MINUTES

CIVIL RULES ADVISORY COMMITTEE

APRIL 18 and 19, 1996

The Civil Rules Advisory Committee met on April 18 and 19, 1996, at the Administrative Office of the United States Courts in Washington, D.C. The meeting was attended by all members of the Committee: Judge Patrick E. Higginbotham, Chair, and Judge John L. Carroll, Judge David S. Doty, Justice Christine M. Durham, Francis H. Fox, Esq., Assistant Attorney General Frank W. Hunger, Mark O. Kasanin, Esq., Judge David F. Levi, Judge Paul V. Niemeyer, Carol J. Hansen Posegate, Esq., Professor Thomas D. Rowe, Jr., Judge Anthony J. Scirica, Judge C. Roger Vinson, and Phillip A. Wittmann, Esq. Edward H. Cooper was present as reporter. Former member John P. Frank, Esq., also attended. Judge Alicemarie H. Stotler attended as Chair of the Committee on Rules of Practice and Procedure; Professor Daniel Coquillette attended as Reporter, And Sol Schreiber, Esq., attended as liaison member, of that Committee. Judge Jane A. Restani attended as liaison representative from the Bankruptcy Rules Advisory Committee. John K. Rabiej and Mark D. Shapiro represented the Rules Committee Support Office, and Karen Kremer of the Administrative Office of the United States Courts also attended. Thomas E. Willging represented the Federal Judicial Center. Other observers and participants are named in the appendix.

Judge Higginbotham welcomed the members of the Committee, other participants, and observers.

The Minutes of the November, 1995 meeting were approved.

RULES PUBLISHED FOR COMMENT IN 1995

Amendments of four rules were published for comment in 1995. Rules 9(h), 26(c), 47(a), and 48 drew substantial written comments. Hearings were held in Oakland, California; Atlanta, Georgia; and New Orleans, Louisiana. All members of the Committee had the complete written comments and transcripts of the hearings. Summaries of the written comments and the hearing testimony also were provided. Action on these proposals came first on the Committee agenda.

Rule 9(h)

The proposal to amend Rule 9(h) would remove an ambiguity in the present rule provision relating to interlocutory appeals in admiralty. It is not clear whether appeal can be taken under § 1292(a)(3) when, in a case that includes both an admiralty claim and a nonadmiralty claim, the court acts on a nonadmiralty claim by an order that would qualify for § 1292(a)(3) appeal if it had involved an admiralty claim. The proposal resolves the ambiguity by permitting appeal. Public comment was sparse, but was approving. The Committee voted unanimously to recommend that the Standing Committee recommend adoption of the amendment to the Judicial Conference.

Rule 26(c)

The proposal to amend Rule 26(c) has been discussed extensively by the Committee. The proposal that was published in 1995 was discussed extensively at the October, 1994 and April, 1995 meetings. The proposal drew substantial written comment and testimony.

Discussion began by observing that the most frequently expressed concern was that the proposal expressly recognizes the common practice of entering discovery protective orders on stipulation of the parties. This reference to stipulated orders rested on the Committee's belief that in creating explicit procedures to modify or dissolve protective orders, existing stipulation practice should be confirmed. In March, 1995, the Judicial Conference asked the Committee to reconsider the proposal. One basis for its concern was that the proposal submitted to the Judicial Conference had been modified from the proposal that was first published. The Committee responded by recommending publication of the same proposal for a new round of public comment. Publication in fact prompted extensive comment that repeated concerns that had become familiar from earlier public comments and Committee deliberations.

The new round of public comment and testimony also focused substantial attention on the reliance factor that was listed in both the first and second published proposals as one element in the determination whether to modify or dissolve a protective order. The fear expressed is that this factor will make it too difficult to get relief. The thread of the comments seems to reflect a desire to require a judge-made finding of good cause before a protective order can be entered, and at the same time to make it easier to modify an order. This combination of desires does not seem likely to be realized in the real world; once a judge has made an express determination of good cause, it is likely to be more difficult to persuade the judge to modify the order.

Exploration of Rule 26(c) was initially prompted by Congressional concern that protective orders may be thwarting access to information that is important to protect the public health and safety. Throughout consideration of the gradually developed proposal, several members of the Committee have been skeptical of the need for any action. This history may help in choosing among the present alternatives: (1) change the proposal still further, perhaps so extensively that another round of public comment should be requested; (2) reject the proposal; (3) send the proposal forward with a recommendation for adoption; or (4) continue to study the proposal in a broader framework that includes study of the Rule 26(b)(1) scope of discovery.

