Summary
holding that the district court's continued supervisory role didn't undermine finality
Summary of this case from Fox v. Saginaw Cnty., Mich.Opinion
No. 78-2036.
Argued September 28, 1978.
Decided February 26, 1979. Rehearing and Rehearing En Banc Denied May 4, 1979.
William J. Harte, Chicago, Ill., of counsel, for plaintiffs-objectors Betty Oswald, Eileen Miller and Phil Miller.
Lawrence Walner, Chicago, Ill., for plaintiffs-appellants.
William J. Scott, Atty. Gen. of Illinois, Springfield, Ill., for plaintiff-appellee State of Illinois; Donald G. Mulack, Chicago, Ill., of counsel.
Charles E. Clark, Birmingham, Ala., for plaintiffs-appellees.
Thomas A. Gottschalk, Kirkland Ellis, Chicago, Ill., for defendant-appellee.
Appeal from the United States District Court for the Northern District of Illinois.
Before FAIRCHILD, Chief Judge, and BAUER and WOOD, Circuit Judges.
In 1976 the defendant, General Motors (GM), began substituting engines produced by its Chevrolet Division in many of the 1977 model year cars produced by its Oldsmobile Division. The discovery of the engine switch culminated in the commencement of a plethora of lawsuits against GM in the state and federal courts. The Judicial Panel on Multidistrict Litigation transferred those actions which had been filed in the federal courts to the United States District Court for the Northern District of Illinois for consolidated pretrial proceedings with several actions which were already pending there. See 28 U.S.C. § 1407. The district court certified that the actions could be maintained as a class action and later approved the settlement of the actions as to one of two subclasses of Oldsmobile purchasers.
This appeal is from the order of the district court approving the subclass settlement. Although the facts are lengthy, the litigation's history complex, and the resolution of the issues difficult, the issues may be stated with relative simplicity:
First, is the district court's order approving the subclass settlement appealable?
Second, should counsel prosecuting the appeal be limited to representing the interests of those class members who objected to the settlement before the district court?
Third, did the district court err by refusing to permit appellants' counsel to inquire into the conduct of the negotiations that led to the settlement?
Fourth, did the district court err by dismissing with prejudice the federal claims of those class members who declined to release their state law claims pursuant to the settlement agreement?
We find that this court does have jurisdiction to entertain the appeal and hold that the trial court erred in approving the subclass settlement. Consequently, we reverse and remand the order of the district court with instructions.
I. Facts
A. The Engine Interchange Litigation
Beginning in 1974, GM planners began considering the manufacturing requirements for GM cars for the 1977 model year. By 1976 various GM management committees began planning for extensive interdivisional engine exchanges. Because the Chevrolet Division had a significant surplus production capacity, GM planners decided to rely on Chevrolet produced engines to meet part of the engine requirements of GM's Buick, Oldsmobile and Pontiac Divisions.
To institute the engine interchange in the Oldsmobile Division, GM used codes to identify the different engines that would be used in its 1977 Oldsmobiles. The Rocket 350 V-8 engine produced by Oldsmobile, for example, was given the code name "L34"; the Chevrolet engine used in place of the Rocket was given the code "LM1." Moreover, GM, over some objections by the Chevrolet Division, decided to adopt a common engine color for all of its engines. Thus, the distinctive red Chevrolet engine became blue. Despite the planned Oldsmobile-Chevrolet engine change, GM's advertising, EPA gas mileage disclosures and communications to Oldsmobile dealers referred to the changes by the use of the codes.
Three Chevrolet produced V-8 engines were used in 1977 Oldsmobiles: the LM1, a 350 engine equipped with a four-barrel carburetor, the L65, a 350 engine equipped with a two-barrel carburetor, and the LG3, a 305 cubic inch displacement engine. The class eventually certified by the district court includes all purchasers of Oldsmobiles with Chevrolet engines regardless of which of the three Chevrolet engines the purchasers actually received.
The switch from standard components to different components in Oldsmobiles was not confined to engines. GM used different components than it had used in previous years for other parts of the power train (the engine, transmission, and drive axle) in some of its Oldsmobiles. For reasons which do not appear with clarity in the record, GM decided in 1976 to install in all 1977 Oldsmobile Delta 88 coupes and sedans the THM 200 transmission instead of the THM 350, the transmission traditionally used in those cars. The THM 200, like the THM 350, is produced by GM's Turbohydramatic Division. The THM 200, originally designed for use in the subcompact Chevette, was used in all 1977 Delta 88 coupes and sedans regardless of whether they contained Oldsmobile or Chevrolet engines. The appellants maintain that GM's advertising materials nevertheless indicated that the THM 350 was standard equipment in all 1977 Deltas.
The case before this court is a subset of the Oldsmobile litigation spawned by the discovery of the engine interchange. After filing suit in the Cook County Circuit Court alleging violations of the Illinois Consumer Fraud and Deceptive Business Practices Act, Ill.Rev.Stat. ch. 121 1/2, §§ 261-272, the Illinois Attorney General filed suit in the federal court for the Northern District of Illinois on behalf of the State of Illinois, which had purchased a 1977 Oldsmobile with a Chevrolet engine, and more than 100 other Oldsmobile purchasers. The complaint alleged that the sale of the Oldsmobiles without disclosure of their engine source violated the Magnuson-Moss Act, 15 U.S.C. § 2301-2312, and sought certification of the action as a nationwide class action. The Oswald and Miller actions were later brought to the federal district court and consolidated with the State of Illinois action before Judge McGarr. Upon GM's petition, the Judicial Panel on Multidistrict Litigation transferred seven actions then pending in other federal courts to the Northern District for consolidated pretrial proceedings.
The Magnuson-Moss Act limits federal court jurisdiction over class actions prosecuted under the Act to those actions in which the amount of each individual claim is at least $25, the total amount in controversy is at least $50,000, and the number of named plaintiffs is at least 100. 15 U.S.C. § 2310(d)(3). Otherwise, presumably every consumer complaint alleging a violation of the Act could have been maintained in the federal courts, without regard to the amount in controversy, under 28 U.S.C. § 1337. Compare Barnette v. Chrysler Corp., 434 F. Supp. 1167 (D.Neb. 1977) (individual action alleging a violation of the Act and seeking recovery of the purchase price of a defective car could not be maintained in federal court, because it failed to meet the $50,000 requirement). On the other hand, the Act's amount in controversy requirements, by lowering from the usual $10,000 to $25 the amount necessary for individual claims but requiring an aggregate amount of at least $50,000, reduce the obstacles normally encountered in meeting the jurisdictional amount necessary to maintain a class action. See Snyder v. Harris, 394 U.S. 332, 89 S.Ct. 1053, 22 L.Ed.2d 319 (1969); Zahn v. International Paper Co., 414 U.S. 291, 94 S.Ct. 505, 38 L.Ed.2d 511 (1973). The number of named plaintiffs required, however, remains a substantial barrier to maintaining class actions under the Act. It was enacted by Congress to prevent "trivial or insignificant" class actions from being brought in the federal courts. H.R. Rep. No. 93-1107, 93d Cong., 2d Sess., reprinted in [1974] U.S. Code Cong. Ad.News 7702, 7724. Although the Illinois Attorney General's complaint was the only complaint to satisfy the last jurisdictional requirement, we attach no particular significance to this fact.
General Motors characterizes the case before this court as "only the tip of a litigation iceberg" over GM's interdivisional engine use. The widespread publicity given to the engine switch by the initial lawsuits bred additional lawsuits. Other Attorneys General soon filed state court actions against GM under state consumer protection statutes. Furthermore, many individual car buyers started state court proceedings seeking individual and sometimes class relief. Altogether, GM estimates, over 300 engine interchange actions were filed against GM since March 1977. Forty-one of the suits were filed as class actions and thirty-three were brought by state Attorneys General. Some of the actions were initiated by purchasers of 1977 Buicks and Pontiacs which, like the Oldsmobiles in this suit, were equipped with Chevrolet engines. At least two suits were filed by owners of 1977 Buicks and Cadillacs, alleging that they received cars equipped with Oldsmobile engines. See In re GMC Engine Interchange Litigation, 441 F. Supp. 933 (J.P.M. D.L. 1977) (transferring actions to the Northern District of Illinois for consolidated pretrial proceedings). GM's interdivisional engine program also prompted investigation by the Federal Trade Commission. See GMC v. FTC, 1978-1 Trade Cas. ¶ 62,005 (N.D.Ohio 1977) (rejecting GM's challenge to the authority of the Commission to undertake the investigation). The bulk of the lawsuits, however, appear to involve 1977 Oldsmobiles, the subject of the litigation before this court.
The Oldsmobile actions that eventually were consolidated for pretrial proceedings are: State of Illinois v. GMC, No. 77-C-927 (N.D.Ill.); Oswald v. GMC, No. 77-C-1006 (N.D.Ill.); Miller v. GMC, No. 77-C-1436 (N.D.Ill.); Skokie Central Traditional Congregation v. GMC, No. 78-C-1457 (N.D.Ill.); State of Alabama ex rel. Baxley v. GMC, No. 77-P-0881-S (N.D.Ala.); Creel v. GMC, No. CA-77-P-0440-S (N.D. Ala.); Natter v. GMC, No. CA-77-P-0659-S (N.D.Ala.); Balog v. GMC, No. 77-443 (W.D. Pa.); Hannan v. GMC, No. 77-C-265 (E.D. Wis.); King v. GMC, No. M-77-24-CA (E.D. Tex.); Levine v. GMC, No. 77-C-849 (E.D.N. Y.); Parker v. GMC, No. S-77-0174(N) (S.D. Miss.).
The various federal actions were consolidated before the district court for pretrial purposes only. Although the actions have not been consolidated for trial purposes, the appellants do not contest, and we do not question, the district court's authority to approve a settlement of all the actions before it. See 15 C. Wright, A. Miller E. Cooper, Federal Practice and Procedure § 3866 at 374-76 (1976); Weigel, The Judicial Panel on Multidistrict Litigation, Transferor Courts and Transferee Courts, 78 F.R.D. 575, 582-83 (1978).
The order certifying the class action found that each of the named plaintiffs would adequately represent the class and confirmed the representative status of each. Therefore we need not decide whether all of the actions are technically before us, because we find that the appeal of some of the named plaintiffs is sufficient to permit this court to consider the interests of all class members. See also Part III of this opinion infra.
On July 22, 1977, the district court entered an order adopting an agreement of the numerous counsel for the plaintiffs in the consolidated cases. The order created an executive committee of six attorneys to represent the plaintiffs in all pretrial proceedings. See generally Manual for Complex Litigation §§ 1.92-1.93. Although the committee was given broad power in the pretrial proceedings, the order provided that the committee could conduct settlement negotiations only with the consent of all counsel for the named plaintiffs.
All citations in this opinion unless otherwise noted are to the Manual's fourth edition. Citations to particular pages follow the pagination of the Wright and Miller edition.
On October 13, 1977, the district court certified the consolidated cases as a class action. The order defined the class as "[a]ll persons . . . who purchased 1977 Oldsmobile automobiles which without their knowledge or consent, contained V-8 engines manufactured by the Chevrolet Motor Division . . ." The court dismissed all federal claims except the Magnuson-Moss claim and declined to exercise its power to take pendent jurisdiction over the related state law claims. The trial court recognized that parallel state court actions were pending, but rejected GM's position that the state proceedings should prevent class certification on the Magnuson-Moss claim. Despite the certification of the class, no notice to class members was mailed to inform them of the pendency of the class action at that time.
B. The Settlement
Sometime during the fall of 1977, General Motors entered into settlement negotiations with representatives of the various state Attorneys General who had filed or were contemplating filing actions against GM. A representative of the Illinois Attorney General who was also a member of the executive committee participated in the negotiations without leave of the district court or other counsel for the plaintiffs in the federal class action. On December 13, 1977, one of the counsel for the plaintiffs received word that a tentative settlement agreement had been reached by GM and the Attorneys General. The attorney, in essence, requested the district court to order immediate disclosure of the progress of the settlement negotiations or any agreements that had been reached. The trial court, however, regarded the motion as premature. Unwilling to interfere with communications between GM and the Attorneys General before an agreement was reached, the district court declined to order the requested relief. The trial judge remarked that he believed he had sufficient power over the approval of any settlement to protect the interests of class members.
GM maintains that the negotiations were begun at the suggestion of the Consumer Protection Committee of the National Association of State Attorneys General.
Six days later on December 19, the Illinois Attorney General in his capacity as one of the class counsel moved that the district court consider the settlement agreement between GM and all but five of the fifty state Attorneys General. The proposed settlement provided that GM would provide to each consumer who had purchased a 1977 Oldsmobile, Buick or Pontiac equipped with a Chevrolet engine on or before April 10, 1977, $200 plus a 36-month or 36,000-mile extended warranty on the power train. In return each purchaser would be required to sign a release of all state and federal claims concerning the substitution of engines, components, parts, and assemblies in the car. GM also agreed to disclose the source of all engines of new GM cars for the next three years. The Attorneys General, in turn, promised to secure dismissals with prejudice of all actions prosecuted by them.
Several other state Attorneys General have since joined in the agreement.
The district court showed itself willing to consider the agreement as a basis for settling the class action. Although the court afforded private counsel time to conduct discovery to determine whether the settlement was fair, it denied the motion of some of plaintiffs' counsel for discovery into the negotiations between the Attorneys General and GM. The court maintained that the negotiation process was irrelevant to the central issue of the fairness of the settlement.
