiffs themselves give up part of their winnings to counsel. Florin, 34 F.3d at 565. "Further, there is no injustice in requiring plaintiff class members to shoulder the burden of compensating counsel for prosecuting the class' case without any assurance of compensation." Id.

In In re Washington Public Power Supply System Lit., 19 F.3d 1291 (9th Cir.1994), the Court of Appeals for the Ninth Circuit reached the same conclusion. "There is nothing unfair about contingency enhancements in common fund cases because of the equitable notion that those who benefit from the creation of the fund should share the wealth with the lawyers whose skill and effort helped create it." Id. at 1300 (citations omitted). Because the division of the fund was "of no concern whatsoever to the defendants," the court found "Dague's reasoning inapposite in the common fund context." Id. at 1301.^[85]

However, the Dague Court had an additional objection to risk enhancements in statutory fee cases:

[T]he interest in ready administrability that has underlain our adoption of the lodestar approach ... and the related interest in avoiding burdensome satellite litigation ... counsel strongly against adoption of contingency enhancement. Contingency enhancement would make the setting of fees more complex and arbitrary, hence more unpredictable, and hence more litigable.

___ U.S. at ___, 112 S.Ct. at 2643 (citations omitted). This objection would seem to apply also to common fund cases. Federal Judicial Center, supra, at 70-71 ("In light of its clear desire to facilitate administration and avoid arbitrariness, it seems likely that the Court would reject enhancements in common fund cases.")

Although the Florin court made no mention of this objection by the Dague Court, the Washington Power court recognized it. That court found that both concerns "drove the [Dague] Court" to disallow contingency enhancements. 19 F.3d at 1300. Nevertheless, finding that these concerns "apply with much less force in common fund cases," the court held that contingency enhancements were appropriate for common funds. Id.

This court reads Washington Power to suggest that the Dague Court's second objection does not, alone, preclude contingency enhancements in statutory fee cases. The Florin court's silence as to the second objection is also significant. Therefore, in the absence of appellate rulings to the contrary, the court finds that contingency enhancements are appropriate in common fund cases.^[86]

The court finds that an multiplier in the range of 1.75-2.25 would be applied to the lodestar under a full-scale analysis.^[87] Multipliers in many cases have been in this range or lower — even 1.0 and less. See, e.g., In re Centocor, 1993 WL 189937, at *6 (no enhancement of lodestar warranted). Still, multipliers in many other cases have been higher. See, e.g., O'Rourke v. Healthdyne, Inc., No. CIV.A. 84-4295, 1986 WL 923, at *2 (E.D.Pa. Jan. 16, 1986) (multiplier of 3.0); Municipal Authority of Bloomsburg v. Pennsylvania, 527 F. Supp. 982, 1000 (M.D.Pa.1981) (multiplier of 4.5); Rabin v. Concord Assets Group, Inc., [1991-1992 Transfer Binder] Fed.Sec.L.Rep. (CCH) ¶ 96,471 at 92,081, 1991 WL 275757 (S.D.N.Y. 1991) ("[i]n recent years multipliers of between 3 and 4.5 have been common"); Behrens v. Wometco Enterprises, Inc., 118 F.R.D. 534, 549 (S.D.Fla.1988) ("the range of lodestar multipliers in large and complicated class actions runs from a low of 2.26 to a high of 4.5"). This is particularly relevant because class counsel in this case fared well under the contingency and quality tests. Among the factors leading the court to choose a multiplier in the 1.75-2.25 range, rather than a larger multiplier, are the following^[88]:

(1) Only in the most exceptional circumstances would this court award a multiplier of 3 or greater. Though many courts do not employ such an approach, this court believes that lodestars enhanced by multipliers less than 3 should adequately compensate even the most talented counsel. See Skelton, 860 F.2d at 258 (a multiplier of 2 is a sensible ceiling for lodestars).

(2) The rates charged in this case were high-end. The quality of an attorney's work is generally reflected in the reasonable hourly rate. Lindy II, 540 F.2d at 117; Silberman v. Bogle, 683 F.2d 62, 64 (3d Cir.1982); Zeffiro, 574 F.Supp. at 450. See also Blum, 465 U.S. at 899, 104 S.Ct. at 1549 (same principle in statutory fee case). The rates in this case were high partly because the petitioning attorneys are so experienced and of such high quality. See Centocor, 1993 WL 189937, at *6 ("in view of the high level of skill possessed and employed by counsel, which is directly reflected ... in their hourly rate, much of what counsel seeks as a multiplier is already being paid for"). Similarly, part of the risk involved in contingency-fee cases is reflected by a higher billing rate. See, e.g., Florin v. Nationsbank of Georgia, 869 F. Supp. 669, 670-72 (W.D.Wisc.1994) (on remand, court awards risk multiplier of 1.01, noting that in light of high hourly rates "[c]ounsel was handsomely rewarded for dealing with the complexities of this case"). Thus, while a multiplier is still appropriate for contingency and quality — indeed, the court must take these into account, Lindy I, 487 F.2d at 168 — the multiplier is lower in this case than it would have been if counsel had charged lower rates.

(3) This multiplier is an average of the multipliers appropriate for the time spent by each attorney and paralegal. If the court were making individual awards, many of the attorneys would receive multipliers higher than the average multiplier; others would receive lower. These differences would reflect, inter alia, the relatively greater risk for attorneys working on the Burroughs case, as well as differences among counsel in the difficulty of work undertaken, the quality of the work performed, and the billing rates. Class counsel similarly incorporated varied multipliers in their tentative allocation.

(4) Many hours were spent by paralegals. This work was necessary to the case and is thus compensable. However, although the paralegals technically risked not being compensated, their work does not justify an enhancement. See, e.g., Edmonds, 658 F.Supp at 1150 (declining to enhance the lodestar amount for paralegal time).

Multiplying the unadjusted aggregate lodestar by the lodestar multiplier produces a lodestar range of $5,328,312.50-$7,402,162.50. An award to Dale Nathan must be added to this range.

4. Dale Nathan

Dale Nathan's role in this litigation for the settlement class has been enigmatic. On the one hand, he only very recently became counsel of record. Moreover, he is the only attorney seeking compensation who did not have an individual retainer agreement with members of the class allowing him to file a complaint and fully represent them throughout the litigation. On the other hand, from the outset of this litigation, he has represented the MRDO; approximately 388 members of the MRDO, 41% of the total membership, are part of the settlement class. In this capacity, he assisted the MRDO in searching for a firm to represent it,^[89] and has expended several thousand hours on the organization's behalf.

Mr. Nathan's substantive contributions to the settlement are similarly difficult to characterize. On the one hand, he gathered information and documents from settlement class members and provided the same to class counsel. He helped to inform these members about the proposed settlement and represented their views in pleadings and at the settlement hearing. Generally, he was of much benefit to these plaintiffs by keeping them abreast of the case and serving as a liaison between them and class counsel.

On the other hand, Mr. Nathan argued strenuously against the proposed settlement. In fact, he has appealed the settlement on behalf of some members of the MRDO not included in the settlement class. His arguments opposing the settlement did help to clarify some of the issues pertinent to the court's approval. Moreover, in so arguing, he espoused the views of many members of the settlement class. However, there is no gainsaying the fact that this particular portion of Mr. Nathan's work, though stemming from his vigorous representation of the entire MRDO, jeopardized the successful outcome for the settlement class.

In light of these considerations, the court will award Mr. Nathan an unenhanced lodestar for time expended on behalf of the settlement class. See Howes, 668 F.Supp. at 1027 (fee award to counsel who played role of "devil's advocate" in objecting to settlement is justified). This award will fairly compensate Mr. Nathan for his efforts without rewarding him.

The court must adjust Mr. Nathan's request. He calculated his total expenses to be $15,913.21 and calculated his total time charges to be $82,710.04. He then deducted from this latter amount one-third of the $98,623.25 sum of his expenses and time charges, or $32,874.42. He thus requests $15,913.21 in expenses and $49,835.62 in fees. However, Mr. Nathan should have applied his one-third reduction equally to his fees and to his costs. Thus, a proper request for him is $10,608.86 in expenses and $55,140.30 in fees.