The first observation expressed a lack of enthusiasm for going forward with the proposal. This subject has been studied extensively, and it is not clear that the proposal is any better than present practice, that it will improve anything. The inquiry began in response to a desire to integrate the Enabling Act rulemaking process with Congressional study. If our conclusion is that there is no real need to act, perhaps it is better to hold the topic for continuing study as part of a broader review of discovery. This view was repeated later, with the observation that there are not many problems in actual practice. The proposal may upset general procedure that now works perfectly well by stipulation, creating a whole series of hearings that are not held now. Other members of the Committee agreed that they simply do not encounter problems in practice.

Kenneth Sherk, representing the Federal Rules of Civil Procedure Committee of the American College of Trial Lawyers, noted that they had concluded that the proposal is innocuous so long as stipulation practice is clearly protected. They could easily agree, however, that there is no need to make any change.

The responding view was that the changes are good and should be sent forward. The decision of the Sixth Circuit in the recent Proctor & Gamble litigation with Business Week may show skepticism about stipulated consent orders that could cause difficulty in the future. But the language relating to stipulated orders should be revised to require that the stipulation show good cause, or that an evidentiary showing be made: "for good cause shown by motion, by stipulation of the parties, or by evidentiary showing."

The need for language referring to an evidentiary showing was questioned. If there is a hearing, the opportunity to advance evidence is clear. The requirement that there be a motion if there is no stipulation carries a hearing opportunity with it. And the Sixth Circuit concerns were thought to arise from the fact that the parties had, by consent, sought to seal pleadings and other materials filed with the court.

Other advantages were urged in support of going ahead with the proposal. Rule 26(c) now seems to require a showing of good cause. Stipulated orders are common, however, and can be beneficial. The stipulation practice should be confirmed by the rule. And the explicit provisions for modification or dissolution clarify many lingering doubts that beset present practice. The factors listed in subdivision (c)(3)(B) also make it clear that if a protective order is entered by stipulation, the court must consider the need for protection de novo when a motion is made to modify or dissolve the order. This change too is good.

Discussion then returned to themes that were sounded at earlier meetings. Protective orders are an integral part of the arrangement that makes tolerable the sweeping scope of discovery allowed by Rule 26(b)(1). Discovery sweeps in much information that otherwise is protected against any public inquiry, and sweeps it in merely on showing that it is relevant to the subject-matter involved in the pending action. There is no need to show that it would be admissible in evidence, so long as it appears reasonably calculated to lead to the discovery of admissible evidence. The proposal that the Committee reconsider this scope of discovery will provide occasion for further consideration of protective orders. The scope of discovery has been approached in the past, but no changes have been recommended. In 1970, the requirement of good cause was dropped from the document-production provisions of Rule 34. The concern with stipulated protective orders today seems to focus in large part on documents produced in discovery, and historically there has been an interaction between the working of document production and protective orders. Perhaps further consideration of protective orders should be integrated with a broader study of the scope of discovery. If the scope of discovery is to be narrowed, it may be important to reappraise the role of protective orders in relation to narrower discovery.

A motion was made to hold the Rule 26(c) proposal for further study in conjunction with study of the broader scope of discovery. Discussion suggested that it should be made clear that the Committee is not backing away from the proposal, which expresses good practice. An additional reason for going slow is that the RAND report on local practice under the Civil Justice Reform Act will be available soon, and likely will bring additional Rule 26 topics to the Committee agenda. The ABA plans a major program on the RAND study early next year.

It was asked whether deferral would require yet another round of publication and public comment if the Committee should decide in the future to recommend adoption of the current Rule 26(c) proposal. It was recognized that at some point, continued delay might give rise to a need for new comment in relation to whatever developments might occur in actual practice. No clear time line was identified for this possibility.

The motion to join further consideration of Rule 26(c) to study of the general scope of discovery provided by Rule 26(b)(1), and the related question whether document discovery should be governed by standards different than those that govern other discovery methods, was adopted by unanimous vote.