Furthermore, the district court entertained GM's motion to redefine the class to include only those Oldsmobile purchasers to whom the settlement agreement contemplated payment. The class originally included all 1977 Oldsmobile purchasers who bought their cars before October 13, 1977, without knowledge that the cars had Chevrolet engines. The settlement agreement contemplated narrowing the class to purchasers before April 11, 1977. In an order dated March 14, 1978, the trial court denied GM's motion to redefine and narrow the class. The court did, however, designate "for purposes of sending the settlement notice" a subclass of pre-April 11 purchasers. Notices informing class members of the pendency of the class action were sent out shortly thereafter. The notice to settlement subclass members, in addition to informing them of the pendency of the action, informed them of the proposed settlement and gave them the opportunity, inter alia, to opt-out of the action or to object to the proposed settlement. The notice to class members not in the settlement subclass merely provided notice of the action and the opportunity to opt-out.
The trial court also agreed with GM to broaden the class in one respect. The court, for the purpose of settlement only, struck the no-knowledge-or-consent requirement of the original class certification as to members of the settlement subclass. This conformed the subclass to the precise class of Oldsmobile purchasers contemplated by the GM-Attorneys General agreement.
In May 1978, pursuant to its authority under Fed.R.Civ.P. 23(e), the district court held a fairness hearing to determine whether it should approve the settlement. Because some of the private counsel objected to the settlement, the hearing was contested and lasted twelve days. The order of proof was irregular. Both sides submitted numerous exhibits. The plaintiff-objectors presented, among others, several 1977 Oldsmobile owners who objected to the settlement and two mechanics who testified that the substituted power train was inferior to the one GM allegedly warranted. GM relied largely on exhibits and the testimony of a Chevrolet staff engineer who testified that the power trains warranted and those provided were comparable.
On July 17, 1978, after considering post-hearing memoranda of the various sides in the litigation, the district court entered an order approving the subclass settlement as fair. Adopting GM's proposed findings of fact almost verbatim, the district court found that the engines and other parts included in the Oldsmobiles were "comparable" to those warranted. Resolving most of the other contested issues in favor of GM, the district court ordered the action dismissed as to all members of the subclass and directed GM to send an approved notice of settlement to each member of the subclass. Before the notice could be mailed, however, some of the plaintiff-objectors prosecuted this appeal.
After the notice of appeal was filed, the Illinois Attorney General made a motion before the trial court requesting permission to send the settlement notice (with additional language indicating the pendency of the appeal) to subclass members. The trial court held that the appeal deprived it of jurisdiction to entertain the motion, but indicated that if it had had jurisdiction, it would have granted the motion. The Attorney General then, with the apparent acquiescence of the plaintiff-proponents and GM, moved this court for relief under Fed.R.App.P. 8(a). Because the contents of the notice were at issue on this appeal, we took the motion under advisement. Our decision on the merits of the appeal necessarily precludes sending out the notice in its present form. Accordingly, we hereby deny the motion.
II. Appealability
The plaintiff-objectors prosecuting this appeal and GM agree that this court has jurisdiction to hear this appeal. The attorney for one of the plaintiffs and an objector to the settlement before the trial court, however, maintains that the trial court's order approving the settlement is neither a final decision nor a collateral order within the meaning of 28 U.S.C. § 1291. Of course, we cannot determine this court's jurisdiction by majority vote of counsel appearing before us and, even if the parties unanimously agreed to appeal the order, we would be required to raise the issue sua sponte. Levin v. Baum, 513 F.2d 92 (7th Cir. 1975).
Disagreement between attorneys for the class, as will become apparent, has become the norm in the conduct of this litigation. For our purposes, counsel for the class may be divided into basically three groups. Those who objected to the proposed settlement in the trial court shall be referred to as plaintiff-objectors. Despite the division over the appealability issue, the attorney contesting the jurisdiction of this court to entertain the appeal is a member of this group. Those private counsel who supported the settlement shall be referred to as plaintiff-proponents. Finally, the Attorneys General from Illinois and Alabama who represented named plaintiffs in the trial court constitute the third group. The latter two groups have aligned themselves with GM on many of the issues in this appeal.
There is only one apparent obstacle to our hearing this appeal. The trial court's division of the class into two subclasses arguably makes this a multi-party action subject to the requirements of Fed.R.Civ.P. 54(b). In an order following its approval of the subclass settlement, the trial court refused to make a determination that there was no just reason for delay and to direct entry of judgment. We hold that, despite the refusal of the trial court to enter judgment pursuant to Rule 54(b), we have jurisdiction to review the order approving the subclass settlement as a collateral order.
There is considerable doubt whether Fed.R.Civ.P. 54(b) was intended to govern the situation when two distinct subclasses are created from a single class and one subclass' right to recover under a settlement neither affects nor is affected by the merits of the other subclass' claim. Aside from the difficulty of construing "multiple parties" to encompass separate subclasses, the settlement of one subclass' suit arguably should be treated as a separate lawsuit outside the ambit of Rule 54(b). This practical view of the position of the subclasses accords with the legal effect of creating subclasses under Fed.R.Civ.P. 23(c)(4). That rule provides that when a class is subdivided "each subclass [shall be] treated as a class, and the provisions of this rule shall then be construed and applied accordingly." Each subclass must independently meet the requirements of Rule 23 in order to be maintained as a class action, 7A C. Wright A. Miller, Federal Practice and Procedure § 1790 at 191-92 (1972), and therefore it seems consistent with the spirit of the rules to treat each subclass action as a separate action for all purposes.
Because we find that even if Rule 54(b) encompasses the present litigation that an independent basis for jurisdiction exists, we need not attempt to reconcile Rule 23 with Rule 54(b). Collateral orders are appealable without the express entry of judgment under Rule 54(b). See Swanson v. American Consumer Industries, Inc., 517 F.2d 555, 560-61 (7th Cir. 1975).
The Supreme Court has taken an "intensely practical" approach when deciding whether judgments are appealable. Mathews v. Eldridge, 424 U.S. 319, 331 n. 11, 96 S.Ct. 893, 47 L.Ed.2d 18 (1976). In close cases the determination must be made by balancing the "inconvenience and costs of piecemeal review" against "the danger of denying justice by delay." Gillespie v. United States Steel Corp., 379 U.S. 148, 152-53, 85 S.Ct. 308, 311, 13 L.Ed.2d 199 (1964). We are cognizant that the federal policy against piecemeal review admits no exception merely because the judgment appealed from affects the conduct of a class action. See Coopers Lybrand v. Livesay, 437 U.S. 463, 98 S.Ct. 2454, 57 L.Ed.2d 351 (1978) (striking the death knell for the death knell doctrine); Weit v. Continental Illinois National Bank Trust, 535 F.2d 1010 (7th Cir. 1976) (order requiring notice to class members is not a collateral order). We believe, however, that although the federal courts have narrowly interpreted the collateral order doctrine established in Cohen v. Beneficial Industrial Loan Corp., 337 U.S. 541, 69 S.Ct. 1221, 93 L.Ed. 1528 (1949), that this case falls within "that small class which finally determine claims of right separable from, and collateral to, rights asserted in the action, too important to be denied review and too independent of the cause itself to require that appellate consideration be deferred until the whole case is adjudicated." Id. at 546, 69 S.Ct. at 1226.
The first requirement of the collateral order doctrine is that the matter appealed from must have been finally determined by the district court. This does not require that the trial court be without power to reverse its ruling; it only requires that no further consideration be likely. 15 C. Wright, A. Miller E. Cooper, Federal Practice and Procedure § 3911 at 470 (1976). The record amply indicates the trial judge's resolve not to reconsider the fairness of the subclass settlement. After the long fairness hearing, the trial court approved the settlement in an order with fairly extensive findings of fact. The order purported to immediately dismiss the claims of all subclass members. Afterward, the trial court on two occasions declined to reconsider its decision. Moreover, although the trial court retained jurisdiction over the settlement subclass action to supervise the implementation of the settlement, this left the trial court with only the ministerial task of executing its judgment. The trial court's order, therefore, is not tentative and it finally determines the matter appealed to this court.
"There are two aspects of the final judgment rule. One is that the order be the final disposition of the entire case. The other is that the order be the final disposition of the issue. The Cohen rule permits a limited exception with respect to the first aspect but not with respect to the second." Rodgers v. United States Steel Corp., 508 F.2d 152, 159 (3d Cir.), cert. denied, 423 U.S. 832, 96 S.Ct. 54, 46 L.Ed.2d 50 (1975).
The second requirement of the collateral order doctrine is that the matter appealed must be "separable from, and collateral to, rights asserted in the action" and neither affect nor be affected by decision on the merits. 337 U.S. at 546, 69 S.Ct. at 1225-1226. Application of this requirement to appeals from decisions on the fairness of a settlement presents some difficulties. Ordinarily settlements of civil litigation are not reviewed by federal courts. Thus, the issue is raised almost exclusively in class or derivative actions. One court of appeals, however, has held that a refusal of a trial court to approve a class action settlement to be "collateral," Norman v. McKee, 431 F.2d 769 (9th Cir. 1970), cert. denied, 401 U.S. 912, 91 S.Ct. 879, 27 L.Ed.2d 811 (1971), and another has reviewed such a refusal without expressly considering the appealability issue, In re International House of Pancakes Franchise Litigation, 487 F.2d 303 (8th Cir. 1973).
Court approval of settlements is also necessary in bankruptcy reorganization proceedings. See, e. g., Protective Committee for Independent Stockholders of TMT Trailer Ferry, Inc. v. Anderson, 390 U.S. 414, 88 S.Ct. 1157, 20 L.Ed.2d 1 (1968).
The Second Circuit has recently rejected the position taken by the Eighth and Ninth Circuits and refused to review a trial court's refusal to approve a settlement of a shareholders derivative action. Seigal v. Merrick, 590 F.2d 35 (2d Cir. 1978). Because this appeal challenges the trial court's approval of a settlement, we need not align this court on one side of this conflict between the Circuits. This appeal because of the subclassing of Oldsmobile purchasers for the purposes of settlement presents a situation unlike those which ordinarily confront class members or shareholders after the trial court's approval or disapproval of a proposed settlement of a representative action. In Seigal the court stated that "[an approved] settlement . . . is not a deviation from the main path of the litigating process. It is a step on that path directly leading to final judgment. An approval of a compromise, after appropriate notice, becomes a final judgment." 590 F.2d at 38. In the case at bar, the trial court's approval of the subclass settlement does not lead directly to final judgment. But unlike a disapproval of a settlement, the trial court's order looks toward neither a renewal of settlement negotiations nor a trial on the merits. Thus, the danger of appellate court interference with proceedings before the trial court is small in comparison with the danger of denying justice by delay.
Although in Norman the court maintained that appellate review of the initial determination of the settlement's fairness was completely divorced from the merits of the claim, adequate review of the fairness of a settlement necessarily requires some examination of the underlying cause of action. 15 C. Wright, A. Miller E. Cooper, Federal Practice and Procedure § 3911 at 385 (1976); see Manual for Complex Litigation § 1.46 at 56. See also Coopers Lybrand v. Livesay, 437 U.S. at 469, 98 S.Ct. at 2458 ("the class determination generally involves considerations that are 'enmeshed in the factual and legal issues comprising the plaintiff's cause of action'"). Nevertheless, several factors bring this appeal within the separateness requirement. First, the Supreme Court has not applied the requirement that the issue be "separate" from the merits to require the precise division of the issues presented on appeal and the elements of the underlying cause of action that a semanticist might expect. See National Socialist Party v. Village of Skokie, 432 U.S. 43, 97 S.Ct. 2205, 53 L.Ed.2d 96 (1977). Moreover, to the extent that this appeal raises issues about the regularity of the conduct of the settlement negotiations or the fairness hearing, consideration of the merits of the plaintiffs' cause of action is unnecessary. Similarly, because appellate courts will reverse a trial court's determination on the fairness of a settlement only if there is a clear abuse of discretion, consideration of the merits is necessarily something less than penetrating.
Finally, the order approving the settlement is, in one sense, completely separate from the merits of the action. The trial court's approval of the settlement precludes any decision on the merits of the settlement subclass' claim because the claim will never go to trial.
The third requirement of the collateral order doctrine is that the rights asserted would be lost, probably irreparably, if review were delayed until the conclusion of proceedings in the district court. It is unlikely that the claims of the post-April 10, 1977, Oldsmobile purchasers will be decided any time soon. GM has made clear its intention not to settle with that subclass. Therefore years of litigation before the entire class action is concluded is possible. In the meantime, the settlement, if executed, contemplates the release of state and federal claims by those class members who accept the settlement package and dismissal of the Magnuson-Moss claims for those who do not. If the settlement is later undone on appeal, ordering reimbursement by those who accepted the $200 and received benefits under the mechanical insurance policy would be practically impossible. Those signing releases might also lose their state claims against GM because of the running of the statutes of limitation. Conversely, those who decline to sign the release, may file and pursue state claims. Any judgment in the state courts may possibly bar subsequent action on their Magnuson-Moss claims.