The court must next determine whether a 33% reduction is sufficient.^[90] Based upon the court's review and Mr. Nathan's own breakdown, approximately 20% of his time is noncompensable either because it was expended exclusively on behalf of the non-settling class or because it pertained to the attorneys' fees petitions.

Approximately 30% of his time was spent on activities concerning members of both the settlement and non-settling classes. Among these activities are searching for a law firm to endorse, gathering information and documents, and communicating with plaintiffs. The court concludes that the number of hours expended on these activities would have been smaller if the case had been brought only by the settlement class. For example, less time would have been expended on the law firm search: it is the court's understanding that Mansfield & Tanick and Meagher & Geer would likely have agreed early on to represent a class that did not contain retirees from the Unisys class. Moreover, Mr. Nathan would have spent fewer hours assisting class counsel with summary judgment and otherwise gathering information regarding regular retirees. As explained supra, regular retiree evidence would still have been necessary, but less of it would have been accumulated and used.^[91] The court finds that, if the case had been brought only by the settlement class, Mr. Nathan would have spent 25%-33% less time on these activities.

This amounts to an additional 7.5%-10% overall reduction. Adding this to the initial 20% reduction amounts to a reduction of 27.5%-30%. The court further finds that reductions for vagueness, excessiveness, and irregularity amount to no more than 3% of the total time expended. Therefore, Mr. Nathans's proposed 33% reduction is reasonable. Accordingly, the court will award Mr. Nathan $55,140.30 in fees and $10,608 in costs.^[92]

The sum of Dale Nathan's award and the lodestar range for class counsel is $5,383,452.80-$7,457,302.80. This is the aggregate lodestar range, i.e. the range within which the court would award fees if a full-scale lodestar analysis were used as the sole determinant. As a percentage of the settlement fund, this range is 4.8%-6.7%.

5. The Percentage Award

The court must now determine a percentage-of-the-fund award. This task is far less time-consuming than the modified lodestar analysis. But it is far from simple; it entails careful consideration of the interplay of numerous factors. Among the most salient of these factors are the following:

(1) Under no circumstances would the court award more than the 7.5% requested by class counsel.

(2) The lodestar range in terms of a percentage recovery is 4.8%-6.7%. The percentage represented by the mean of the lodestar range is 5.75%.

(3) The court must preserve meaningful benefits for the settlement class. Fee awards are designed partly "to provide protection to class members whose distribution and losses will be directly affected by the amount of such awards." First Fidelity, 750 F.Supp. at 164. Awards from a common fund must thus be "made with moderation and a jealous regard to the rights of those who are interested in the fund." Greenough, 105 U.S. at 536. The plaintiffs in the settlement class have recovered approximately 65% of what they lost and, to the extent that a lodestar is guiding the award, are paying expensive rates of compensation to counsel. Expenses amounting to roughly .04% will also be deducted from the fund. See infra.

(4) The court must also adequately compensate counsel to ensure that ERISA continues to be vigorously enforced. As one federal court reasoned in the context of a securities case:

[I]ndividuals damaged by violations of the federal securities laws should have reasonable access to counsel with the ability and experience necessary to analyze and litigate complex cases. Enforcement of the federal securities laws should be encouraged in order to carry out the statutory purpose of protecting investors and assuring compliance.... A large segment of the public might be denied a remedy for violations of the securities laws if contingent fees awarded by the courts did not fairly compensate counsel for the services provided and the risks undertaken.

In re Union Carbide, 724 F.Supp at 169 (citation omitted). See also Deposit Guaranty Nat'l Bank v. Roper, 445 U.S. 326, 338, 100 S. Ct. 1166, 1174, 63 L. Ed. 2d 427 (stressing importance of class actions generally), rehg. denied, 446 U.S. 947, 100 S. Ct. 2177, 64 L. Ed. 2d 804 (1980); Warner Communications, 618 F.Supp. at 750-51 ("Fair awards in cases such as this encourage and support other prosecutions, and thereby forward the cause of securities law enforcement and compliance"); Bateman Eichler, Hill Richards, Inc. v. Berner, 472 U.S. 299, 310, 105 S. Ct. 2622, 2628-29, 86 L. Ed. 2d 215 (1985) (stressing importance of private actions as a supplement to SEC action). The same reasoning applies with equal force to ERISA cases. The filing of lawsuits such as this one should not be deterred by inadequate fee awards.^[93]

(5) In calculating the lodestar, full value was given to Category Three hours, even though class counsel expended these hours partly on behalf of the regular retirees, who have not yet made a recovery.

(6) A fee percentage negotiated beforehand would likely have been 7.5%-12.5% for a $111 million recovery.

(7) Fee awards for recoveries of $100-$200 million typically constitute 4%-10% of the fund.

The court awards counsel 6.25% of the settlement fund. This amounts to $6,937,500.00 — $55,140.30 to Dale Nathan and $6,882,359.70 to class counsel. This amount is fair and reasonable in light of all the circumstances surrounding this case. The court believes that this amount preserves a significant recovery for the Burroughs and Sperry incentive retirees while adequately rewarding counsel for a job well done.

D. REIMBURSEMENT OF EXPENSES

The reimbursement of expenses to counsel who create a common fund is appropriate. SmithKline Beckman, 751 F.Supp. at 534. These expenses must be reasonable and necessary. See Danny Kresky, 716 F.2d at 219-20; Domestic Air, 148 F.R.D. at 357.

Class counsel seek reimbursement for the following aggregate expenditures:

    Berger & Montague: $171,299.86
    Davis, Miner: $66,913.75
    Joseph Roda: $5,961.24
    Mansfield & Tanick: $28,313.98
    Meagher & Geer: $51,342.07
    Hirsch & Associates: $301.00
    Heller, Kapustin: $1,243.87
    Milberg, Weiss: $6,851.27
    Rossbacher & Associates: $14,079.35
    Schiffrin & Craig: $962.50
    Levin, Fishbein: $411.86
    Gottlieb & Goren: $11,465.99
    Sommers, Schwartz: $33,777.86
    Miller, Faucher: $21,719.22
    The Segal Company (actuarial consultants):
    $162,785.46

The court finds that these expenses are reasonable, although the amounts expended on travel often bordered on excessive, and some of the annotations were overly vague. However, despite their reasonableness, some expenses should not be reimbursed because they were not expended on behalf of the settlement class.

Seven of the twelve firms that made Category Two hour reductions in their fee petitions failed to also reduce their expenses. Of the five firms that did reduce their expenses, four firms tried to identify or estimate expenses not related to work on behalf of the settlement class and ultimately reduced their expenses by a smaller percentage than by which they had reduced their lodestar hours. In contrast, Joseph Roda employed a pro rata formula by which he reduced the total amount of expenses by the same percentage by which the lodestar hours were reduced.

Expenses corresponding to work that was not expended on behalf of the settlement class are similarly unnecessary to the settlement class and should not be reimbursed. But expenses typically are not memorialized with the same specificity as are billable hours. For example, entries of expenses will usually be as general as "photocopies" or "postal service." This makes it extremely difficult for the firms, or the court, to ascertain which expenses correspond to which tasks; attempts to eliminate expenses in this way will seldom be accurate.