Rule 47(a)

The Rule 47(a) proposal published in 1995 would establish a right for lawyers to participate in voir dire examination of prospective jurors, subject to reasonable limits set by the court in its discretion. This proposal drew extensive comment. Almost all of the many federal judges who commented on the proposal spoke in opposition. Comments from the bar were not as nearly unanimous, but the very large majority of bar comments supported the proposal.

Discussion opened with the observation that in an ideal world, virtually all federal judges would allow lawyer participation in voir dire under present Rule 47(a). The common theme of most comments by federal judges is the fear that they will lose control if they lose the unlimited right to deny any lawyer participation in voir dire. There also is a hint of the "random selection" philosophy that there is no real value in jury selection, that any group of six or more jurors will do as well as any other, although this view is seldom made explicit. Many of the adverse comments reflect direct experience with state systems in which the right of lawyer participation has run riot.

As compared to judicial comments, many lawyers say that selection practices are inadequate in many courts. Judges do not adequately understand the case, and fail to appreciate the importance of direct lawyer questioning to supplement initial questioning by the judge. Written questions submitted to the judge simply to not provide sufficient opportunity to follow up answers with further questions. The lawyers recognize that they will not be allowed an open field with the jury.

These competing visions of reality make it difficult to write a rule.

Most federal judges now do what the draft would have them do. They can share their experience with other judges, encouraging them to test the waters. The Federal Judicial Center can be encouraged -- and indeed seems receptive -- to put voir dire on its educational agenda for new judges and for judge workshops. The workshops may be vital. Simply bringing judges together with small numbers of respected local attorneys for frank discussion can prove highly productive.

The comments from the bench and bar before and during the comment period have proved most useful. They can set the stage for new educational efforts and improved communication on these issues. In addition, they identified the potential problems that arise from the use of questionnaires to supplement oral voir dire. Questionnaires can be quite useful. But they also can become quite extensive, seeking information for a psychological profile to be used by "jury consultants." This is cause for concern.

Discussion turned to the most effective means of encouraging education of both bench and bar. The first step should be an information report to the Standing Committee, for the Judicial Conference, describing the problems that have been reported to the Committee. Significant problems with jury selection have been clearly identified by comments from the bar, and the conclusion that the best present solution may not involve amendment of Rule 47(a) does not justify complete inaction. The Committee should encourage informal meetings between groups of judges and respected local lawyers for frank discussion of the problems. The Committee also should consider whether there is some other means of spreading the information gathered during the public comment period. There may be some room for systematic experimentation to test the information provided by the Federal Judicial Center survey of federal judges.

Concern was expressed that Rule 47(a) was published for comment in tandem with identical proposed changes in Criminal Rule 24(a). The Criminal Rules Advisory Committee had not yet met to discuss the public comments -- most of which were addressed alike to both rules -- and might reach a different conclusion as to the wisdom of pursuing rules amendments now. The need for lawyer participation in voir dire examination may seem even stronger in criminal prosecutions, particularly in capital cases. The Committee anticipated, however, that the Criminal Rules Committee also would conclude that education is the better part of immediate reform efforts.

The Committee concluded unanimously that it should continue to study Rule 47(a), while encouraging the Federal Judicial Center to go ahead with educational efforts and also encouraging further study of jury questionnaires.

Rule 48

The 1995 proposal would amend Rule 48 to require that all civil juries begin trial with 12 members, absent agreement by the parties on a smaller number. As under present practice, there would be no provision for alternates, and the unanimity requirement would remain unchanged. This proposal drew substantial public comment. Much of the comment approved the proposal. No part of the comment suggested that 12-person juries are intrinsically inferior to the 6- or 8-person juries commonly used in civil actions today. Concerns were expressed about cost and delay, however, focusing on the need to assemble larger panels, select and pay more jurors, and meet the problem arising from the fact that some magistrate-judge courtrooms have jury boxes too small to accommodate 12-person juries. Some concern also was expressed with the prospect that failure to agree on a verdict might be more common with 12-person juries than with smaller juries.

Discussion began with reflections on the great divergences among estimates of the marginal costs associated with moving to 12-person juries, and on the equally great uncertainties of all the estimates. None of the plausible estimates, however, seem to threaten undue additional cost. The problem of inadequate jury boxes can be addressed in various ways, including scheduling magistrate-judge civil trials in district court rooms that are equipped for 12-person juries; when that is not possible, the parties will need to add the need for agreement on a smaller jury to the factors that influence the decision whether to consent to magistrate-judge trial. The available data, including most persuasively the comparison between 6-person civil juries and 12-person criminal juries, indicate that there is no measurable difference in the failure-to-agree rate.