These characteristics of the settlement approved by the trial court distinguish this appeal from the appeal which was dismissed for lack of an appealable order in Rodgers v. United States Steel Corp., 541 F.2d 365 (3d Cir. 1976). In Rodgers the trial court permitted the defendant to communicate to individual members of the class an offer to enter into individual settlements. See Rodgers v. United States Steel Corp., 70 F.R.D. 639 (W.D.Pa. 1976). See also Part VI of this opinion infra. The trial court merely approved the communication of the offer; it did not finally determine the rights of any member of the class. See 541 F.2d at 370. In the present case, the trial court dismissed the federal claims of all settlement subclass members and effectively terminated their participation in the class action whether they released their claims or not. Moreover, the settlement offer in Rodgers merely promised payment of back pay in return for signed releases. The Court of Appeals, dismissing the appeal, noted that the parties could be returned to their original positions if the release was subsequently invalidated. Id. at 371. Here, we cannot say with any degree of certainty that we could later return to GM the benefits that class members received under the mechanical insurance policy.
We conclude that "delay of perhaps a number of years in having [their] rights determined might work a great injustice" to the subclass members. Gillespie v. United States Steel Corp., 379 U.S. 148, 153, 85 S.Ct. 308, 311, 13 L.Ed.2d 199 (1964). They "cannot make important decisions about . . . further participation in this suit without having [their] rights determined now." Diaz v. Southern Drilling Corp., 427 F.2d 1118, 1123 (5th Cir.), cert. denied, 400 U.S. 878, 91 S.Ct. 118, 27 L.Ed.2d 115 (1970). The possibility that later appellate review would be effective is simply too slight.
Cf. Pettway v. American Cast Iron Pipe Co., 576 F.2d 1157, 1221 (5th Cir. 1978), cert. denied, ___ U.S. ___, 99 S.Ct. 1020, 59 L.Ed.2d 74 (1979):
The court's November 20 order required awardees wishing to opt into the settlement to do so by December 15, 1975 or be deemed to have opted out of the subclass. This created a dilemma for dissatisfied subclass members, who were faced with the equally unpalatable alternatives of opting into a possibly invalid settlement or being relegated to individual lawsuits. A decision to opt into the settlement by endorsing the back pay check and thereby releasing the company of all liability for past discrimination might preclude entitlement to a share in a new agreement or award if the settlement were invalidated on appeal. On the other hand, a decision to opt-out of the subclass by failing to cash the tendered check would create the possibility of receiving no back pay award if the appeal were unsuccessful and an individual lawsuit proved unrealistic. . . .
The procedure adopted by the district court, by requiring claimants to choose whether or not to opt into the settlement before they could exercise their right to appellate review, unfairly burdened the rights of awardees to appeal the settlement and thereby significantly undermined one of the most important procedural protections associated with the approval of a settlement. We hold that the ability of subclass members to opt into a back pay settlement may not be terminated before a final determination of the propriety of that settlement is made.
A final requirement of the collateral order doctrine is that the order must present "important and unresolved legal questions." Weit v. Continental Illinois National Bank Trust Co., 535 F.2d 1010, 1015 (7th Cir. 1976); Weight Watchers, Inc. v. Weight Watchers International, Inc., 455 F.2d 770, 773 (2d Cir. 1972). We think this appeal raises at least two important questions concerning the proper balance between the general policy of encouraging settlements and a court's specific duty to insure the fairness of class action settlements. The first question involves the scope of discovery which should be afforded to objectors to proposed class settlements which were negotiated under questionable circumstances. Because adequate representation is the foundation of all representative actions, see Fed.R.Civ.P. 23(a)(4), Hansberry v. Lee, 311 U.S. 32, 61 S.Ct. 115, 85 L.Ed. 22 (1940), we think this question is appropriately reviewed at this time. The second question concerns the nature of the "settlement" that Rule 23(e) authorizes the trial court to approve. Because this question goes to the power of the district court in the settlement of representative actions, we believe it is sufficiently important to receive appellate consideration now.
In conclusion, the trial court's order is not tentative; it is capable of review without extensive examination of the merits; it raises issues which could not be effectively reviewed later; and it presents important, unresolved legal questions for consideration by this court. We hold that the trial court's order approving the subclass settlement is an appealable collateral order.
III. Motion to Limit the Appeal
Before oral argument, the attorney representing the State of Alabama in this litigation presented to this court a "motion to limit appeal to certain named appellants." The motion seeks to have the effect of this court's decision limited to (1) only the named plaintiffs, Oswald and Miller, the plaintiff-objectors prosecuting this appeal or, alternatively, (2) only those class members who filed objections to the proposed settlement in the district court. We consider the arguments in support of the second alternative first.
It is argued that this court's decision in Research Corp. v. Asgrow Seed Co., 425 F.2d 1059 (7th Cir. 1970), compels this court to restrict the representative standing of the named plaintiffs who prosecute this appeal to those class members who objected to the settlement in the trial court. In Research, the appellants were members of a defendant class represented in the district court by numerous named defendants. Despite adequate notice, the appellants failed either to request exclusion from the defendant class or to object to a proposed settlement negotiated by the named defendants; the appellants attacked the fairness of the settlement for the first time on appeal. This court held that the failure of the appellants to intervene in the action foreclosed their right to appeal. Here it is argued by analogy that each individual subclass member who failed to object to the settlement before the trial court has waived the right to appeal and the right to be represented by others on appeal. We think the argument is without merit.
There is no doubt that the named plaintiffs, Oswald and Miller, preserved the right to appeal. They are parties to the lawsuit; intervention was obviously unnecessary. Moreover, through their attorneys they vigorously objected to the settlement in the district court and created a record adequate for appellate review. Thus, the issue raised by the motion may be refined to whether Oswald and Miller through their counsel may represent the interests of absent subclass members on this appeal.
We would be reluctant to hold that absentee class members waive appellate review merely because they failed to take affirmative action when their interests were already being adequately represented by participants in the lawsuit. Cf. Ace Heating Plumbing Co. v. Crane Co., 453 F.2d 30, 32-33 (3d Cir. 1971) (objectors' failure to opt-out of a class action does not preclude appellate review). To do so would unnecessarily restrict the representational character of all class actions. We need not reach the issue here, however, because the notice of the proposed subclass settlement informed subclass members that if they neither opted out of the subclass nor intervened in the lawsuit that "attorneys for the named plaintiffs will represent your interest in these suits." We think subclass members who received the notice could reasonably rely on class counsel to protect their interests by prosecuting an appeal from the judgment of the district court if necessary. See Gonzales v. Cassidy, 474 F.2d 67 (5th Cir. 1973) (failure to appeal approval of an unfair settlement constitutes inadequate representation). We therefore decline to hold that absentee subclass members waived their right to have the settlement reviewed by this court.
The second argument advanced in favor of limiting the representative capacity of the plaintiff-objectors on this appeal is that the pretrial order of the trial court vested the power to conduct all pretrial actions on behalf of the class in the attorneys' executive committee. Because the executive committee did not authorize the prosecution of the appeal, it is argued, the authority of counsel for the plaintiff-objectors must be confined to representing the individual named plaintiffs before this court.
We question initially the premise that it is the attorney, not the named plaintiff, who possesses the power to appeal the approval of a settlement. "[T]he decision to appeal a class action judgment must rest with the class plaintiffs," not class counsel. Pettway v. American Cast Iron Pipe Co., 576 F.2d 1177-78 (5th Cir. 1978), cert. denied, ___ U.S. ___, 99 S.Ct. 1020, 59 L.Ed.2d 74 (1979). Since the pretrial order did not purport to restrict the representative capacity of the named plaintiffs prosecuting this appeal, it would seem that the argument misses the mark. The court in Pettway, however, acknowledged that "no clear concept of the allocation of decision-making responsibility between the attorney and the class members has yet emerged." Id. at 1176. Consequently, assuming arguendo the premise that the class attorney is the dominus litus, we consider and reject the argument that the pretrial order prohibits counsel for Oswald and Miller from representing the interests of the class before this court.
The pretrial order does not on its face vest the power to appeal in the executive committee. The order itself only lists the committee's various duties and powers relating to pretrial proceedings. We would be extremely reluctant to imply a provision that restricts the right to appeal decisions of the trial court. Furthermore, even if the pretrial order contemplated giving the executive committee the power to prohibit individual attorneys from appealing, whether the executive committee has done so is unclear. The minutes of the committee meeting show that the committee did pass a motion that no appeal be taken from the trial court's approval of the settlement. Nevertheless, those minutes also indicate that before passage of the motion "[t]he chair ruled that the motion does not proclude [ sic] anyone from appealing but states the position of the majority of plaintiffs' counsel."
We believe that the question of whether an appeal should be made and the scope of that appeal should be answered by determining the best interests of the class. The plaintiff-proponents maintain that the settlement is fair, that the approval of the trial court is correct, and that the matter is best left unreviewed by this court. Plaintiff-objectors, of course, disagree. The purpose of Fed.R.Civ.P. 23(e) is to protect the interests of absentee class members; the danger of abuse is high and the protection of their interests cannot be left to class counsel alone. Rule 23 imposes on the trial court in the first instance, and on this court eventually, the duty to examine the fairness of proposed settlements. Limiting the representative capacity of the appellants on this appeal would effectively negate this court's obligation to act as the guardian of the class. We do not believe that the interests of class members are best served by leaving the settlement unreviewed. Cf. McDonald v. Chicago Milwaukee Corp., 565 F.2d 416, 417 n. 1 (7th Cir. 1977) (permitting briefs and oral arguments by parties who failed to file a separate notice of appeal because the case involved "issues inextricably bound up with" those properly before the court). Restricting the appeal would only leave the door open to additional individual appeals by those who decline to accept the settlement offer. A series of individual and possibly conflicting appellate decisions on the propriety of the settlement would undermine the representative nature of class actions significantly and sacrifice the public's interest in judicial economy unnecessarily. We hold that plaintiff-objectors Oswald and Miller are parties who through their counsel will fairly and adequately protect the interests of the class in this appeal. See Fed.R.Civ.P. 23(a)(4) (requirement for class certification).
We do not hold "that each individual plaintiff and lawyer must be permitted to do what he pleases in litigation as complex as this, and can behave in total disregard of the interest of other litigants and of the class . . . ." Farber v. Riker-Maxson Corp., 442 F.2d 457, 459 (2d Cir. 1971). We note the following factors which convince us that the interests of the class will be well represented on this appeal. Cf. Pettway v. American Cast Iron Pipe Co., 576 F.2d 1157, 1178-80 (5th Cir. 1978), cert. denied, ___ U.S. ___, 99 S.Ct. 1020, 59 L.Ed.2d 74 (1979) (discussing factors relevant to determining whether the named plaintiff may appoint new counsel to appeal the approval of a settlement negotiated by former class counsel). First, the named plaintiffs and their counsel were among the first to file engine switch suits against GM. Second, counsel for the appellants was a member of the class executive committee and is well acquainted with the litigation. Despite suggestions and innuendoes of ulterior motives in some of the briefs which we can only regard as symptoms of "the 'brief writer's hyperbole' syndrome," United States ex rel. Sims v. Sielaff, 563 F.2d 821, 824 n. 6 (7th Cir. 1977), nothing in the record indicates that appellants' counsel has acted with other than the best interests of the class in mind. Third, although vocal objection to the settlement among class members was not widespread, "the sentiment of the class is but one factor in our analysis of the appealability question." Pettway, 576 F.2d at 1178. In Patterson v. Stovall, 528 F.2d 108 (7th Cir. 1976), this court heard the appeal of objectors to a class action settlement even though the objectors constituted only .0018% of all class members and their claims constituted only .0022% of all claims. Id. at 109 n. 1. See also Mandujano v. Basic Vegetable Products, Inc., 541 F.2d 832 (9th Cir. 1976) (reversing settlement even though only 4% of the class was in active opposition to it). Fourth and finally, we find that the issues raised on appeal are far from meritless.
In Patterson v. Stovall, 528 F.2d at 109 n. 1, we noted:
"Although in terms of the class and settlement [appellants'] number and size might be considered miniscule, the serious issues raised before this Court are not reduced in their magnitude."
We conclude that the best interests of the class warrant that this court review the fairness of the settlement as it affects the entire class. Consequently, we consider the merits of the objections to the trial court's approval of the proposed settlement.
IV. Conduct of the Settlement Negotiations
The plaintiff-objectors challenge the refusal of the trial court to permit them to conduct discovery into the settlement negotiations. They contend that the trial court's order prohibiting discovery and the court's limitation of examination of the Assistant Illinois Attorney General during the fairness hearing prevented them from being able to determine whether the proposed settlement was fair, reasonable and adequate. The trial court's order limiting discovery evidences its belief that how the settlement was reached was irrelevant to the issue of the fairness of the settlement. The court's findings of fact, although finding the irregular method of negotiating the settlement did not prejudice subclass members, reaffirmed the court's belief that the objection was irrelevant to the adequacy of the settlement "and would not constitute sufficient grounds to withhold an otherwise fair settlement from consideration by the subclass members."
The plaintiffs' second set of interrogatories requested that GM identify all documents that it relied upon during the course of the negotiations. The interrogatories also asked GM to state "the highest demand made by the various State Attorneys General in the course of the negotiations with defendant and identify all factual support for such demand, as well as any documents which relate to such demand or factual support." The trial court entered an order ruling that the process of the negotiations was not open to discovery. During the fairness hearing, although the court permitted some questioning of the Assistant Illinois Attorney General about the time, place and other aspects of the negotiations, it refused to permit inquiry into what transpired during the negotiations.
GM maintains that the plaintiff-objectors waived this issue by failing to recall the Assistant Illinois Attorney General after being given the opportunity to do so. The record, however, clearly indicates that, given the trial court's limitation on the scope of examination, any further questioning by the objectors would have been futile. The objectors brought the issue to the attention of the trial court and cannot be deemed to have waived it.