So long as the reduced hours can generally be expected to have entailed corresponding expenses, reducing the expenses by the same percentage by which the lodestar hours were reduced serves as the best proxy for making an accurate reduction of unnecessary expenses. Accordingly, the court will multiply the overall expenses for each firm by an appropriate percentage corresponding to each firm's voluntary Category Two hour reductions to make an initial reduction of Category Two expenses^[94]:

  Berger & Montague: $179,238 × .93 =
  $166,691.34
  Davis, Miner: $71,928.28 × .87 = $62,577.60
  Joseph Roda: $6,754.18 × .8826 =
  $5,961.24
  Mansfield & Tanick: $30,883.89 × .88 =
  $27,177.82
  Meagher & Geer: $54,571.92 × .91 = $49,660.45
  Hirsch & Associates: $301.00 × .90 =
  $270.90
  Heller, Kapustin: $1,243.87 × 1.0 =
  $1,243.87
  Milberg, Weiss: $6,851.27 × 86 =
  $5,892.09
  Rossbacher & Associates: $14,079.35 × .88
  = $12,389.83
  Schiffrin & Craig: $962.50 × .89 =
  $856.63
  Levin, Fishbein: $411.86 × .82 = $337.72
  Gottlieb & Goren: $11,465.99 × .93 =
  $10,663.37
  Sommers, Schwartz: $33,777.86 × .99 =
  $33,440.08
  Miller, Faucher: $21,719.22 × .90 = $19,547.30
  The Segal Company: $162,785.46 × 1.0 =
  $162,785.46[95]

Further reductions should be made to correspond with the additional Category Two reductions of lodestar hours made by the court regarding summary judgment, discovery, trial, and settlement.^[96] Using the means for all ranges, the court deducted 3,437.5 hours from the lodestar corresponding to work on these activities. This amounts to roughly 17% of all hours remaining after the firm's own cuts. A 17% reduction of the remaining aggregate expenses would thus be a reasonable approximation of the summary judgment, trial, and settlement expenses which were not incurred on the settlement class's behalf. The court will apply this reduction to each firm^[97]:

  Berger & Montague: $166,691.34 × .83 =
  $138,353.81
  Davis, Miner: $62,577.60 × .83 = $51,939.41
  Joseph Roda: $5,961.24 × .83 = $4,947.83
  Mansfield & Tanick: $27,177.82 × .83 =
  $22,557.59
  Meagher & Geer: $49,660.45 × .83 = $41,218.17
  Hirsch & Associates: $270.90 × .83 =
  $224.85
  Heller, Kapustin: $1,243.87 × .83 =
  $1,032.41
  Milberg, Weiss: $5,892.09 × .83 =
  $4,890.43
  Rossbacher & Associates: $12,389.83 × .83
  = $10,283.56
  Schiffrin & Craig: $856.63 × .83 =
  $711.00
  Levin, Fishbein: $337.72 × .83 = $280.31
  Gottlieb & Goren: $10,663.37 × .83 =
  $8,850.60

  Sommers, Schwartz: $33,440.08 × .83 =
  $27,755.27
  Miller, Faucher: $19,547.30 × .83 = $16,224.26
  The Segal Company: $162,785.46 × .83 =
  $135,111.93

Accordingly, the court will reimburse class counsel for expenses in the amount of $464,381.42. Including the reimbursement to Dale Nathan, the aggregate total reimbursement is $474,990.29.^[98]

E. ALLOCATION BETWEEN SPERRY AND BURROUGHS ACCOUNTS

Class counsel requested fees and expenses from the Sperry settlement amount that are approximately 66% of their total (lodestaradjusted) fee requests, and approximately 73% of their total expense requests. Their requested (unadjusted) lodestar attributable to the Sperry settlement is approximately 78% of their total requested lodestar.^[99]

Class counsel may wish to alter this allocation in light of the fee and expense payment granted to Dale Nathan, and/or in light of the court's reductions of fees and expenses to class counsel and the court's reasons for these reductions. The court therefore will allow class counsel to determine the allocation of fees and expenses between the two settlement accounts. The percentage of this final allocation attributable to the Sperry settlement amount must not, however, exceed 80% or fall below 60% for either fees or expenses.

An appropriate order follows.

ORDER

AND NOW, this 20 day of March, 1995, it is hereby ORDERED that:

(1) Defendant Unisys Corporation shall pay $6,937,500.00 in attorneys' fees to Berger & Montague, P.C. Defendant shall attribute this amount as between the Sperry settlement account and the Burroughs settlement account as directed by class counsel and in conformity with the court's opinion.

(2) Berger & Montague, P.C., shall pay from these fees $55,140.30 to Nathans and Associates and shall disburse the remainder as agreed upon by class counsel. In deciding upon this disbursement, class counsel shall apply the same criteria to Miller, Faucher, Chertow, Cafferty, & Wexler that it applies to all other firms.

(3) Defendant shall pay $474,990.29 in expenses to Berger & Montague, P.C. Defendant shall attribute this amount as between the Sperry settlement account and the Burroughs settlement account as directed by class counsel and in conformity with the court's opinion.

(4) Berger & Montague, P.C., shall pay from this reimbursement $10,608.86 to Nathan and Associates and shall disburse the remainder in conformity with the court's opinion.

(5) Dale Nathan's Motion to Amend is GRANTED.

(6) All other outstanding motions are DENIED.

NOTES

[1] Those retirees who did not retire under such a program have been called "regular retirees" throughout this litigation. Also, unless otherwise noted, "incentive retirees" refers to the Sperry and Burroughs settlement class, not to the Unisys incentive retirees.

[2] The court found that the explicit statements in the Sperry SPDs that retirement benefits would continue for life were internally inconsistent with the reservation provisions in SPDs, and thus rendered the SPDs ambiguous.

[3] The Sperry regular retirees' motion for reconsideration allowed the court to restore the breach of fiduciary claims.

[4] At the settlement hearing, the issue of attorneys' fees was argued, and members of the settlement class had fair notice and an adequate opportunity to be heard. See In re Fine Paper Antitrust Litigation, 751 F.2d 562, 584 (3d Cir. 1984).

[5] Unisys will establish two accounts: a Sperry Incentive Retiree Medical Account and a Burroughs Incentive Retiree Medical Account. Although no funds will actually be set aside, the establishment of these accounts will enable the incentive retirees and Unisys to account for Unisys' financial obligations to the incentive retirees over time. The starting balances for the accounts will be the settlement amounts. One of the first deductions will be for those attorneys' fees and expenses for class counsel which are approved by the court. The remaining balances, with interest, will be used to pay for the incentive retirees' health benefits, as well as the reasonable costs of administering the plan. Medical claims payments and reasonable administrative expenses will be deducted from the accounts as they are paid. The contributions received from incentive retirees, and interest on the outstanding balance in each account at a fixed annual rate of 7 3/8 %, will be added as credits to the accounts.

[6] Under a lodestar approach and the tentative allocation of the requested fees, each firm would receive a multiplier larger or smaller than 2.13 corresponding to its role in the litigation.

[7] Counsel from Miller, Faucher suggest that their fees would total at least $375,000 under the approach recommended in the first petition.

[8] Mr. Nathan miscalculated his request in his original petition. The aforementioned figures are correct. The court grants his Motion to Amend.

[9] Most counsel in this case worked completely on a contingency basis. Therefore, they have received no compensation as of this date. Two exceptions are Mr. Nathan, who has received $35,226.58 from the MRDO, and the firms of Gottlieb & Goren, P.C., and Sommers, Schwartz, Silver & Schwartz, which were advanced $60,750 by the Burroughs Alumni Retirement Group. Mr. Nathan and counsel from the two abovementioned firms have represented that they will reimburse their clients for these expenditures should they receive an award from this court.

[10] Arguably, defendant has no standing to object to the fee petitions. See, e.g. Florin v. Nations-bank of Georgia, 34 F.3d 560, 562 n. 1 (7th Cir.1994) ("Defendants have satisfied their obligation to pay into the settlement fund, and thus have no interest in the amount of fees class counsel want to extract from the fund"); Boeing Co. v. Van Gemert, 444 U.S. 472, 481-82, 100 S. Ct. 745, 750-51, 62 L. Ed. 2d 676 (1980) (defendant has no interest in fund, and "any right defendant may establish to the return of money eventually unclaimed is contingent on the failure of ... class members to exercise their present rights of possession"). But see Jackson v. United States, 881 F.2d 707, 709 (9th Cir.1989) (government has "ancillary standing" by virtue of its status as party to case in chief). In any event, the court will consider defendant's arguments. Counsel from Miller, Faucher and settling members of the MRDO, all of whom do have standing, have objected. Also, the court is obligated independently to determine the reasonableness of the award of attorneys' fees. Thus, the absence of an objection by defendant would not prevent the court from examining the fee petitions. Arguments and exhibits furnished by defendant will therefore be accepted, in effect, as "amicus" submissions.