In voicing tentative support for the proposal, one reservation was noted arising from trials in sparsely populated rural areas. It may prove difficult to assemble sufficiently large jury panels to ensure 12-person juries, particularly in cases that involve frequent acquaintances between potential jurors and the parties or people related to the parties. If 30 jurors are summoned, it is never certain how many will appear. In at least some of these cases -- especially civil actions brought by prison inmates -- the parties are not likely to stipulate to smaller juries. Perhaps, as with Rule 47(a), it would be better simply to encourage the use of 12-person juries.

Reservations were expressed on the basis of the public comments. Some suggest that the comments do not reflect a groundswell of support for the change: Eight-person juries have become common in civil trials of any expected length because of the abolition of alternate jurors, and 12-person juries are common in complex cases because of the fear that jurors will be lost as the trial extends to several days or even weeks. And we may have underestimated the costs, including the burdens imposed on the jurors themselves, their employers, and others. So long as we have a unanimity requirement, defendants will always prefer 12-person juries, and will not stipulate to smaller juries simply to have an earlier trial before a magistrate judge.

The magistrate-judge concern was met by reference to data showing that in the most recent year available, the average was 1.1 civil jury trials per magistrate judge. Many magistrate judges never try jury cases. Most jury trials before magistrate judges occur only in specific parts of the country. Concerns about prisoner litigation should not control a matter of such general importance. There is also some hope that the prisoner litigation problem will be eased by proposals pending in Congress.

It was observed that the entire cost of the jury system, including both civil and criminal cases, is less than the cost of one manned bomber. It is not so much as a blip on the screen of the national budget, and is a tiny fraction even of the budget for the judiciary.

Six-person juries have been used only since Chief Justice Burger, by extra-curial comment, effectively directed their use as a cost-saving measure, and perhaps also with some sense of hostility to jury trial. "Six is half-way to zero." To say that people are comfortable with the system is not comforting; those who have experience with 12-person juries in civil cases often are less sanguine about smaller juries than those whose experience has been only with smaller juries.

Unanimity is a false issue. In criminal cases, studies show that the unanimity requirement affects the dynamics of deliberation, but not the rate of hung juries. Hung juries are very rare, both in civil and criminal trials.

It is incontestable that 12-person juries more than double the probability that a particular jury will include representatives of various minority groups. The increase in representativeness is almost exponential. Many lawyers have commented as well that it is easier for a single forceful person to dominate a smaller jury, lending anecdotal support to the regular findings of psychologists and sociologists. The dynamics of jury deliberations are different in larger juries. The jury studies that lent support to the initial proposal remain convincing. The actual experience of a 12-member jury trial is more reassuring. Putting aside any mystical qualities, the 12-person jury developed and was adhered to for centuries, distilling the wisdom of vast experience. "Carpentry costs" should not stand in the way.

The question whether bankruptcy-judge and magistrate-judge trials should be exempted from a 12-member jury requirement was discussed briefly. It was concluded that it is better to encourage scheduling in 12-person jury courtrooms, so as not to complicate the choice between district-judge and other-judge trials. Consent to smaller juries can resolve such scheduling difficulties as remain.

The motion to recommend that the Standing Committee recommend adoption of the proposal to provide for 12-person juries in Rule 48 was approved by vote of 11 for, 2 against.

Rule Not Yet Published

Rule 23

Discussion of Rule 23 began with an invitation to consider the draft by asking what can be achieved by (b)(3) class actions that cannot be achieved by consolidation and other tools. The 1966 version of Rule 23 came into being as the Advisory Committee worked through concerns about civil rights injunction class actions. What would the world look like if (b)(3) were abrogated? Is (b)(3) desirable for single event disasters, such as airplane crashes? What of the securities field, where private enforcement often takes the form of a (b)(3) class action? And what of other fields of litigation that amass large numbers of small claims into a (b)(3) class?