We think that the conduct of the negotiations was relevant to the fairness of the settlement and that the trial court's refusal to permit discovery or examination of the negotiations constituted an abuse of discretion. In addition, we do not think that the record adequately supports the court's conclusion that the seemingly irregular conduct of the negotiations did not prejudice the interests of the class. We must, therefore, reverse the trial court's order approving the settlement.
Neither GM nor the Illinois Attorney General has argued that the conduct of the settlement negotiations is protected from examination by some form of privilege, and we find no convincing basis for such an objection here. Although particular documents or discussions conceivably could be immune from discovery as attorney work product or as privileged attorney-client communications, the existence of such privileges is best determined in the context of particular demands for discovery. Inquiry into the conduct of the negotiations is also consistent with the letter and the spirit of Rule 408 of the Federal Rules of Evidence. That rule only governs admissibility. It simply bars admission of evidence of compromise negotiations to prove liability or damages and expressly provides that it "does not require exclusion when evidence is offered for another purpose. . . ." The rule is grounded on the policy of encouraging the settlement of disputed claims without litigation. That policy is not undermined by our decision here. Participants in negotiations to settle class actions are aware that Rule 23(e) requires the trial court's approval of any settlement reached. Moreover, they are or should be aware that the court will inquire into the conduct of the negotiations. See Manual for Complex Litigation § 1.46 at 53-54. To the extent such inquiry discourages settlements, it should only discourage those negotiated in circumstances so irregular as to cast substantial doubt on their fairness.
This court has several times commented on the trial court's continuing duty to undertake a stringent examination of the adequacy of representation by the named class representatives and their counsel at all stages of the litigation. McDonald v. Chicago Milwaukee Corp., 565 F.2d 416, 419 (7th Cir. 1977); Susman v. Lincoln American Corp., 561 F.2d 86, 89-90 (7th Cir. 1977). The trial court's duty to undertake such an inquiry arises from the requirement that it find that "the representative parties will fairly and adequately protect the interests of the class." Fed.R.Civ.P. 23(a)(4). The trial court's duty is heightened by its responsibility to review the fairness of any compromise of the class action. Id. 23(e).
"Before approving a settlement, therefore, the judge must assure himself that the class has been adequately represented during the settlement talks, a conclusion which will not follow automatically from a finding of adequacy for litigation purposes." Developments in the Law — Class Actions, 89 Harv.L.Rev. 1318, 1537-38 (1976). See also Wolfram, The Antibiotics Class Actions, 1976 A.B. Foundation Research J. 251, 361.
The Manual for Complex Litigation provides that inquiry into the conduct of settlement negotiations is pertinent to the court's examination of the settlement. Manual for Complex Litigation § 1.46 at 53-54. It recommends that before sending a notice to class members of a proposed settlement and before considering the substantive fairness of the settlement, the trial court should conduct a preliminary hearing to determine whether the proposed settlement is "within the range of possible approval." Id. Among the questions which merit judicial examination at the "probable cause hearing," the Manual lists:
We recognize that the Manual does not provide "an inflexible formula or mold into which all . . . pre-trial procedure must be cast." Manual for Complex Litigation at xix; see McDonald v. Chicago Milwaukee Corp., 565 F.2d 416, 420 (1977). In appropriate cases, however, the Manual does provide a rough guide by which to measure whether the trial judge acted within his discretion. We rely on it in that manner here.
Who were the negotiating parties and to what extent were they authorized to proceed with the settlement of their class' claims and possibly those of other classes?
Manual for Complex Litigation § 1.46 at 53 (Consideration 4).
Among the reasons for examining whether settlement negotiations were authorized is the danger of defendant "attorney-shopping."
[A] person who unofficially represents the class during settlement negotiations must be under strong pressure to conform to the defendants' wishes . . .. [A]n individual, lacking official status, knows that a negotiating defendant may not like his "attitude" and may try to reach a settlement with another member of the class.
Id. at 59 quoting Ace Heating Plumbing Co. v. Crane Co., 453 F.2d 30, 33 (3d Cir. 1971). Thus, unauthorized settlement negotiations create the possibility of negotiation from a position of weakness by the attorney who purports to represent the class. In addition, the prestige attendant upon negotiating a large settlement against a corporate defendant and thereby acquiring reputations as consumer advocates may place public attorneys in a situation analogous to private counsel who hope to win large fee awards. The possibility of such a conflict of interest as a general rule warrants judicial scrutiny of unauthorized settlement negotiations. Furthermore, settlement negotiations with less than all class counsel weaken the class' tactical position even if the attorney who enters into the negotiations attempts to represent the class' interests vigorously.
The court, to be sure, will not approve a settlement if it is unfair, but "fairness" may be found anywhere within a broad range of lower and upper limits. No one can tell whether a compromise found to be "fair" might not have been "fairer" had the negotiating [attorney] possessed better information or been animated by undivided loyalty to the cause of the class. The court can reject a settlement that is inadequate; it cannot undertake the partisan task of bargaining for better terms. The integrity of the negotiating process is, therefore, important.
Haudek, The Settlement and Approval of Stockholders' Actions — Part II: The Settlement, 23 Sw.L.J. 765, 771-72 (1969).
Cf. Developments in the Law — Class Actions, 89 Harv.L.Rev. 1318, 1552 (1976) (noting the conflict of interest created not only by counsel seeking large fees after settlement, but also by counsel pursuing "his own ideological goals without regard to the desires of class members").
A time-honored litigating tactic for a defendant encircled by multiple claimants is to weaken the total force of the attack little by little. The defendant first enters into settlements with the strongest of the plaintiffs. Then it faces the remaining plaintiffs, now isolated and abandoned, with the threat of long and lonely litigation to force a final round of settlements at terms favorable to the defendant.
Wolfram, The Antibiotics Class Actions, 1976 A.B. Foundation Research J. 251, 264.
Finally, unauthorized settlement negotiations deny other class counsel access to information about the negotiations which is helpful in evaluating the fairness of the settlement. "[T]he options considered and rejected, the topics discussed, the defendant's reaction to various proposals, and the amount of compromise necessary to obtain a settlement" were all matters which class counsel excluded from the negotiations needed to consider before exercising their fiduciary duties to the class by accepting the settlement.
Developments in the Law — Class Actions, 89 Harv.L.Rev. 1318, 1562 (1976).
Cf. Girsh v. Jepson, 521 F.2d 153, 157 (3d Cir. 1975):
It is little comfort to objector Frackman that plaintiffs' counsel may have examined the documents sought by objector during the course of . . . discovery. As an objector, Frackman was in an adversary relationship with both plaintiffs and defendants and was entitled to at least a reasonable opportunity to discovery against both.
See also National Conference of Commissioners on Uniform State Laws, Proposed Uniform Class Action Act § 12(c)(4) reprinted in 32 Bus.Law. 83, 94 (1976) (notice of proposed settlement to class members shall include "a description and evaluation of alternatives considered by representative parties").
The record before this court contains facts which cast some doubt on the adequacy of the representation of the class during the settlement negotiations and the fairness of the resulting settlement. These facts warranted in this instance more probing into the conduct of the settlement negotiations than the trial court permitted.
The record establishes that the settlement presented to the court by the Illinois Attorney General was either (1) negotiated without the permission of the other class counsel in the federal action as required by the court's first pretrial order or (2) negotiated by the Attorney General's office in a capacity other than class counsel in this action. The pretrial order prohibited the class counsel executive committee from entering into settlement negotiations without the consent of all plaintiffs' attorneys. The Attorney General's Assistant was a member of the committee and therefore subject to the pretrial order's restrictions. Nevertheless, he participated in negotiations with GM without the consent of other counsel.
If the negotiations did proceed in violation of the trial court's pretrial order, we think that the plaintiff-objectors were entitled to discovery to determine whether the negotiations may have prejudiced the interests of the class. Moreover, even if discovery failed to reveal identifiable prejudice, the exclusion of the private counsel from the settlement negotiations should weigh heavily against approval of the settlement. "[T]he excluded plaintiff might well have improved the settlement terms, and while this may be hard to demonstrate, the proponents of the compromise should not be helped by a difficulty of proof created by their improper conduct." Haudek, The Settlement and Approval of Stockholders' Actions — Part II: The Settlement, 23 Sw.L.J. 765, 770 (1969).
The trial court found that at least some private counsel knew of the negotiations between GM and the Attorneys General in advance of the settlement. The knowledge of some counsel, however, falls short of the authorization contemplated by the trial court's pretrial order. That order authorized the class counsel executive committee to conduct negotiations, but only with the consent of all counsel for the named plaintiffs. The trial court made no finding that all class counsel were aware of the negotiations between GM and the Attorneys General. Moreover, knowledge of the existence of the negotiations does not necessarily indicate consent to the negotiations for the purpose of settling the federal action. We do not question the right of the state Attorneys General to settle their parallel state lawsuits against GM without the approval of private counsel in the federal class action. Their authority to do so is unquestioned even though the settlement of state actions may have some collateral impact on the federal action, e. g., reducing the size of the class by affording relief to some class members. Here, however, the negotiations were conducted not only to settle the state actions, but also to settle the federal class action. We find no indication in the record that private counsel were aware that the negotiations would have such a broad effect until immediately before the announcement of the GM-Attorneys General agreement.
After the submission of the proposed settlement agreement, six of the private counsel in the federal action did agree to support the settlement. The district court relied on the plaintiff-proponents' support as a factor indicating both the absence of prejudice from the circumstances of the settlement's negotiation and the settlement's fairness. See Manual for Complex Litigation § 1.46 at 53 (Consideration 5). The support of some private counsel after being presented with the agreement as a fait accompli does not amount to a ratification of the conduct of the negotiations. As noted supra, class counsel should know the options considered and the topics discussed during the negotiations before supporting a settlement as fair. In the absence of such familiarity of counsel with the conduct of the settlement negotiations, the inference of fairness to be drawn from their support is weak. Cf. id. at 64 ("a plan should not be approved simply because counsel on both sides recommend it").
Thus, although the proponents of any class settlement always bear the burden of proof on the issue of fairness, Manual for Complex Litigation § 1.46 at 56, proponents who improperly negotiate a settlement should bear the heavier burden of establishing fairness by clear and convincing evidence. This does not unduly hamper settlements since the disapproval of the settlement always permits the renewal of negotiations between all of the proper participants in the class action. The question of prejudice aside, it is clear that the trial court did not require the proponents of the settlement proposed here to meet such a heavy burden. In fact, the trial court accepted the proposed settlement as prima facie fair and shifted to the objectors at least the burden of producing evidence disproving the fairness of the settlement. Whether the trial court shifted the burden of persuasion to the objectors as well is unclear. The objectors complain that it did and the Illinois Attorney General's brief seems to concede the point. The trial court's conclusions of law, however, recite that it placed the burden of persuasion on the proponents. Our comparison of the record with the findings of fact leads us to believe that as to some of the court's findings that it may indeed have misplaced the burden.
The Assistant Illinois Attorney General maintains, however, that his participation in negotiations between the state Attorneys General and GM did not violate the pretrial order because he was not negotiating as a class representative in the action in the federal court but rather was negotiating as a representative of the State of Illinois in the parallel state proceedings in the Circuit Court of Cook County. The motion of the Illinois Attorney General for leave to file the settlement took this position also, although the motion's first paragraph based the Attorney General's capacity to present the motion on his status as counsel for the State of Illinois, one of the designated class representatives in the federal action. Also consistent with his position that he did not participate in the settlement negotiations as a federal class representative, the Assistant Attorney General admitted during the fairness hearing that the Illinois Attorney General's office did not obtain consent to the settlement from the over 100 named private plaintiffs that the Illinois Attorney General represented in the federal action.
During the fairness hearing, Mr. Mulack, the Assistant Illinois Attorney General, described his position as one in which he wore "two hats."
[T]he Attorney General filed a State Court action . . . in the Circuit Court of Cook County, on March 7th of 1977. . . . [T]wo weeks later we filed the Federal Action. So as I told the Court, on several occasions, as we had appeared here during the motions on behalf of the class certification, I was wearing two hats — and the Attorney General of Illinois was, likewise, wearing two hats; one as a plaintiff, under the State Court action, under the Consumer Fraud Act, in the Circuit Court of Cook County, and the other as a punitive class representative in the Federal Court Action. . . .
I was perfectly aware of the limitations in Pre-Trial Order No. 1, that prohibited either myself or any representative of the Attorney General's office from taking part in nationwide negotiations on this particular class action. With that particular concern and that understanding, I had approached the posture of the overall negotiations.
Now, attendant at those meetings were Assistant Attorneys General literally from every state that had a major action going against General Motors. Each of those Attorneys General were there in their state capacity only — they were only concerned about their state lawsuits, as I was concerned, only, about my state lawsuit.
At the opening salvo — the opening introductions of the settlement negotiations — as people were being introduced, and from which state they attended, and as General Motors' attorneys were being introduced, as I was being introduced, I made this caveat on the record, that "I'm here only as an Assistant Attorney General on behalf of the State of Illinois case; I am not here, at all, as any class representative, or on behalf of the nationwide action; and if any discussions are brought up about the nationwide class action, I cannot participate, because that is not my function." With that caveat, we proceeded to discuss those particular matters attendant to the settlement.