[11] The court received 241 written objections from settlement class members. Because retirees' spouses are included as class members, it is possible that one partner of a marriage objected on behalf of both spouses. There was at least one instance where both spouses objected individually; the court will disregard this fact, however, for the sake of making the following calculation. If the court assumes that each objection was also an objection from any eligible spouse, it must multiply 241 by 1.76, since 76% of the incentive retirees have living spouses. This results in a range of 241-424 objections out of roughly 7200 settlement class members.

[12] "In support of the American rule, it has been argued that since litigation is at best uncertain one should not be penalized for merely defending or prosecuting a lawsuit, and that the poor might be unjustly discouraged from instituting actions to vindicate their rights if the penalty for losing included the fees of their opponent's counsel." Fleischmann Distilling Corp. v. Maier Brewing Co., 386 U.S. 714, 718, 87 S. Ct. 1404, 1407, 18 L. Ed. 2d 475 (1967).

[13] The other exceptions pertain to statutory fee-shifting provisions, see Ruckelshaus v. Sierra Club, 463 U.S. 680, 684, 103 S. Ct. 3274, 3277, 77 L. Ed. 2d 938 (1983), such as the one in ERISA, 29 U.S.C. § 1132(g); willful violation of a court order, Toledo Scale Co. v. Computing Scale Co., 261 U.S. 399, 426-28, 43 S. Ct. 458, 465-66, 67 L. Ed. 719 (1923); and bad faith or oppressive litigation practices, Vaughan v. Atkinson, 369 U.S. 527, 530-31, 82 S. Ct. 997, 999-1000, 8 L. Ed. 2d 88 (1962).

[14] "[I]n most class actions, the named plaintiffs [have] contracted with their attorney[s] to pay [them] a contingent fee. Although the absent class members obtain the same benefit from the settlement in this case as do the named plaintiffs, they of course have assumed no such independent fee obligation to plaintiffs' counsel." Bowen v. Southtrust Bank of Alabama, 760 F. Supp. 889, 893 n. 10 (M.D.Ala.1991).

[15] "Today the rationale has been fortified by an economic analogy to contingent fee litigation — the market reflects the fact that plaintiffs will pay a percentage of their recovery in return for the work of attorneys in obtaining the result." Jerold Solovy et al., Attorneys' Fees in Common Fund Cases: The Percentage Method is Back in Securities Class Actions: Abuses and Remedies 145 (Edward J. Yodowitz et al., eds. 1994).

[16] See also In re Fine, 751 F.2d at 583 (allowing recovery of attorneys' fees from common fund after settlement of case brought under Clayton Act); Sutton v. Medical Service Ass'n of Pennsylvania, No. CIV.A. 92-4787, 1994 WL 246166, at *3 (E.D.Pa. June 8, 1994) (Judge Broderick implies in dicta that recovery of attorneys' fees from settlement fund would be allowed under ERISA).

[17] Heightened scrutiny is warranted because of an important difference between common fund and statutory fee cases. In the latter, the defendant has a stake in the outcome of attorneys' fees issues, and adversarial litigation or close scrutiny by the defendant results. In re Fine, 751 F.2d at 582-83. In common fund cases, on the other hand, challenges to fee petitions are less common and less rigorous because those parties with a stake in the matter, the plaintiffs, are usually unorganized or lack the incentive to make challenges, see Matter of Continental Illinois Securities Litigation, 962 F.2d 566, 573 (7th Cir.1992) (each individual class member's gain from a reduction in fees is small). For this reason, and because there exists at the fee petition stage "an inevitable conflict of interest between [class action] attorneys ... and the class members," Id. at 583, the court is often called a "fiduciary" for the plaintiffs. See, e.g., Skelton v. General Motors Corp., 860 F.2d 250, 253 (7th Cir.1988), cert. denied, 493 U.S. 810, 110 S. Ct. 53, 107 L. Ed. 2d 22 (1989); Report of the Third Circuit Task Force, Court Awarded Attorney Fees, 108 F.R.D. 237, 251 (1985) [hereinafter Task Force Report]. The case at bar is a common fund case converted from a statutory fee case by the creation by defendant of the settlement fund in exchange for the release of defendant from liability for damages and statutorily authorized fees. Thus, "[i]nstead of serving as an arbiter in an adversarial setting, as is the case in a statutory fee controversy, the [court] must determine a reasonable fee by weighing the appropriate interests of the beneficiaries in light of the efforts of counsel on their behalf." Brown v. Phillips Petroleum Co., 838 F.2d 451, 456 (10th Cir.1988), cert. denied, 488 U.S. 822, 109 S. Ct. 66, 102 L. Ed. 2d 43 (1988). The court is fortunate that a well-organized group of plaintiffs, as well as a concerned defendant and a petitioning law firm with an alternative viewpoint, have raised some issues and crystallized others. This has helped the court to scrutinize the petitions and thus better protect the settlement class.

[18] The court's role in this regard was anticipated by the Settlement Agreement itself: "[The] awards must be approved by the Court following an application by Class Counsel." Settlement at 31. Moreover, "reasonableness" is a proven benchmark for judging an award of attorneys' fees. See Hensley v. Eckerhart, 461 U.S. 424, 433, 103 S. Ct. 1933, 1939, 76 L. Ed. 2d 40 (1983).

[19] For example, in Lindy, the district court awarded a percentage of part of the settlement fund to those petitioning attorneys responsible for that part of the fund. In determining the percentage, the court considered the percentage of recovery awarded as attorneys' fees in other cases, the amount of recovery from which fees were being awarded, the amount received by the petitioning attorneys from their clients under private agreements, and the time spent by those attorneys in connection with the litigation. Lindy Brothers Builders, Inc. of Philadelphia v. American Radiator & Standard Sanitary Corp., 341 F. Supp. 1077, 1089-90 (E.D.Pa.1972).

[20] In Hensley, the Supreme Court approved the Lindy approach in statutory fee cases. Id.; see also Blum v. Stenson, 465 U.S. 886, 888, 104 S. Ct. 1541, 1543-44, 79 L. Ed. 2d 891 (1984).

[21] For example, the Court of Appeals for the Fifth Circuit adopted a 12-factor approach in Johnson v. Georgia Highway Express, Inc., 488 F.2d 714 (5th Cir.1974). The Court of Appeals for the Ninth Circuit subsequently adopted the Johnson approach. Kerr v. Screen Extras Guild, Inc., 526 F.2d 67 (9th Cir.1975), cert. denied, 425 U.S. 951, 96 S. Ct. 1726, 48 L. Ed. 2d 195 (1976). The 12 factors are:

(1) the time and labor required;

(2) the novelty and difficulty of the questions involved;

(3) the skill requisite to perform the legal service properly;

(4) the preclusion of other employment by the attorney due to acceptance of the case;

(5) the customary fee;

(6) whether the fee is fixed or contingent;

(7) time limitations imposed by the client or the circumstances;

(8) the amount involved and the results obtained;

(9) the experience, reputation, and ability of the attorneys;

(10) the "undesirability" of the case;

(11) the nature and length of the professional relationship with the client;

(12) awards in similar cases.

Johnson, 488 F.2d at 717-19. "Most commentators consider Johnson to be little different from Lindy because the first criterion of the Johnson test, and indeed the most heavily weighted, is the time and labor required. Similarly, many of the Johnson factors are subsumed within the initial calculation of hours reasonably expended at a customary hourly rate." Task Force Report, 108 F.R.D. at 244.

The Johnson approach has been criticized for being too imprecise and redundant. Copeland v. Marshall, 641 F.2d 880, 889-94 (D.C.Cir.1980) (court of appeals evaluates Johnson and Lindy and selects Lindy).

[22] This acknowledgment of the percentage approach is consistent with the Supreme Court's use of such an approach in every prior case in which it addressed the computation of a common fund fee award. See, e.g., Sprague v. Ticonic National Bank, 307 U.S. 161, 59 S. Ct. 777, 83 L. Ed. 1184 (1939); Central R.R. & Banking Co. v. Pettus, 113 U.S. 116, 5 S. Ct. 387, 28 L. Ed. 915 (1885).

[23] See, e.g., Dorfman v. First Boston Corp., 70 F.R.D. 366, 375 (E.D.Pa.1976); John C. Coffee, Jr., Rescuing the Private Attorney General: Why the Model of the Lawyer as Bounty Hunter is Not Working, 42 Md.L.Rev. 215 (1983) [hereinafter Rescuing the Private Attorney General].

[24] The lodestar approach in effect places the court in the "position of a public utility commission that regulate the `fair' return the attorney receives by both determining the attorney's normal billing rate and assessing whether the attorney's time was reasonably expended." John C. Coffee, Jr., Understanding the Plaintiff's Attorney: The Implications of Economic Theory For Private Enforcement of Law Through Class & Derivative Actions, 86 Colum.L.Rev. 669, 691 (1986) [hereinafter Understanding the Plaintiff's Attorney].

[25] The Court of Appeals for the Fifth Circuit appears to be staying with the lodestar approach. See Longden v. Sunderman, 979 F.2d 1095, 1099 (5th Cir.1992). Courts of Appeals in the remaining circuits have not addressed the issue. However, some district courts within these circuits have endorsed the percentage method. See, e.g., Edmonds v. United States, 658 F. Supp. 1126, 1144 (D.S.C.1987); In re Par Pharmaceutical, Inc. Securities Litigation, [1992 Transfer Binder] Fed.Sec.L.Rep. (CCH) ¶ 96,962, at 94,111, 1992 WL 150632 (S.D.N.Y.1992).

[26] However, as one district court judge recently pointed out, "[p]anels of our Court of Appeals have ... cited the Task Force Report without the expression of reservation." In re U.S. Bioscience Securities Litigation, 155 F.R.D. 116, 118 n. 4 (E.D.Pa.1994) (citing In re Busy Beaver Building Centers, 19 F.3d 833, 849 (3d Cir.1994); Rode v. Dellarciprete, 892 F.2d 1177, 1191 (3d Cir.1990); Student Public Interest Research Group of New Jersey v. AT & T Bell Laboratories, 842 F.2d 1436, 1450 (3d Cir.1988)).

[27] Counsel from Miller, Faucher, however, prefer the use of the lodestar method to better reflect the time expended on behalf of the settlement class.

[28] Some method for determining a percentage is necessary, since no one percentage automatically suggests itself. As one court surmised, "[t]here is no general rule of what percentage of a common fund may reasonably be awarded as a fee because the amount of any fee must be determined upon the facts of each case." Mashburn v. National Healthcare, Inc., 684 F. Supp. 679, 692 (M.D.Ala.1988).

[29] Cf. In re Oracle Securities Litigation, 131 F.R.D. 688 (N.D.Cal.1990) (court conducts a bidding process among plaintiffs' firms to determine class counsel and attorneys' fees).

[30] The court realizes that, especially in multidistrict litigation, obtaining such an agreement among so many attorneys and plaintiffs will often be difficult.

The court also notes that class counsel's reference to a 7.5% fee request in the notice of settlement sent to all members of the settlement class does not suffice as a negotiated percentage. Cf. Edmonds, 658 F.Supp. at 1142 (while not the pre-judgment negotiation envisioned by the Third Circuit Task Force, the ceiling of 5% on fees established at the time of notice to class members that the class action had been certified, where the class members could request exclusion from the class, "reflects knowledgeable assent, in advance, to the fee").

[31] See Continental Illinois, 962 F.2d at 572-73 (it is simpler to argue over what arrangement would have been negotiated than over details of lodestar).

[32] The court will, however, consider what percentage might have been negotiated as one equitable consideration in determining a reasonable percentage for class counsel.

[33] See, e.g., In re MGM Grand Hotel Fire Litigation, 660 F. Supp. 522 (D.Nev.1987) ($205 million settlement fund; 7% fee award); In re Baldwin-United Corp. Litigation, [1986-87 Transfer Binder] Fed.Sec.L.Rep. (CCH) ¶ 92,918, 1986 WL 12195 (S.D.N.Y.1986) ($183.8 million settlement fund; 4.1% fee award); In re "Agent Orange" Product Liability Litigation, 611 F. Supp. 1296 (E.D.N.Y.1985) ($180 million settlement fund; 5.5% fee award); Sioux Nation of Indians v. United States, 650 F.2d 244, 227 Ct. Cl. 404 (1981) ($106 million settlement fund; 10% fee award); In re Folding Carton Antitrust Litigation, 84 F.R.D. 245 (N.D.Ill.1979) ($200 million settlement fund; 6.6% fee award).

[34] "To avoid the problems and criticisms of a common fund fee award being summarily granted as a percentage of the fund recovered without any findings supporting its reasonableness ... courts are now supporting their percentage awards with particular findings showing factors considered." Attorney Fee Awards, supra, § 2.08 at 47.

[35] Again, however, in choosing an appropriate percentage the court will consider the percentages that other courts have found reasonable in cases with funds comparable to this $111 million fund.

[36] The court also observes that in virtually every case where the court notes a lodestar but awards fees based upon a percentage, the lodestar multiplier converted from this percentage is in the range of 1 to 4. This speaks to an implicit awareness by courts that the percentage should bear a reasonable relationship to the time expended by counsel. It is also not surprising. Since many court have adopted percentage ranges — ranges which were developed in large part by converting the results in lodestar cases into percentages — it is to be expected that percentage awards today will bear a resemblance to typical lodestar awards.

[37] "The public would be appalled at a fee which equals the total fund obtained even if justified by the time devoted, because of a lack of relationship to the result achieved." First Fidelity, 750 F.Supp. at 164.

[38] Class counsel have apparently used a similar maneuver. They have asked for a percentage award of 7.5. The court assumes that they then converted this requested award into a lodestar request by dividing the award by their aggregate lodestar. The resulting lodestar enhancement multiplier is 2.13, a figure that would look relatively arbitrary standing on its own. See Harman v. Lyphomed, Inc., 787 F. Supp. 772, 781 (N.D.Ill. 1992).

[39] The Warner court partially based its use of this factor on the Supreme Court's language in Blum regarding the percentage approach.

[40] The Brown court held that although the "`time and labor involved' factor need not be evaluated using the lodestar formulation" when other factors predominated, nevertheless "it is a relevant factor and the availability of contemporaneous time records enhances the trial court's ability to properly evaluate it." Id. at 456 & n. 3.

[41] The Diamond court noted that although the Court of Appeals for the Second Circuit had not yet addressed the Plan, "the instant award is the same as would be awarded if the lodestar method were used and then checked against the percentage of fund it represented, or if the classic Grinnell lodestar approach were used." Id. at 94,244 n. 1 (citations omitted).

[42] "A request for attorney[s'] fees should not result in a second major litigation," Hensley, 461 U.S. at 437, 103 S.Ct. at 1941, or otherwise "consume[] a disproportionate share of the court's time," In re Activision, 723 F.Supp. at 1374. In fact, a modified lodestar analysis might actually be closer in spirit to Lindy than are many of today's full-scale analyses. See Lindy II, 540 F.2d at 116 (the assessment of the fee should not "assume massive proportions, perhaps even dwarfing the case in chief"). For an illuminating example of the rigor entailed by a full application of the lodestar approach in a large case, see In re "Agent Orange" Product Liability Litigation, 611 F. Supp. 1296 (E.D.N.Y.1985). To assist the court in that case, three temporary law clerks were hired expressly to work on the fee petitions, and they, along with two senior law clerks, worked full-time for more than three months to help the judge examine the petitions and relevant legal issues. Id. at 1319.