We note that the written settlement agreement between GM and the Attorneys General devoted much space and went into considerable detail reciting the rights and obligations of the parties to the negotiations with respect to the settlement of the federal action. For example, the agreement, mentioning the Magnuson-Moss class action by name, required the Attorneys General, inter alia, to seek amendment of the class certification to conform with that group of consumers to whom GM would extend its offer, to represent to the trial court that the proposed settlement was fair and reasonable, and to recommend that the court approve the settlement of the entire action in accordance with the terms of the agreement.
The State of Illinois is a representative party in this suit solely because it purchased a 1977 Oldsmobile with a Chevrolet engine. The Illinois Attorney General's ability to maintain the suit on Illinois' behalf as a class action is governed solely by Rule 23. In the absence of statutory authorization, Illinois cannot maintain this action in federal court as a parens patriae action. Assuming arguendo that the Attorney General's office did not violate the pretrial order and thus participated in the negotiations solely as a representative in the parallel state court action, we believe, nevertheless, that the trial judge should have opened up the negotiations to scrutiny, if only to dispel the questions which naturally arise from the unusual posture of the case. If the settlement was not negotiated by authorized class counsel in the capacity of class counsel in this action, then it was negotiated in the name of, at best, only one of the named plaintiffs in the federal action, the State of Illinois. This stretches the theory of representation of absentee interests by the named plaintiff to its limit and warrants searching judicial examination of the circumstances surrounding and the matters discussed during the settlement negotiations before acceptance of the proposed settlement for possible approval.
See State of Iowa v. Union Asphalt Roadoils, Inc., 281 F. Supp. 391, 401-02 (S.D.Iowa 1968); State of Minnesota v. United States Steel Corp., 44 F.R.D. 559, 576 (D.Minn. 1968).
Cf. Hawaii v. Standard Oil Co., 405 U.S. 251, 266, 92 S.Ct. 885, 893, 31 L.Ed.2d 184 (1972) (" Parens patriae actions may, in theory, be related to class actions, but the latter are definitely preferable in the antitrust area. Rule 23 provides specific rules for delineating the appropriate plaintiff-class, establishes who is bound by the action, and effectively prevents duplicative recoveries").
The class action, although it also provides a vehicle for furthering the substantive policies behind legislation, is primarily a device to vindicate the rights of individual class members. We also note that the Magnuson-Moss Act does provide that the United States Attorney General and the Federal Trade Commission may go to federal court to enjoin violations of the Act. 15 U.S.C. § 2310(c). Thus the Act provides its own mechanism for protecting the general public's interest in enforcement of its provisions. It does not leave protection of the public interest up to the Attorneys General of the fifty states. Compare 15 U.S.C. § 15a-15h (explicitly vesting power in state Attorneys General to maintain actions against persons engaged in anti-competitive practices which harm state consumers).
In their briefs and during oral argument the parties devoted a good deal of time to a discussion of whether a settlement could be approved over the objections of some of the named plaintiffs. We agree with General Motors that the unanimous approval of all named plaintiffs is not a prerequisite to judicial approval of a settlement approved by some of the named plaintiffs. See McDonald v. Chicago Milwaukee Corp., 565 F.2d 416 (7th Cir. 1977). This case does not present, and we need not here decide, GM's admittedly extreme position taken during oral argument that the trial court can approve a settlement offered unilaterally by a class action defendant with the approval of neither a class representative nor class counsel. Here, at least the State of Illinois, a named plaintiff, agreed to settle.
Several additional facts suggest that the representation of the class during the negotiations was less than vigorous. The class settlement was reached relatively early in the course of the action. The federal action had been filed about nine months before; the class had been certified only two months before; and notice to class members of the pendency of the action had not even been mailed. Although discovery had commenced, GM's answers to many of the requests were less than completely responsive. Moreover, because the proposed settlement contemplated the release of all claims relating to component substitutions, not just the engine interchanges, the range of possible damages to class members was unclear. It is not possible to tell from the record how fully informed the Attorneys General may have been about the value of the claims they were surrendering.
See Manual for Complex Litigation § 1.46 at 53 (Consideration 1).
Id. (Considerations 2 3). The record does not reveal and the briefs of the parties do not detail the extent to which the Attorneys General had proceeded with discovery in their parallel state actions or whether they examined the value of the claim for the entire power train. The trial court's order precluding discovery of the conduct of the settlement negotiations, of course, prevented the objectors from making such a record. To this day, we have no idea how the participants in the negotiations arrived at the settlement package of $200 plus the extended power train warranty.
Not only was the settlement arguably hasty, but also the settlement agreement contemplated the abandonment of the prosecution of the claims of post-April 10 class members. The settlement agreement entered into by the Attorneys General obligated them to seek settlement of the entire class action even though the agreement obligated GM to offer payments to only part of the certified class. The agreement contemplated narrowing the class certified to those who purchased Oldsmobiles before April 11, 1977, despite the original certification of the class to include those who purchased before October 13. GM subsequently formally moved the court for such a revised class definition to conform to the settlement agreement. The court denied GM's motion, but did decide to create a subclass for settlement purposes. Although the abandonment by the Attorneys General of the claims of post-April 10 purchasers does not by itself warrant the reversal of the settlement of the claims of the pre-April 11 purchasers, it does indicate that the representation during the negotiations may have been inadequate as to all Oldsmobile purchasers who constituted the original class.
See id. at 54 (Consideration 6).
We must note that the means by which the trial court attempted to create a subclass also may have seriously jeopardized the rights of the post-April 10 purchasers. Aside from the tactical disadvantage of having their claims separated from the claims of the other class members, the subclassing technique chosen by the court raises doubts about whether those outside the ambit of the settlement could maintain a class action after the settlement with the pre-April 11 subclass.
The trial court has broad discretion in determining whether to allow a class action to be maintained, Jimenez v. Weinberger, 523 F.2d 689 (7th Cir. 1975), cert. denied, 427 U.S. 912, 96 S.Ct. 3200, 49 L.Ed.2d 1204 (1976), King v. Kansas City Southern Industries, Inc., 519 F.2d 20 (7th Cir. 1975), and must necessarily have an equally broad range of discretion in determining whether to create subclasses pursuant to Fed.R.Civ.P. 23(c)(4)(B). Division of a class or potential class into subclasses to account for differences in proof that may be required at trial is clearly permissible. See, e. g., Dorfman v. First Boston Corp., 62 F.R.D. 466, 476 (E.D. Pa. 1973) (creating subclasses to account for differences between class members who purchased before and after relevant information received wide public circulation). The trial court's discretion, however, is bounded by the requirements of the applicable law and in this case we believe that the trial court overstepped the bounds of the Federal Rules of Civil Procedure.
The trial court's order creating the settlement subclass did not conform to the requirements of Rule 23 which provides in pertinent part that when appropriate "a class may be divided into subclasses and each subclass treated as a class, and the provisions of this rule shall then be construed and applied accordingly." Fed.R.Civ.P. 23(c)(4)(B). The rule contemplates that at least two subclasses will be formed and requires that each independently meet the requirements of Rule 23 for the maintenance of the class action. See Monarch Asphalt Sales Co. v. Wilshire Oil Co., 511 F.2d 1073, 1077 (10th Cir. 1975). The trial court made no finding that the post-April 10 subclass could be maintained as a class action. The record shows instead that the trial court attempted to create a single subclass for the settlement, leaving the post-April 10 purchasers in the original class. No attempt was made to test whether the nonsettlement subclass action met the requirements of Fed.R.Civ.P. 23(a) (b). Furthermore, the record does not indicate whether any named plaintiff in the current action is even in the nonsettlement subclass. The subclass could be "headless," thus raising serious questions about whether the trial court could proceed to consider the post-April 10 claims. See generally Winokur v. Bell Federal Savings Loan Association, 560 F.2d 271 (7th Cir. 1977), cert. denied, 435 U.S. 932, 98 S.Ct. 1507, 55 L.Ed.2d 530 (1978); Susman v. Lincoln American Corp., 587 F.2d 866 (7th Cir. 1978); Satterwhite v. City of Greenville, 578 F.2d 987 (5th Cir. 1978) (en banc); Goodman v. Schlesinger, 584 F.2d 1325 (4th Cir. 1978). Even if a named plaintiff is before the trial court, no showing has been made that he desires to or will adequately represent the subclass.
The uncertainly about the viability of the subclass action on behalf of class members who purchased their cars after April 10, 1977, is significant. The notice to these subclass members informing them of the pendency of the action has been sent out. The subclass members, therefore, may rely on the federal class action to vindicate their interests. If it is later determined that the action cannot be maintained, the statutes of limitation may preclude individual lawsuits in the state courts.
The questions raised about the viability of the subclass action if the settlement of the other subclass action is executed illustrate the inadvisability of creating tentative subclasses for settlement purposes without careful examination of the adequacy of the representation of each subclass. Cf. Manual for Complex Litigation § 1.46 at 59-61 (condemning tentative classes for settlement purposes).
One final matter casts doubt upon the circumstances in which the settlement was negotiated by the Attorneys General. The settlement agreement contains GM's promise to compensate the Attorneys General $150,000 "for all the expenses they have incurred in connection with the subject matter of this Agreement." Allocation of the proceeds is left solely to the Attorneys General. The agreement also commits GM to pay private attorney's fees in the federal action "in an amount no greater than the amount of documented time actually expended . . . multiplied by the hourly fee prevailing . . . in the community." These amounts were in addition to the amounts promised class members accepting the settlement. The notice to subclass members informed them of even less than was provided by the agreement, and the record does not provide any reliable estimate of the aggregate amount of attorneys' fees and expenses that GM will eventually pay. We think the proposed settlement's estimate of attorneys' fees and expenses is so vague that subclass members could not determine the possible influence of attorneys' fees on the settlement in considering whether to object to it.
The notice to subclass members merely stated:
As part of the Agreement with the Attorneys General, General Motors agreed to pay an aggregate amount of $150,000, to be divided among those Attorneys General, including the Attorney General of Illinois, accepting the Agreement, in payment for expenses claimed to have been incurred in connection with the subject matter of their litigation. The amount of any attorneys' fees, costs or expenses to be paid to the attorneys for the private plaintiff purchasers in the class litigation will be subject to the review and approval by the Court. Any award of costs, expenses and/or fees to the private plaintiff purchasers and their counsel in the class litigation will be in addition to, and not deducted from, the $200.00 offered by General Motors per automobile purchased as part of the proposed settlement.
The record does indicate that GM and six of the nine teams of private attorneys have reached an understanding, if not agreement, about attorneys' fees. The understanding is that GM will not object to a request by those counsel for fees up to $360,000, but that private counsel are free to request that the court award a larger amount. Like the provision for expenses of the Attorneys General, this understanding leaves the allocation of the payment a matter for determination by the recipients of the payment. The agreement apparently contemplates that no requests for fees will be made until the end of all of the litigation, including that concerning the rights of post-April 10 purchasers.
See Manual for Complex Litigation § 1.46 at 54 (Consideration 7).
Aside from some doubt about whether Attorneys General who, of course, are compensated by the public may ever recover attorneys' fees and expenses, we believe that the method by which the GM-Attorneys General agreement contemplates payment of private attorney's fees and expenses is questionable. The Manual condemns settlement agreements which provide
The Manual regards the question of whether publicly employed counsel may be allowed reimbursement for expenses as an "interesting" and apparently open one. Id. § 1.44 at 42. It notes that expenses and attorneys' fees have been allowed to state Attorneys General in several class action settlements. See also In re Coordinated Pretrial Proceedings in Antibiotic Antitrust Actions, 410 F. Supp. 706 (D.Minn. 1975). On the other hand, several district courts have preferred state Attorneys General as counsel in class actions, in part because the Attorneys General presumably would not seek attorneys' fees. See State of Illinois v. Harper Row Publishers, Inc., 301 F. Supp. 484, 494-95 (N.D.Ill. 1969); State of Minnesota v. United States Steel Corp., 44 F.R.D. 559, 577 (D.Minn. 1968); cf. State of Ohio v. Richter Concrete Corp., 69 F.R.D. 604 (S.D.Ohio 1975) (permitting state Attorney General to communicate with putative class members after class certification was denied because salaried Attorney General, unlike private attorneys, had no interest in soliciting litigation or fees). We need not meet the question posed by the Manual. On this record it is not clear that the expenses of the Attorneys General to be reimbursed are those incurred in the litigation before the federal court. It is fair to assume that a large proportion of the expenses, if not all, are due to state court litigation.
that the fees and sometimes expenses of plaintiffs' counsel are to be paid separately by the defendant(s) over and above the settlement. Frequently, the amount thereof is not disclosed at the time the settlement is proposed. Such an arrangement should not be permitted. All amounts to be paid by the defendant(s) are properly part of the settlement funds and should be known and disclosed at the time the fairness of the settlement is considered.
The effect of such an arrangement is to neutralize the court's power and responsibility to pass upon the reasonableness of the amounts to be paid to plaintiffs' counsel since any reduction by the court in the amount counsel agree upon after the class settlement has been approved will simply go to reduce the aggregate amount defendant(s) will pay and will not increase the amount to be paid to the plaintiffs. As a result, there is little incentive for the judge to reduce the agreed upon fees. On the other hand, the effect of such an arrangement may be to cause counsel for the plaintiffs to be more interested in the amount to be paid as fees than in the amount to be paid to the plaintiffs. Only if the aggregate of all payments to be made by defendants is disclosed in the proposed settlement can the class members and the court make any intelligent judgment as to the fairness and reasonableness of a proposed settlement.