The court notes, however, that the amount of work done by courts in lodestar cases appears to vary greatly. Thus, while the modified lodestar represents significant savings in judicial resources when compared to what would be entailed if this court were to undertake a full-scale lodestar analysis, its workload under this modified approach was apparently greater, and no less rigorous, than for some courts employing a "full-scale" lodestar analysis.

[43] To the extent that using even a modified lodestar analysis might encourage inefficiency, this development does not present a problem in this case. All counsel have performed their duties with integrity. The court does not believe that any work was performed with the intent of padding the time charges. Class counsel's preference for a percentage approach rather than the traditional lodestar analysis strengthens this belief.

[44] The Steiner court, noting the difficulty of "reaching truly reliable conclusions on factual issues" presented by the lodestar approach, such as proper billing rates and the nature, quality, and efficiency of the large quantity of work performed, reached this apt conclusion: "It may be that the discretion provided the Court in looking to a percentage of recovery method would lead to just as reasonable a finding on a fee and one that more clearly recognizes the amount of judgment and discretion that goes into making that decision." 835 F.Supp. at 792. To this, the court adds only that, to the extent that too much discretion is seen as a drawback of both the lodestar and percentage approaches, the use of a negotiated fee percentage in future cases will minimize such discretion. See First Fidelity, 750 F.Supp. at 163. The negotiated fee would thus also make the attorneys' fee process more predictable.

[45] See also Dunne, 915 F.2d at 545 (allowing use of either lodestar or percentage approach, court "require[s] only that fee awards in common fund cases be reasonable under the circumstances"); Rawlings, 9 F.3d at 513 (district court retains discretion to pick an appropriate method "in light of the unique ... circumstances of the actual cases").

[46] "The best short statement may be that the size of an award should be reasonably arbitrary — meaning that there is no magic in the precise final numbers but that they must be reasonable, which is to say, the result of considered, articulated and fair judgments." Superior Beverage, 133 F.R.D. at 128.

[47] The court will deal separately with the petition by Dale Nathan. After determining a reasonable award to Mr. Nathan, the court will add that award to the aggregate lodestar range when determining a total percentage award to all counsel.

[48] While checking the addition of dollar amounts in several of the fee petitions, the court found errors amounting to an overstatement of $2500. Although the addition of hours in these petitions was accurate, the court considered subtracting a number of hours from the aggregate, especially since the petitions scrutinized by the court were from firms where the number of entries to be added were fewer. The court cautions firms to carefully prepare future fee petitions.

In a few instances, firms listed an aggregate number of hours spent on this case and an aggregate lodestar, pointed out that they had subtracted time from the petition, and then gave a revised lodestar without listing the actual number of hours for which compensation is sought. In response, the court calculated the average hourly rate for the aggregate hours, and then calculated the revised hour total by subtracting from the aggregate hour total the subtracted dollar amount divided by the average hourly rate. While refreshingly evocative of school-day mathematics, this type of exercise speaks to the mechanistic, and often complex, nature of the lodestar calculation.

[49] The court believes, based on the attendance of in-house counsel in court and in other phases of the litigation, that in-house counsel spent a significant amount of time assisting Unisys' outside counsel, the firm of Morgan, Lewis & Bockius.

[50] Class counsel further contends that "[i]t does not matter to [the settlement class] how many lawyers and paralegals were needed to achieve the result, only the fact that it was achieved." Response of Class Counsel to Memoranda of Dale Nathan, Firm of Miller, Faucher and Unisys re Petition for Award of Attorneys' Fees at 14. However, a great victory for the settlement class would be hollow if too much of the resulting fund was given to class counsel. While class counsel's proposition might ring true where all concerned parties recognized that attorneys' fees would be awarded on the basis of a negotiated percentage, the use of a lodestar warrants concern by the settlement class and the court for an appropriate deployment of legal staff.

[51] While the court did not scrutinize every entry in every petition — to do so for the 20,000 hours in this case would have defeated the purpose of the modified lodestar — it checked a sufficient number of entries to make reasonable calculations. See Evans v. City of Evanston, 941 F.2d 473, 476-77 (7th Cir.1991) (upholding sampling procedure by district court), cert. denied, ___ U.S. ___, 112 S. Ct. 3028, 120 L. Ed. 2d 899 (1992).

[52] See, e.g., Bell v. United Princeton Properties, Inc., 884 F.2d 713, 721 (3rd Cir.1989) (district court legitimately reduced time spent by attorney because it was excessive in light of the attorney's expertise in ERISA law).

[53] Because the medical plans for Unisys retirees had been in effect for only a few years, the amount of variation in these plans pertinent to this case was considerably less for this class than for the Sperry and Burroughs classes. The work needed to be performed on behalf of this class was therefore considerably less as well.

[54] Time spent on fee petitions may not be compensated in common fund cases. Lindy II, 540 F.2d at 111. Cf. Danny Kresky Enterprises Corp. v. Magid, 716 F.2d 215, 219 (3d Cir.1983) (compensating time spent on fee petitions in statutory fee case); Basile, 640 F.Supp. at 704-05 (allowing time spent on fee petitions in common fund cases because of the importance of the petitions in protecting the class).

[55] Two firms — Heller, Kapustin, and Sommers, Schwartz — eliminated little or no time from their petitions for Category Two hours. However, most of the work that these firms performed pertained to the settlement class; virtually none of their work was spent on the activities covered by the guidelines.

[56] Some petitioning firms either did not include or eliminated a large percentage of such time.

[57] However, class counsel's response, though similar, would not have included the entire response actually made in this case, since there would be no need to address specifically Unisys regular or incentive retirees. Also, as evidenced by the post-trial briefs, such a response would have entailed much more than the summary judgment response, since it would have dealt with arguments unique to the incentive retirees.

[58] These facts were not presented at trial by affidavit, as they were at the summary judgment stage. However, only a small percentage of summary judgment time was spent translating the facts into these formats.

[59] The task of determining how many hours were spent on each activity is made difficult not only by the huge number of hours expended in this case, but also by the vagueness of many entries and the failure in many petitions to adequately signal to the court which activity was being furthered by a given entry. As mentioned previously, counsel should give clearer and fuller descriptions in their entries.

[60] Similar arguments could be made regarding the filing of the original actions, class certification activities, and the consolidation into multidistrict litigation. However, the court will not reduce for these activities which initiated this case.

[61] The court finds some evidence of these propositions in the actual deployment of regular retirees and incentive retirees at trial. On the Sperry side, where both regular retirees and incentive retirees proceeded to trial, 30 incentives and 26 regulars were used as witnesses. On the Burroughs side, where only the incentives proceeded to trial — the regulars had lost on summary judgment — 10 incentives and only 3 regulars were used as witnesses. This can be only partly explained by the fact that Burroughs had a higher percentage of incentive retirees.

[62] Dale Nathan argued for reductions of 50% for discovery, 50% for settlement negotiations, and 25% for trial activities. "[I]n setting the amount of any reduction, the court will inevitably be required to engage in a fair amount of `judgment calling' based upon its experience with the case and its general experience as to how much time a case requires." Bell v. United Princeton Properties, Inc., 884 F.2d at 721.

[63] Miller, Faucher's formula actually calculates the actual fee for Category Three hours, and then makes the reduction. The court, on the other hand, is reducing the number of hours and then calculating the fee. However, because the court is using an average rate for all hours and an across-the-board reduction of hours, the final fee under these two methods will not differ.

[64] The rule favored by class counsel was dealt with by the three preceding cases in statutory fee cases. However, the court does not see a reason for treating those cases differently than this common fund case where this rule is concerned.