Manual for Complex Litigation § 1.46 at 62. This court has previously declined to upset a settlement agreement merely because some problems regarding fees and expenses remained unresolved. See McDonald v. Chicago Milwaukee Corp., 565 F.2d 416, 426 (7th Cir. 1977). We do not overrule that decision, but do regard the questionable provision made for expenses and attorneys' fees as one factor requiring examination of the settlement negotiations.
In conclusion, we hold that the trial court abused its discretion by failing to undertake a careful examination of the conduct of the settlement negotiations and by preventing the plaintiff-objectors from showing that the negotiations prejudiced the best interests of the class. Regardless of which of the two possible capacities the Illinois Attorney General's office assumed in negotiating the proposed settlement, the conduct of the negotiations was irregular and the record contains too much evidence tending to indicate prejudice to the class to permit us to allow the trial court's order to stand. Because, however, our decision upsets a settlement of considerable magnitude and because complex class actions are often, although not always, settled before trial, we conclude with a discussion of what we do not hold.
We do not hold that irregular settlement negotiations may never form the basis for a judicially acceptable class action settlement. In fact, a prior decision of this court has approved a settlement negotiated in somewhat similar circumstances. See McDonald v. Chicago Milwaukee Corp., 565 F.2d 416 (7th Cir. 1977). We realize that the system of state and federal courts often generates simultaneous litigation over the same subject matter. We recommend that an attorney who is counsel in both state and federal actions request leave of court before entering into settlement negotiations. In addition, the trial court should probably require as a condition to such leave at least that the attorney inform other counsel in the proceedings of the matters discussed during the separate negotiations. Although this practice is preferable, the failure to follow it is not necessarily reversible error if the record clearly indicates that representation of the class during the negotiations was adequate and that the settlement itself is fair.
In McDonald the objectors to a settlement contested, inter alia, the negotiations conducted in connection with a related state court action. The negotiations had begun prior to the commencement of the federal action which was filed only after the negotiations broke down, 565 F.2d at 420. Negotiations resumed prior to class certification, but largely because the trial court delayed certification of the class during the negotiations. Significantly, the trial court was never afforded an opportunity to pass on the issue of the propriety of the negotiations because the objector failed to raise the issue there.
In this case, a pretrial order expressly limited the conduct of settlement negotiations. The objectors raised the issue before the trial court by seeking discovery and by questioning the Assistant Illinois Attorney General during the fairness hearing. The trial court when given a chance to consider the conduct of the negotiations ruled that the matter was irrelevant. Finally, the record contains some evidence suggesting that the settlement negotiations prejudiced the class.
Although the trial court concluded that the settlement of the subclass action was fair, our discussion of the conduct of the settlement negotiations necessarily casts doubt upon that conclusion. Moreover, that matter aside, we are not convinced that the court's conclusion finds clear support in the record.
The most important factor relevant to the fairness of a class action settlement is the strength of plaintiff's case on the merits balanced against the amount offered in the settlement. Manual for Complex Litigation § 1.46 at 56. Conceptually, this requires a comparison of the amount offered with the product of (1) the probability of plaintiff's prevailing on the merits times (2) the present value of probable damages plaintiff would recover if he did prevail. We do not expect the trial court's conclusions to be set forth with mathematical precision. A fairness hearing is not a trial on the merits. The trial court, however, does have a duty to members of the class and to the reviewing court to assess, if not decide, the issues of law which weigh heavily in the above calculus and to consider the most probative evidence bearing on those issues.
The trial court's findings contain no express discussion of the merits of the Magnuson-Moss claim. Indeed, with respect to the alleged transmission switch in Delta 88s, the court apparently misapprehended the nature of the objectors' claims. The court noted that all Delta 88 coupes and sedans contained the THM 200 regardless of whether they had Chevrolet or Oldsmobile engines. The gist of objectors' claim, as we understand it, is that the transmissions used simply were not those warranted. Thus, the fact that all Delta 88 sedan and coupe purchasers received the smaller transmission is irrelevant. If objectors' contention is correct, GM breached its warranty to all Delta purchasers, not just those who received Chevrolet engines.
On the issue of compensatory damages, the trial court framed the issue as the "comparability" of the Oldsmobile engines allegedly warranted and those Chevrolet engines received. The findings then recite a mass of technical data indicating that the durability, performance and fuel economy of the Chevrolet and Oldsmobile engines were not materially different. The evidence on these technical issues was conflicting, but we are more concerned by the district court's failure to apply the ordinary measure of damages for breach of warranty: "the difference . . . between the value of the goods accepted and the value they would have had if they had been as warranted. . . ." U.C.C. § 2-714(2) (emphasis added). This is presumably the measure of damages contemplated by the drafters of the Magnuson-Moss Act. Yet, the court found it unnecessary to resolve an evidentiary conflict on the value of the engines. The objectors presented evidence tending to establish a difference in value of over $400. GM presented evidence that the cost of manufacture was virtually the same. Although neither form of evidence was the "best" evidence of value, this is a matter upon which the proponents of the settlement had the burden of proof. Manual for Complex Litigation § 1.46 at 56. The trial court should have made a more precise estimate of probable compensatory damages. Cf. id. at 61 ("in view of the complexity which ordinarily attends settlement issues, it is wise in most cases to rely upon proven facts, particularly economic facts").
Finally, we question the court's resolution of the possibility of recovering punitive damages against GM. The court declined to consider whether punitive damages are recoverable under the Magnuson-Moss Act because it found the evidence insufficient to permit an inference that GM acted in willful disregard of the rights of Oldsmobile purchasers. We think the objectors presented substantial evidence tending to show that GM deliberately concealed the source of the engines in the cars that it sold as Oldsmobiles and that it did so to increase profits. Moreover, we cannot say that the possible recovery of punitive damages should not have received any weight because they were unavailable under the Magnuson-Moss Act. Although one opinion published after the trial court's approval of the settlement intimates that the Act does not permit punitive damages, it does not resolve the issue. See Novosel v. Northway Motor Car Corp., 460 F. Supp. 541 (N.D.N.Y. 1978). In any event, that decision is binding on neither this court nor the district court. The Act itself provides "for damages and other legal and equitable relief." 15 U.S.C. § 2310(d)(1). Although this broad language falls short of express statutory authorization for an award of punitive damages, we do not believe as GM does that punitive damages are never recoverable under federal law unless expressly authorized. See Globus v. Law Research Service, Inc., 418 F.2d 1276, 1284 (2d Cir. 1969), cert. denied, 397 U.S. 913, 90 S.Ct. 913, 25 L.Ed.2d 93 (1970); Comment, Punitive Damages Under Federal Statutes: A Functional Analysis, 60 Calif.L.Rev. 191 (1972). Although the legislative history of the Act is silent on the matter, we think it is not unlikely that Congress intended to provide at least the same relief available under state law for breach of warranty. Although punitive damages are usually unavailable for actions sounding in contract, see U.C.C. § 1-106(1), McGrady v. Chrysler Motors Corp., 46 Ill.App.3d 136, 4 Ill.Dec. 705, 360 N.E.2d 818 (1977); Hibschman Pontiac, Inc. v. Batchelor, 340 N.E.2d 377 (Ind.App. 1976), this general rule is subject to exceptions. Punitive damages may be awarded, for example, when the breach amounts to an independent tort or is accompanied by fraudulent conduct. See Sullivan, Punitive Damages in the Law of Contract: The Reality and the Illusion of Legal Change, 61 Minn.L.Rev. 207 (1977); 3 Williston on Sales § 25-13 (4th ed. 1974); R. Nordstrom, Sales § 155 (1970).
We do not decide here that an award of punitive damages is appropriate under the Magnuson-Moss Act or that if it were that class members would be entitled to them. We do believe, however, that the possibility of such a recovery is not insubstantial and that this possibility as well as the probable compensatory damages were given insufficient weight by the trial court in the calculus of the fairness of the settlement.
Similarly, we do not hold that the failure of the trial court to hold a preliminary hearing prior to the mailing of the notice of the proposed settlement is inevitably reversible error. Although we believe such a hearing is better practice and the Manual for Complex Litigation recommends it, this court has gone as far as to affirm the approval of a settlement when no evidentiary hearing on its fairness was held before or after the notice to the class. See Patterson v. Stovall, 528 F.2d 108 (7th Cir. 1976). We do hold the record in this case raises so many questions about the adequacy of representation during the settlement negotiations that we cannot say the record clearly supports the trial court's conclusion that the negotiations did not prejudice the interests of the settlement subclass.
Thus, we do not hold that the representation of the class members during the negotiations was in fact inadequate. The record simply does not provide any basis for us to tell. We do note, however, that this is not the first class action in which the State of Illinois has negotiated a settlement without the participation of other counsel representing the class. See Liebman v. J. W. Petersen Coal Oil Co., 73 F.R.D. 531 (N.D.Ill. 1973).
We noted in McDonald v. Chicago Milwaukee Corp., 565 F.2d 416, 422 (7th Cir. 1977), that " Per se rules often represent the abdication of judicial discretion rather than its informed exercise." Consequently, this court has declined to adopt per se rules rigidly confining the trial court's exercise of its discretion in the supervision of class actions. This does not relieve us, however, of our duty to reverse the trial court's judgment when we are convinced that there has been a clear showing of an abuse of that discretion. On the facts of this case, the irregular conduct of the negotiations, the failure of the trial court to examine the irregularities thoroughly, and the evidence in the record indicating that the irregularities may have damaged the interests of the class convince us that such a clear showing has been made. The judgment of the trial court approving the settlement, accordingly, must be reversed.
V. Form of the Settlement
Even if we were not constrained to reverse the trial court's approval of the settlement because of the circumstances surrounding its negotiation, we would have to find the settlement defective in another respect. Although the defect may affect only a small portion of those to whom GM's offer would be extended, convenience and expediency cannot justify the disregard of the individual rights of even a fraction of the class. As an appellate court we are without power to rewrite the settlement of the parties. We only have the authority to approve or disapprove the settlement in the form it is presented to us.
Patterson v. Stovall, 528 F.2d 108, 111 (7th Cir. 1976).
The settlement order gives subclass members two options. If the subclass member signs a release he will receive the settlement package and his Magnuson-Moss claim will be dismissed. But even if the subclass member refuses to accept GM's offer and refuses to sign the release, the order nevertheless dismisses with prejudice the subclass member's federal claim. The subclass member is presented with an accept-or-else situation: if he does not accept, his federal claim is lost even though he cannot receive the benefits of the settlement package. We have searched the reported decisions in vain for precedent for such a settlement. Finding none and being of the opinion that the dismissal of the action is fundamentally unfair to nonconsenting subclass members, we cannot permit the settlement in its present form to stand.
The signed release, of course, operates to preclude the accepting subclass member from proceeding on any state claims he may have against GM.
The relevant paragraphs of the trial court's order provide:
4. The action on behalf of subclass members who accept and receive the settlement shall be and is hereby dismissed as to defendant General Motors with prejudice.
5. The action on behalf of subclass members who do not accept the settlement shall be and is hereby dismissed as to defendant General Motors. Dismissal as to those persons shall be without prejudice solely to their rights to pursue such other remedies as may be otherwise available to them.
GM argues that the form of the settlement is not unusual. It argues that nonconsenting class members are bound by a class settlement even if it is approved over their objections. Moreover, it argues, the very purpose of the 1966 amendments to Rule 23 was to eliminate the spurious class action in which potential class members could obtain the rewards of a favorable suit, but escape being bound by an unfavorable outcome. Thus, GM would have us hold that the dismissal of the Magnuson-Moss claims of nonconsenting subclass members is permissible. Finally, GM goes on to argue that "[t]he settlement does allow class members, even at this late stage, to reject it and pursue state law remedies. To the extent nonconsenting class members are allowed to pursue any future litigation rights by the settlement . . . it is more favorable to them than federal law or policy require." We do not disagree with GM's arguments in the abstract. In the context of the particular settlement here which attempts to settle both state and federal claims, however, we must disagree.
We consider GM's last argument first. A fundamental characteristic of the federal courts is their limited jurisdiction. In the same pretrial order in which the trial court certified the class, it also expressly declined to take pendent jurisdiction over the state claims presented by the pleadings. Therefore GM's contention that the settlement was more favorable than federal law requires presumably because the trial court could have forced subclass members to accept the settlement package in return for all state and federal claims is without merit. The trial court, having declined jurisdiction over the state claims, was without power to extinguish them. The form of settlement with its unusual use of individual releases was apparently agreed to by GM and the Attorneys General in recognition of the federal court's inability to settle the state claims of subclass members. The opt-out provision which permits nonconsenting subclass members to pursue state remedies is a necessary consequence of the limited jurisdiction of the federal courts.
The use of individual releases to effectuate a class action settlement, although unusual, is not unprecedented. See 3 H. Newberg, Class Actions § 5620p (1977).
We do not disagree with GM's statement that class members can be bound by a settlement over their objections and that the same is true of objecting named plaintiffs. Similarly, we agree that Rule 23 was amended to eliminate the spurious class action. We do not think that it follows, however, that the trial court has the power under Rule 23 to dismiss with prejudice the Magnuson-Moss claims of those subclass members who refuse to accept the settlement package. As to them, the "settlement" is not a settlement; it is merely an offer to settle with a penalty, the dismissal of their federal claims, if they do not accept. We decline to put every subclass member to such an unfair choice.
In a brief amicus curiae the Congressional sponsors of the Magnuson-Moss Act, Senator Warren G. Magnuson and Representative John E. Moss, also attack the form of the settlement approved by the trial court. The Congressional sponsors maintain that the class members' federal rights under the Act cannot be settled or compromised by a class representative without each class member's individual consent. They would have us hold that to the extent that Fed.R.Civ.P. 23(e) authorizes the settlement of class actions over the objections of some class members, it is inapplicable to class actions maintained under the Magnuson-Moss Act. Because we find that the form of settlement in the case at bar was not authorized by the Federal Rules, discussion of this argument is not strictly necessary to our decision. We discuss the issue raised, however, so as not to discourage settlement of the present action after its return to the district court.
The Federal Rules of Civil Procedure provide, with exceptions not important here, that they shall "govern the procedure in the United States district courts in all suits of a civil nature . . .." Fed.R.Civ.P. 1 (emphasis added). Although Congress unquestionably has the power to supersede any federal rule either in its entirety or in particular types of civil actions, we think that the proper rule of construction is that the Congressional intent to repeal a federal rule must be clearly expressed before the courts will find such a repeal. See United States v. Gustin-Bacon Division, Certainseed Products Corp., 426 F.2d 539, 542 (10th Cir.), cert. denied, 400 U.S. 832, 91 S.Ct. 63, 27 L.Ed.2d 63 (1970). We think neither the language of the Magnuson-Moss Act nor its legislative history clearly manifests Congress' intent to supersede Rule 23(e).
The Act itself refers to Rule 23 twice. In both cases, however, it merely provides that in class actions maintained in the federal courts, Rule 23 will govern whether the named plaintiff is a proper party to represent the class. 15 U.S.C. § 2310(a)(3), 2310(e). The explicit mention of the applicability of Rule 23 bolsters our conclusion that Rule 23(e) is applicable to class actions maintained under the Act. We do not find the negative pregnant that the Congressional sponsors find. Nor do the Act's provisions encouraging informal dispute resolution necessarily preclude the later settlement of a class action without individual consent by each class member. Indeed, it would be unreasonable to construe an act whose purpose is to encourage settlement to preclude settlement as a practical matter after a class action is commenced.
The legislative history of the Act also fails to evince a Congressional desire to prohibit class action settlements without the consent of every class member. That history instead suggests that Congress had precisely the opposite intention.
Generally speaking, with specific exceptions set forth in the bill, the procedures are to utilize Rule 23 of the Federal Rules of Civil Procedure. For instance, in negotiating the use of any complying informal dispute settlement procedure or any other settlement procedure the representative party would negotiate on behalf of the 100 named plaintiffs and any other class members.
120 Cong.Rec. 40712 (1974) (remarks of Sen. Moss). The legislative history does indicate some dissatisfaction with the Supreme Court's decision in Eisen v. Carlisle Jacquelin, 417 U.S. 156, 94 S.Ct. 2140, 40 L.Ed.2d 732 (1974), and perhaps indicates Congress' intention to make Rule 23(c)(2) inapplicable in some class actions maintained under the Magnuson-Moss Act. See H.R. Rep. No. 93-1107, 93d Cong., 2d Sess., reprinted in [1974] U.S. Code Cong. Ad.News 7702, 7724. No issue about the need for notice, however, has been raised in this appeal so we need not decide this question. We decide simply that the Magnuson-Moss Act does not alter the general rule that the trial court may approve a class action settlement without the consent of every member of the class.
This court on two occasions has noted that the essence of a settlement is a bilateral exchange. "The inherent nature of a compromise is to give up certain rights or benefits in return for others." McDonald v. Chicago Milwaukee Corp., 565 F.2d 416, 429 (7th Cir. 1977). "A settlement by its very nature is an agreement where both sides gain as well as lose something." Patterson v. Stovall, 528 F.2d 108, 115 (7th Cir. 1976). By the terms of the order of the trial court, subclass members who do not sign the release give up their Magnuson-Moss claims and the opportunity to be represented in the class action in return for nothing. The right to pursue state remedies is not a benefit, because, as discussed above, the class members possessed state causes of action against GM independently of the federal litigation and the federal court is without power to extinguish those state-created remedies. GM gains the dismissal of each subclass member's federal claim, but surrenders nothing in return.
The form of settlement in the case at bar is quite different than a settlement in which the defendant's liability is stipulated and class members must make claims against the settlement fund. In the latter case, the cause of action of a class member who fails to file a claim is extinguished by the settlement, and his right to a recovery is lost because he sleeps on his rights. In this case, the cause of action of a subclass member is extinguished and his right to a recovery is lost because he stands on his rights under state law.
The federal claims of individual class members cannot be extinguished with neither adequate consideration in return nor a hearing on the merits of their claims. The dismissal of nonconsenting subclass members' claims would serve solely to benefit GM or those subclass members who accept the settlement. Reconciling such a "settlement" with notions of fair play and justice is impossible. To permit the trial court to exercise its power to approve class action settlements in this manner would contravene the Rules Enabling Act, 28 U.S.C. § 2072, by abridging the substantive rights of those who did not accept the settlement offer.
Our objection to the form of settlement in this case is similar to the Second Circuit's objection to "fluid class recovery." See Eisen v. Carlisle Jacquelin, 479 F.2d 1005 (2d Cir. 1973), vacated and remanded on other grounds, 417 U.S. 156, 94 S.Ct. 2140, 40 L.Ed.2d 732 (1974); Van Gemert v. Boeing Co., 553 F.2d 812 (2d Cir. 1977). See also In re Hotel Charges, 500 F.2d 86 (9th Cir. 1974). In Eisen the Second Circuit's rejection of the use of fluid class recovery rested at least in part on the court's concern that that form of recovery would drastically increase the class action defendant's substantive liability. Cf. Beecher v. Able, 575 F.2d 1011, 1016 n. 3 (2d Cir. 1978) (defendant may agree to a settlement which provides for fluid class recovery). In this the converse situation, the form of settlement drastically reduces, in fact extinguishes, the subclass member's substantive cause of action under the Magnuson-Moss Act. We hold the trial court's approval of the form of settlement here was unauthorized by the Federal Rules and was inconsistent with the trial court's responsibility to act as the protector of the interests of absentee class members.
Although we note the similarity of our reasoning with that of the Eisen opinion, we express no opinion on whether the fluid class recovery technique itself is inconsistent with the Rules Enabling Act.
We cannot hold that the dismissal of the federal claims of those who refuse to accept the settlement offer was insignificant because it merely closed one of the two avenues of recovery against GM. Relegating the nonconsenting subclass member to his state remedies severely reduces his chances of obtaining an adequate recovery on his claim.
The nonconsenting subclass member loses the advantages and economies of having his interest represented in the class action. This tends to defeat the purpose of the class action device to vindicate the interests of the victims of mass production wrongs. "Generally, unless the anticipated recovery exceeds the sum of the measure of the injury and the cost of litigation, multiplied by the probability of a successful decision, the aggrieved person will not seek to vindicate his rights." Note, Judicial Prerequisites to Class Actions in Illinois: Policy, Practice, and the Need for Legislative Reform, 1976 U.Ill.L.F. 1159, 1167. The letters of those subclass members who objected to the settlement proposal indicate the illusory value of the right to pursue their claims individually:
I will go along with the majority. I can't afford to spend any money on a personal law suit.
* * * * * *
Reguardless [ sic] of the decision of the Court, I will accept it, because I cant [ sic] whip a giant like General Motors, but you do have the powers of your Judgeship and your Court to set things stright [ sic] as they should be.
This is not to be accepted as notice of withdrawal of Class or Subclass membership.
These letters also refute GM's argument that we can countenance the dismissal of the Magnuson-Moss claim of a nonconsenting subclass member because he was aware of the settlement's terms at the time he made his election to remain in or opt-out of the subclass. The opportunity to opt-out was not a very realistic one. Furthermore, we fail to see that a subclass member's knowledge that he may be treated unfairly excuses committing the injustice.
Even if the subclass member does pursue his state remedies, he is still prejudiced by the dismissal of his Magnuson-Moss claim. "From a consumer protection point of view, the Warranty Act is clearly preferable to the Uniform Commercial Code, which is difficult to apply to consumer sales transactions and is full of pitfalls for consumers seeking recovery for defective products." Smith, The Magnuson-Moss Warranty Act: Turning the Tables on Caveat Emptor, 13 Cal.W.L.Rev. 391, 429 (1977). In addition to providing a more certain path to recovery, the Magnuson-Moss Act provides the consumer with a more adequate remedy. It provides that the successful plaintiff may also recover the costs of litigation (subject to the court's discretion not to award attorneys' fees). 15 U.S.C. § 2310(d)(2). Thus, the dismissal of the subclass member's Magnuson-Moss claim, leaving him to pursue his state remedies individually, reduces both the probability that the consumer will pursue those remedies and, if he does, the probability that his remedy will be adequate.
The dismissal of the subclass members' claims pursuant to the unusual form of settlement here would also tend to undermine the purpose of the Magnuson-Moss Act. As to nonconsenting subclass members, the purpose of the Act to provide a more certain remedy than is provided under state law would be totally defeated. We think that the settlement provides a unique example of how class action settlements may tend to defeat, rather than promote, the policies and purposes of the laws sought to be enforced. See generally DuVal, The Class Action as an Antitrust Emorcement Device: The Chicago Experience ( Part II), 1976 A.B. Foundation Research J. 1273.
GM maintains that we should approve the settlement because it has the "overwhelming" support of the settlement subclass members. GM argues that because only fifteen subclass members or .03% of the subclass opted out of the action or objected to the settlement after notification of its terms, 99.97% of the subclass members support the settlement. Although the support of class members is one factor which should be considered in determining the fairness of a settlement, see Manual for Complex Litigation § 1.46 at 56, we are not as willing as GM to infer support from silence.
When a court evaluates the settlement of a class action brought on behalf of individual shareholders or consumers, it should be reluctant to rely heavily on the lack of opposition by alleged class members. Such parties typically do not have the time, money or knowledge to safeguard their interests by presenting evidence or advancing arguments objecting to the settlement.
Factors Considered in Determining the Fairness of a Settlement, 68 Nw.U.L.Rev. 1146, 1153 (1974). Accord, Developments in the Law — Class Actions, 89 Harv.L.Rev. 1318, 1567-68 (1976); cf. Simon, Class Actions — Useful Tool or Engine of Destruction, 55 F.R.D. 375, 377-79 (1973) (discussing the tendency of class members not to respond to court communications. Acquiescence to a bad deal is something quite different than affirmative support. In any event, even if a majority of the subclass did favor the settlement, we do not believe that the preferences of the majority can justify the substantial injustice to the individual rights of the minority that the form of settlement proposed here would work.
Because the bulk of the class consists of individual consumers, this case is unlike State of West Virginia v. Chas. Pfizer Co., 314 F. Supp. 710, 743 (S.D.N.Y. 1970), aff'd, 440 F.2d 1079 (2d Cir.), cert. denied, 404 U.S. 871, 92 S.Ct. 81, 30 L.Ed.2d 115 (1971), in which the court stated that support by class members was entitled to "great weight." Many of the class members in Pfizer were large public or private institutions with large stakes in the litigation. Thus, they could be expected to come forward to protect their interests. The Pfizer settlement, however, may not have been in the best interest of those individual consumers represented in the action. See In re Coordinated Pretrial Proceedings in Antibiotic Antitrust Actions, 410 F. Supp. 706 (D.Minn. 1975) (approving subsequent settlement offering consumers substantially higher payments). See generally Wolfram, The Antibiotics Class Actions, 1976 A.B. Foundation Research J. 251.
GM's brief indicates that only 26 individuals wrote to the trial court to express their approval of the settlement.
VI. Directions on Remand
In response to a question from the bench at oral argument, GM represented to the court that even if the settlement of the federal class action is not effectuated, GM may still seek to extend its offer to individual members of the class. Local Rule 22 appears to require the trial court's approval of any such communication. The question thus presented is whether the trial court can approve the communication of the offer, despite our reversal of the court's order approving the settlement.
Indeed, the agreement between GM and the Attorneys General may obligate GM to extend the offer. Paragraph 11 of the agreement provides:
while failure by [the district] court to allow General Motors to make such offer to such offerees shall relieve General Motors of the obligation under this Agreement to make such offer, failure by such court to approve settlement of such action . . . shall not relieve General Motors of such obligation if the court has nevertheless allowed General Motors to make such offer in exchange for a Release. . . .
Local Rule 22 of the Northern District of Illinois, captioned "For Prevention of Potential Abuse of Class Actions," provides:
In every potential and actual class action under Rule 23, FRCivP, all parties thereto and their counsel are hereby forbidden, directly or indirectly, orally or in writing, to communicate concerning such action with any potential or actual class member not a formal party to the action without the consent of and approval of the communication by order of the Court. Any such proposed communication shall be presented to the Court in writing with a designation of or description of all addressees and with a motion and proposed order for prior approval by the Court of the proposed communication and proposed addressees. The communications forbidden by this rule, include, but are not limited to, (a) solicitation directly or indirectly of legal representation of potential and actual class members who are not formal parties to the class action; (b) solicitation of fees and expenses and agreements to pay fees and expenses, from potential and actual class members who are not formal parties to the class action; (c) solicitation by formal parties to the class action of requests by class members to opt out in class actions under subparagraph (b)(3) of Rule 23, FRCivP; and (d) communications from counsel or a party which may tend to misrepresent the status, purposes and effects of the action, and of actual or potential Court orders therein, which may create impressions tending, without cause, to reflect adversely on any party, any counsel, the Court, or the administration of justice. The obligations and prohibitions of this rule are not exclusive. All other ethical, legal and equitable obligations are unaffected by this rule.