That these courts compensate counsel for time that "aids" or "contributes" to successful claims of plaintiffs, does not change this court's analysis of time spent by counsel on discovery, summary judgment, trial, and settlement. Many of these hours aided or contributed to the claims of this class. However, as explained, either they wouldn't have been necessary had the settlement class proceeded alone or the benefits that they provided were rather incidental. Thus, the test of whether hours were "expended on behalf of" the settlement class provides a more useful mechanism for compensating attorneys in this particular case. Compare Lindy II, 540 F.2d at 112 (test is whether time "benefitted" the fund), with Hughes v. Repko, 578 F.2d 483, 487 (3d Cir.1978) (to be compensable, hours must have been "actually devoted" to the successful claims).

[65] This issue is addressed after the previous reductions have been made to avoid deducting hours that have already been eliminated.

[66] It was decided among class counsel in April, 1993 that Alan Sandals of Berger & Montague would serve as Lead Coordinating Counsel. In this capacity, Mr. Sandals coordinated all court filings and discovery, and generally coordinated the litigation process to ensure that required tasks were performed. With respect to pretrial preparation, trial, and settlement, Mr. Sandals shared strategic decision-making with Sarah Siskind of Davis, Miner, and Joseph Roda regarding the Sperry case, and with Charles Gottlieb of Gottlieb & Goren, and Joseph Golden of Sommers, Schwartz regarding the Burroughs case.

[67] For example, the Brubaker deposition was attended by Joseph Roda and Alan Sandals. At this deposition, Sandals handled issues not germane to claims of the Sperry retirees whom Roda represented. Thus, the presence of both attorneys was necessary. Cf. Lugo, 1994 WL 45100, at *3 (because it was unnecessary to have to have two attorneys present at deposition, court excluded 34 hours from the fee petition).

[68] The court applies these deductions "across-the-board" rather than specifying which particular hours are not compensable from the fund. The rule in this Circuit, developed under the lodestar regime, is that across-the-board reductions generally are impermissible; instead, the district court must identify with specificity which hours are being eliminated. In re Fine, 751 F.2d at 594-95. Cf. Tomazzoli v. Sheedy, 804 F.2d 93, 98 (7th Cir.1986) (it is within the district court's discretion to chop off a percentage of the hours requested "as a practical means of trimming fat from a fee application"). However, this rule should apply with much less force where, as here, a lodestar is being calculated only to guide an overall percentage award to be allocated by class counsel among themselves. Moreover, even under a lodestar regime, such a rule should not prevent the court from making across-the-board cuts where the hours are eliminated as a matter of equity rather than because specific hours were unnecessary, poorly documented, or otherwise non-compensable. Thus, the across-the-board reductions of hours for summary judgment legal research and preparation, and any general reductions of Category Three hours, would pass traditional lodestar muster. Finally, the across-the-board reductions of summary judgment factual work, discovery, settlement, and trial hours might also be acceptable under a traditional lodestar approach, since it is difficult to specify which hours would have been expended if the case had been brought only on behalf of the settlement class. See Duke v. Uniroyal, Inc., 928 F.2d at 1425 (estimate of time spent on certain non-compensable activities not clearly erroneous); In re Fine, 751 F.2d at 598 (district court did not abuse its discretion in eliminating 50% of damage study hours where it found that the total number of hours expended on the study was excessive); Allen v. City of Los Angeles, No. CV. 91-2497, slip op. at 19 (C.D.Cal. Jan. 13, 1995) (across-the-board percentage reduction of 30% for overstaffing "may lack complete precision, but no better solution presents itself"). See also Student Public Interest Research Group of New Jersey v. Monsanto Co., 721 F. Supp. 604, 612 n. 4 (D.N.J.1989) ("it is more appropriate to decrease the number of hours claimed by plaintiffs' counsel rather than attempt to combat the myriad permutations" regarding the several attorneys' billing rates, levels of experience, and participation in different activities; reduction of 20% of lodestar as estimate of excessive hours). Reductions of vague, irregular, and duplicative hours might require further specification under traditional lodestar criteria.

[69] "[I]ndividual determinations of reasonable billing rates are required for the lodestar determination." In re Fine, 751 F.2d at 590.

[70] Current billing rates were used in the petitions. This is appropriate in this case. See Skelton, 860 F.2d at 255 n. 5 (endorsing use of current rates to compensate for delay in payment); In re Union Carbide Consumer Products Business Securities Litigation, 724 F. Supp. 160, 163-64 (S.D.N.Y.1989) (use of current rates is appropriate, especially when the litigation was concluded within two to three years). See also Missouri v. Jenkins, 491 U.S. 274, 283-84, 109 S. Ct. 2463, 2469, 105 L. Ed. 2d 229 (1989) (endorsing use of current rates in fee petition filed pursuant to 42 U.S.C. § 1988 to compensate for delay); Kyreakakis v. Paternoster, 732 F. Supp. 1287, 1296 n. 12 (D.N.J.1990) (same).

[71] One associate used a billing rate of $80.00. Also, approximately fifteen hours of work performed by a senior partner of one firm were billed at a rate of $500. While the court ordinarily would be concerned about $500 hours being charged to plaintiffs such as the settlement class, such is not the case here. The deployment of a very small number of these hours by a senior partner at a top-flight law firm is not unusual or unreasonable in class actions. Moreover, the amount ($1125) that might be deducted from the fee request would be too small to have an impact on the overall percentage.

[72] This category includes law clerks.

[73] This category includes self-employed attorneys.

[74] This category includes "of counsel."

[75] The total time charges used for this calculation is $3,971,500, which is the total fee requested by class counsel minus the $2500 in overstatements spotted by the court.

[76] The court prefers, as does class counsel, to use a flat hourly rate for each attorney, rather than differentiating among tasks.

[77] It is for these reasons that the rates for some paralegals are actually higher than those for some of the attorneys in this case. Until the market no longer is the litmus test for fee petitions, this seemingly anomalous result will occur.

[78] The only submissions by class counsel regarding the reasonableness of their rates were their own affidavits. Thus, they did not provide the court with sufficient information to make a determination. See Blum, 465 U.S. at 896 n. 11, 104 S.Ct. at 1547 n. 11 (fee applicant must produce sufficient evidence in addition to affidavits to demonstrate that the requested rates are reasonable). The court thus needed to refer to other cases and sources of rate information to judge reasonableness. Class counsel should provide additional information in future cases where a lodestar is relevant in order to comply with Blum and lighten the heavy workload for courts that (even modified) lodestar analyses entail.

The court also finds that class counsel did a satisfactory job of using less expensive employees on tasks that did not call for considerable expertise. There is thus no need to reduce the billed rates for particular tasks. See Bell, 884 F.2d at 722 (using an attorney to perform ministerial services is a legitimate ground for decreasing a lodestar); In re Fine, 751 F.2d at 591-93 (district court did not abuse discretion in reducing time charges for certain work performed by partners).

[79] As previously mentioned, $60,750 was advanced by one group of retirees to two firms. However, because the size of the payment is small, and counsel have represented that this money will be refunded once fees are awarded, this fact will not affect an overall contingency enhancement for class counsel. See Basile, 640 F.Supp. at 702 & n. 5.

[80] The Burroughs SPDs contained no explicit lifetime promises. Thus, the different ranges in percentage of recovery reflect the higher risks of continued litigation for Burroughs incentive retirees.

[81] Cf. Steiner, 835 F.Supp. at 790 ("If counsel obtain a good result with a minimum of time expended, the Court may use a higher multiplier in recognition that counsels' efficiency should not work to their disadvantage and result in a lower fee.")

[82] See Trans World Airlines, Inc. v. Hughes, 312 F. Supp. 478, 480, 483-84 (S.D.N.Y.1970) (in determining fee, court considers the quality of the opposition as well as the standing of plaintiffs' counsel), aff'd as modified, 449 F.2d 51 (2d Cir. 1971), rev'd on other grounds, 409 U.S. 363, 93 S. Ct. 647, 34 L. Ed. 2d 577 (1973). All of the attributes of class counsel in this case are shared by defense counsel.