This rule does not forbid (1) communications between an attorney and his client or a prospective client, who has on the initiative of the client or prospective client consulted with, employed or proposed to employ the attorney, or (2) communications occurring in the regular course of business or in the performance of the duties of a public office or agency (such as the Attorney General) which do not have the effect of soliciting representation by counsel, or misrepresenting the status, purposes or effect of the action and orders therein.
The rule was adopted in accordance with the Manual's recommendation for preventing unauthorized communications with class members, see Manual for Complex Litigation § 1.41, and follows almost verbatim the local rule contained in the Manual's appendix. See id., Appendix § 1.41. (Suggested Rule No. 7). See also Dole, The Settlement of Class Actions for Damages, 71 Colum.L.Rev. 971, 993-97 (1971).
Questions concerning the district court's authority to promulgate the rule pursuant to Fed.R.Civ.P. 83 have not been raised by the parties and we do not consider them here. See generally Manual for Complex Litigation § 1.41 (4th ed. 1977 Cum.Supp. 1978). In any event, Rule 23(d), see Weight Watchers, Inc. v. Weight Watchers International, Inc., 455 F.2d 770, 775 (2d Cir. 1972), and the court's inherent power to control the conduct of the litigation before it, see Vernon J. Rockler Co. v. Minneapolis Shareholders Co., 425 F. Supp. 145, 150 (D.Minn. 1977), provide additional sources for the district court's power to control this particular communication with class members.
We think that the trial court can. GM's offer to settle, if accepted by individual class members, would not amount to a settlement of the class action itself. Individual class members would be free to reject it and continue to have their interests represented in the federal class action. Thus, the communication falls outside the language and the purpose of Rule 23(e). See Weight Watchers, Inc. v. Weight Watchers International, Inc., 455 F.2d 770 (2d Cir. 1972); Rodgers v. United States Steel Corp., 70 F.R.D. 639 (W.D.Pa.), appeal dismissed, 541 F.2d 365 (3d Cir. 1976); Dickerson v. United States Steel Corp., 11 Empl. Prac. Dec. ¶ 10,848 (E.D.Pa. 1976); Vernon J. Rockler Co. v. Minneapolis Shareholders Co., 425 F. Supp. 145 (D.Minn. 1977); 7A C. Wright A. Miller, Federal Practice and Procedure § 1797 at 238-39 (1972). But see In re International House of Pancakes Franchise Litigation 1972 Trade Cas. ¶ 73,864 (W.D.Mo. 1972); Developments in the Law — Class Actions, 89 Harv.L.Rev. 1318, 1548 n. 66 (1976). Rule 23(e) requires judicial approval of class action settlements to guard against possible ineffective representation of absentees' interests by the representative parties. This danger does not inhere in offers to settle with individual class members, which the class members are free to accept or reject. Accordingly, a proposed offer to settle with individual class members requires a lesser degree of judicial scrutiny than a proposed settlement of a class action.
This case does not present the question, and we need not decide, whether Rule 23(e) would be applicable if so many class members accepted GM's offer that the class action could no longer be prosecuted as a class action. Compare American Finance System Inc. v. Harlow, 65 F.R.D. 572, 576-77 (D.Md. 1974), with Vernon J. Rockler Co. v. Minneapolis Shareholders Co., 425 F. Supp. 145, 150 (D.Minn. 1977).
Predicting the number of class members who might accept GM's offer at this time is admittedly speculative, but even if enough named plaintiffs accept the offer to reduce the number of named plaintiffs below the jurisdictional prerequisite, see 15 U.S.C. § 2310(d)(3) note 2 supra, the trial court's jurisdiction to decide the class action would remain unaffected. The general rule is that the jurisdiction of the federal court is determined at the time of the filing of the complaint. See Mullen v. Torrance, 22 U.S. (9 Wheat.) 537, 539, 6 L.Ed. 154 (1824) (diversity not defeated when party subsequently becomes a citizen of the same state as his opponent. "It is quite clear, that jurisdiction of the court depends upon the state of things at the time of the action brought, and that after vesting, it cannot be ousted by subsequent events."); St. Paul Mercury Indemnity Co. v. Red Cab Co., 303 U.S. 283, 58 S.Ct. 586, 82 L.Ed. 845 (1938) (court is not ousted of jurisdiction if plaintiff reduces claim to less than jurisdictional amount subsequent to removal from state court); cf. Rosado v. Wyman, 397 U.S. 397, 402-05, 90 S.Ct. 1207, 25 L.Ed. 442 (1970) (federal court may decide pendent claim even after claim which provided the basis for jurisdiction becomes moot). We see no reason why the general rule should be changed under the Magnuson-Moss Act, particularly when Congress intended that section 110(d) be "construed reasonably to authorize the maintenance of a class action." H.R. Rep. No. 93-1107, 93d Cong., 2d Sess., reprinted in [1974] U.S. Code Cong. Ad.News 7702, 7724. The class action can in no sense be regarded as "trivial or insignificant" merely because some of the named plaintiffs have accepted the benefits which the class action has brought forth. Thus, a reduction in the number of named plaintiffs would not preclude the trial court from proceeding to the merits of the class' Magnuson-Moss claims.
Similarly, even if nearly all the offerees accepted GM's settlement offer — a rather unlikely possibility since the offerees number approximately 70,000 — those who rejected the offer would not be denied the benefit of class adjudication of their claims in federal court. Their claims could be adjudicated along with those of the 66,000 post-April 10, 1977, class members to whom GM will not extend the offer. Therefore, the class will not be decertified for lack of the numerosity required by Fed.R.Civ.P. 23(a)(1). See Rodgers v. United States Steel Corp., 541 F.2d 365, 370 n. 11 (3d Cir. 1976).
The court in Weight Watchers expressly reserved the precise question that we decide here. See 455 F.2d 773 n. 1. In Weight Watchers the appellant sought review of an order of the trial court permitting communication between the defendant and individual putative class members. Unlike the present case, the pending action had not yet been certified to proceed as a class action. Although the Second Circuit dismissed the appeal from the order for want of appellate jurisdiction, accord, Rodgers v. United States Steel Corp., 541 F.2d 365 (3d Cir. 1976), its reasoning is plainly applicable to the present case: "[W]e are unable to perceive any legal theory that would endow a plaintiff . . . with a right to prevent negotiation of settlements between the defendant and other potential members of the class who are of a mind to do this; it is only the settlement of the class action itself without court approval that F.R.Civ.P. 23(e) prohibits." 455 F.2d at 773.
The Manual for Complex Litigation provides no standards for judicial approval of communications with individual class members, but we think that the degree of judicial review should be concomitant with the potential for abuse that such communications create. The danger that the offer to settle individual claims would create is the possible misleading of class members about the strength and extent of their claims and the alternatives for obtaining satisfaction of those claims. Thus, an offer to settle should contain sufficient information to enable a class member to determine (1) whether to accept the offer to settle, (2) the effects of settling, and (3) the available avenues for pursuing his claim if he does not settle. In contrast to judicial examination of a proposed class action settlement which entails consideration of the fairness of the settlement itself, judicial examination of the offer to settle individual claims largely entails only consideration of the accuracy and completeness of the disclosure. See, e.g., Vernon J. Rockler Co. v. Minneapolis Shareholders Co., 425 F. Supp. 145 (D.Minn. 1977) (tender offer which met with preliminary approval of SEC contained sufficient information to allow shareholders-potential class members to make an informed and intelligent decision); American Finance System, Inc. v. Harlow, 65 F.R.D. 572, 576 (D.Md. 1974) (permitting the defendant to sent only "a neutrally worded notice of settlement containing no more than the terms of the proposed compromise, the position of both parties and a copy" of the court's order). Whether the offer to settle should contain a statement by the plaintiff-objectors of their opinion of the adequacy of the settlement package in order to make the communication a full and complete disclosure is a matter left to the trial court's discretion. We do believe, however, that the trial court should insist that the notice state that the court's permission to communicate the offer does not indicate any opinion or finding by the trial court that the settlement package is fair or adequate consideration for the release of a subclass member's claim. See American Finance System Inc. v. Harlow, 65 F.R.D. at 576 n. 5.
This is not to say that the amount of the proposed consideration for the settlement is entirely irrelevant. An offer to settle which offers only nominal consideration in return may amount to little more than a request that the class members opt-out of the class. See Manual for Complex Litigation § 1.41 at 27 (condemning unauthorized solicitations to opt-out). Solicitations to opt-out tend to reduce the effectiveness of (b)(3) class actions for no legitimate reason. Offers to settle, however, both provide redress to individual class members and reduce the burden on the courts of trying massive class suits. Determining the difference between the two kinds of communications necessarily requires some judicial examination of the amount of consideration offered by the defendant. Moreover, the amount offered may be so unrealistically low that the consideration itself tends to mislead class members about the strength and extent of their claims. Thus the trial court should examine the amount tendered in settlement before approving the offer to settle. Yet, because each class member may judge for himself whether the amount offered is acceptable, the court need not determine that the amount is "fair, reasonable and adequate." The court need only find that the proposed exchange provides each individual class member with a meaningful opportunity to obtain satisfaction of his claim. See Rodgers v. United States Steel Corp., 70 F.R.D. 639, 644 (W.D.Pa.), appeal dismissed, 541 F.2d 365 (3d Cir. 1976).
See also Chrapliwy v. Uniroyal, Inc., 71 F.R.D. 461, 464 (N.D.Ind. 1976) ("although the class action itself may not be voluntarily dismissed without Court approval and scrutiny, an individual claim in a 23(b)(3) action may be settled and dismissed at the class member's own initiative. . . . Because the ability to settle an individual class member's claim could be misused, the Court must be careful to exercise control over the communications of all parties to the suit so that undue influence is prevented"); Dole, The Settlement of Class Actions for Damages, 71 Colum.L.Rev. 971, 995-97 (1971); Developments in the Law — Class Actions, 89 Harv.L.Rev. 1318, 1549-50, 1601-04 (1976).
We do not intend to recommend individual settlements as preferable to a fair settlement of the action for all class members. Given the present posture of this litigation, however, we recommend that the district court consider the advisability of permitting the communication if GM decides to extend its offer to individual members of the class. This procedure would provide those class members who wish to settle the benefit of the settlement package already negotiated, minimize further litigation and discovery on issues collateral to the merits of the Magnuson-Moss claim, and permit those who desire to prosecute their claims to do so. Our discussion here is not intended to resolve all questions raised by GM's offer, these matters are best left to the district court for determination in the first instance.
Specifically, because a class action defendant may communicate an offer to settle individual claims without the agreement or consent of the named plaintiffs or their counsel, the court need not permit discovery into the conduct of the settlement negotiations before approving the communication of the offer.
In particular, we leave to the district court the difficult question of the entitlement of the class counsel to attorneys' fees for their part in encouraging GM to extend the offer. If the district court decides attorneys' fees are appropriate, it must then grapple with the even more difficult questions of the allocation of fees among the attorneys and the allocation of the burden of the fees between GM and the class or among class members themselves. See generally Dole, The Settlement of Class Actions for Damages, 71 Colum.L.Rev. 971, 997-1000 (1971); Developments in the Law — Class Actions, 89 Harv.L.Rev. 1318, 1547 n. 59 (1976).
VII. Conclusion
Our reversal of the district court's approval of the proposed settlement is a decision that we reach with considerable reluctance. We do not seek to discourage a full settlement of this litigation. More than a year has passed since the Illinois Attorney General presented the settlement agreement to the district court for its consideration. Most likely little has been done since then, aside from some additional discovery, to advance toward a trial on the merits. In the meantime, members of the settlement subclass must be wondering whatever became of the $200 and the mechanical insurance policy each had been promised. Our reluctance to unscramble on review what has been accomplished in the trial court, however, must yield when what has been done not only creates a substantial doubt about whether the interests of the class were adequately represented during the settlement negotiations, but also unjustifiably prejudices the rights of individual members of the class. We believe that approval of what has been done here would establish a precedent inconsistent with the proper functioning of the class action device.
We do not question in the least the good faith of the group of state Attorneys General who negotiated the settlement. We are well aware of the increasingly important role that state Attorneys General have taken in protecting consumers' rights. We are also acutely aware of the difficulties which confront litigants attempting to settle consumer class actions based on the Magnuson-Moss Act. The Act by adopting in substantial part, but not preempting state law remedies provides a legal environment conducive to competing state and federal court actions. The myriad lawsuits make settlement desirable, but simultaneously make achieving an acceptable settlement extraordinarily difficult for all concerned. We hold merely that the method of reaching a settlement that GM and the Attorneys General chose warranted greater scrutiny than the trial court permitted and that the form of effecting the settlement permitted by the trial court was unauthorized. Accordingly, the order of the district court is
See, e. g., Mooney, The Attorney General as Counsel for the Consumer: The Oregon Experience, 54 Ore.L.Rev. 117 (1975); Tongren Samuels, The Development of Consumer Protection Activities in the Ohio Attorney General's Office, 37 Ohio St.L.J. 581 (1976); Note, The Role of the Michigan Attorney General in Consumer and Environmental Protection, 72 Mich.L.Rev. 1030 (1974); Note, Consumer Protection by the State Attorneys General: A Time for Renewal, 49 Notre Dame Law. 410 (1973). See also 15 U.S.C. § 15a-15h.
Reversed.