[83] As the court noted supra, the presence of a fee-shifting statute does not prevent recovery from a common fund when the case is converted by settlement. That proposition, although instructive, is not dispositive of the issue at hand. Courts could allow counsel to recover from the settlement fund, but apply statutory fee principles to the lodestar analysis.

[84] In dicta, the Court of Appeals for the First Circuit suggested that common fund contingency enhancements do not survive Dague. In re Nineteen Appeals Arising Out of the San Juan Dupont Plaza Hotel Fire Litigation, 982 F.2d 603, 619 (1st Cir.1992) (Lay, J. concurring); Nensel, 815 F.Supp. at 29-30 n. 5 ("Nineteen Appeals does not now prescribe any binding rule on the issue").

[85] See also Shalala, 1 F.3d at 1267-69 (rejecting the argument that Dague mandates an unenhanced lodestar rather than the percentage method in common fund cases). Arguments distinguishing common fund and statutory fee cases were also made in pre-Dague cases in which the availability of risk multipliers was in issue. See, e.g., Skelton, 860 F.2d at 254.

[86] In the context of statutory fee cases, the Supreme Court has also held that quality enhancements should seldom be employed. See Pennsylvania v. Delaware Valley Citizens' Council for Clean Air, 478 U.S. 546, 566, 106 S. Ct. 3088, 3099, 92 L. Ed. 2d 439 (1986) ("[T]he overall quality of performance ordinarily should not be used to adjust the lodestar, thus removing any danger of `double counting.'"); Blum v. Stenson, 465 U.S. 886, 898, 901, 104 S. Ct. 1541, 1548-49, 1550, 79 L. Ed. 2d 891 (1984). However, this does not preclude a quality enhancement in this common fund case. As with contingency enhancements, there is nothing unfair about requiring the plaintiffs to compensate counsel from the fund which they created. Moreover, the Supreme Court has intimated that quality enhancements are permissible in common fund cases. In Blum, the Court went on to deny that "number of persons benefitted" is a proper consideration in statutory fee cases. Id. at 901 n. 16, 104 S.Ct. at 1550 n. 16. In this regard, the Court distinguished common fund cases from statutory-fee cases in that percentage awards are calculated in the former, whereas reasonable fees in the latter "reflect[] the amount of attorney time reasonably expended on the litigation." Id. The Court's clear implication that the "number of persons benefitted" is a proper consideration in common fund cases, where the outcome is of greater significance than the time spent by counsel, applies broadly to the quality of work performed by counsel. Courts thus have refused to restrict quality enhancements in common fund cases. See, e.g., Sherin, [1987-88 Transfer Binder] Fed.Sec.L.Rep. (CCH) ¶ 93,582, at 97,608 n. 2, 1987 WL 18851.

[87] If this were a full-scale lodestar analysis, the court likely would determine a multiplier, and the number of compensable hours, for each attorney and paralegal. The court would also make individual awards based on these determinations. Of course, the firms then would be free to redistribute these awards among themselves.

Because the lodestar here is being used simply to guide a percentage calculation, such individual determinations are not strictly necessary. Nor are individual awards. Indeed, in a climate where an increasing number of courts are paying lip service to the percentage approach devoid of the lodestar, it is difficult to imagine how these courts could make individual awards without referring to lodestar-type considerations. See Superior Beverage, 133 F.R.D. at 124. Allowing the firms to divide an overall award among themselves provides the solution. See Attorney Fee Awards, supra, at 53-54 ("When multiple law firms participated jointly in recovering of a common fund for the benefit of the class, the court should reach an overall reasonable fee award for all counsel based on a fair percentage of a fund, and rule that allocation of the fund is a private matter among counsel"); Domestic Air, 148 F.R.D. at 357. In cases where firms could not reach an agreement, the court would have to step in and might then consider the lodestars.

Although this court has chosen to refer to the lodestar to determine a percentage, it nevertheless prefers to allow class counsel to divide up the total award among themselves. This approach significantly decreases the workload for the court. Based on the tentative allocation by class counsel in this case, it also has the effect that two firms which would receive an individual multiplier no greater than 3 from this court are in effect receiving a multiplier closer to 5 from class counsel. (These multipliers employed by class counsel will decrease as the total percentage awarded by the court falls below 7.5.) Other firms might also receive slightly more or less from this court than if individual calculations were employed.

This outcome does not trouble the court. Any windfall or shortfall to the individual firms is at the expense of the other firms with whom an agreement was made. It is not at the expense of the settlement class. This is because the court bases its percentage award in part on an average lodestar among all of the firms. Thus, although the court is slightly uncomfortable with the fact that a few firms might receive such a large enhancement of their compensable hours, much simplicity is gained without sacrificing fairness.

[88] Of course, part of the disparity between the ranges selected in this case and in others is the unpredictable, non-homogeneous nature of the multiplier selection process.

The court also notes that part of this range of multipliers exceeds the 2.13 multiplier requested by class counsel. If the lodestar approach alone were being employed, the court would not use a multiplier greater than 2.13. The court would fail its obligation to the plaintiff class by using a multiplier greater than was even requested. (The court would likewise fail the class by awarding more than the requested 7.5% of the settlement fund.) However, because the court is using the lodestar to guide a percentage-of-the-fund award of fees, it is proper to calculate what multiplier would be used by the court independent of what is requested.

[89] This search ultimately was unsuccessful because the MRDO insisted that one firm represent all members of the class. Local firms were willing to represent only some of the classes. Only Berger & Montague met this requirement for most of the formative stage of this case. However, the MRDO decided not to retain this firm.

[90] Mr. Nathan's and paralegals' billing rates are extremely reasonable. Furthermore, the hours to be reduced have approximately the same proportion of attorney hours to paralegal hours; the court may therefore use the percentage of hours as a proxy for the percentage of fees when making reductions.

[91] Indeed, the MRDO would have had significantly fewer members.

[92] Expenses reasonably necessary for the case may be reimbursed. See REIMBURSEMENT OF EXPENSES, infra. In general, the court will reduce these expenses by the same percentage by which hours — or in this case, fees — were reduced.

[93] Regarding the performance of counsel, one simple point that often gets lost amidst the fee petitions bears repeating: "[A]bsent counsel's efforts, there would be no fund." First Fidelity, 750 F.Supp. at 164.

[94] For the sake of simplicity, the court determines the appropriate percentage by which to reduce a firm's expenses by taking the percentage by which the firm eliminated Category Two hours, rounding to the nearest whole number, and subtracting from 1.0. Rounding is not necessary in the case of Joseph Roda, who submitted a more precise calculation.

[95] The Segal Company did no work that fits under the rubric developed by class counsel to eliminate Category Two hours. Moreover, their hourly rates were reasonable.

[96] The court will not make any reductions to account for expenses which might correlate with other reductions in lodestar hours, such as those for duplicative or irregular tasks. Nor will it reduce expenses out of a concern for Category Three hours.

[97] The court is aware that some firms spent zero or proportionally few hours on certain of these activities. Each firm, however, did at least spend a significant percentage of their time on discovery. Also, some firms spent relatively less time than other firms, on work pertaining to regular retirees. Rather than finely parsing the actual time spent by each firm on each activity and then comparing these times across the firms — an enterprise whose duration and complexity would exceed its usefulness — the court will make across-the-board reductions. Class counsel may subsequently wish to redistribute the reimbursements among themselves.

[98] The court notes that this amount is roughly .04% of the settlement fund. Any need to reduce the total fee and expense recovery for class counsel based on this additional amount has already been considered in determining the fee percentage.

[99] As class counsel pointed out in their petition, under a lodestar format their total fee request amounts to a multiplier of 1.78 being applied to the lodestar attributable to the Sperry settlement and a multiplier of 3.28 being applied to the lodestar attributable to the Burroughs settlement.