One of the changes that emerged from the November, 1995 meeting was an addition to (b)(3) of a required finding that a class action be "necessary" for the fair and efficient adjudication of the controversy. The purpose was to serve a heuristic function by encouraging courts to look beyond "efficiency," to emphasize the fairness of trying individual traditional cases in traditional ways. The combination of "necessary" with "superior" is awkward, however, seeming to require denial of certification for want of necessity, even though a class action might seem superior. In informational discussion with the Standing Committee in January, 1996, moreover, some concern was expressed about the tangled history of "necessary" parties in Rule 19. The present draft suggests elimination of "necessary" from the required (b)(3) findings, and substitution of a new subparagraph (A) that requires consideration of the need for certification as one factor bearing on the findings of predominance and superiority.

Another of the November changes led to alternative provisions requiring consideration of the probable outcome on the merits as part of the required (b)(3) findings. Increasing concerns have been expressed about the impact of this requirement. One concern arises from the prospect that a prediction of the merits must be supported by extensive discovery, protracting the certification determination and adding great expense. Another concern arises from the effects of the finding; however tentatively and subordinately it may be expressed, the prediction of the merits may affect all future proceedings in the case and may have real-world consequences as well. Impact on market evaluation of a company's stock was one frequently offered illustration. Various responses are suggested by the new drafts -- to require a finding of probable merit only if requested by a party opposing class certification; to eliminate the requirement that there be a finding, but to leave the probable outcome on the merits as one of the factors bearing on predominance and superiority; to consider probable outcome on the merits only as part of an evaluation of the value of "probable class relief"; or to adhere to present practice that, at least nominally, prohibits consideration of the merits in determining whether to certify a class.

The November changes also included in the (b)(3) factors consideration whether the public interest and private benefits of probable relief to individual class members justify the burdens of the litigation. Class actions have become an important element of private attorney-general enforcement of many statutes. In considering the problem of class actions that yield little benefit to class members, the problem is cynicism about the process that generates such remedies as "coupons" that may provide more benefit to the defendants and class lawyers than to class members. Yet there may be indirect benefits to the public at large in deterring wrongdoing, and in some cases it may be desirable to force disgorgement of wrongful profits without regard to individual benefits. The question is in part whether it is wise to rely on private enforcement through Rule 23 rather than specific Congressionally mandated private enforcement devices -- and whether the question is different as to statutes enacted before Rule 23 enforcement had become well recognized than as to more recent statutes.

Settlement classes were discussed extensively in November, but without reaching even tentative conclusions that could be embodied in a revised draft. One of the most difficult questions is whether it is possible to provide meaningful guidance on the use of "futures" classes of people who have not yet instituted litigation, may not realize they have been injured, and indeed may not yet have experienced any of the latent injuries that eventually will arise from past events. Classes of future claimants can achieve orderly systems for administering remedies that avoid the risk that present claimants will deplete or exhaust defense resources -- including liability insurance -- and preempt any effective remedy for the future claimants. There are serious questions that remain to be resolved, however, and that will be addressed in actions now pending on appeal.

Rule 23(f): Interlocutory Appeals

Specific discussion of the multiple drafts provided in the agenda turned first to the interlocutory appeal provision in the "minimum changes" draft, Rule 23(f). This provision has endured with no meaningful changes through several drafts, and has encountered little meaningful opposition. Initial concerns about expanding the opportunities for discretionary interlocutory appeals have tended to fade on close study of the limits built into the draft.

The most commonly expressed reservations were revisited. Courts of appeals have actively used mandamus review in several recent cases, providing the needed safety valve for improvident class certifications. If an explicit interlocutory appeal provision is added, every case will generate an attempted appeal. A heavy burden will be placed on appellate courts. The cost and delay will be substantial. No lawyer worthy of pursuing a class action will let pass an opportunity to appeal.

The common responses also were revisited. The extraordinary writs should not be subject to the pressures generated by Rule 23 certification decisions. Mandamus should remain a special instrument. The burden of applications for permissive appeals under § 1292(b) is not heavy; court of appeals screening procedures are effective. Motions for leave to appeal will be handled in the same way as other motions. And early review is desirable.

It was noted that the Appellate Rules Advisory Committee is engaged in drafting an Appellate Rule that would implement proposed Civil Rule 23(f). The initial proposal would have amended Appellate Rule 5.1 to include Rule 23(f) appeals as well as appeals from district court review of final magistrate-judge decisions. On consideration, the Appellate Rules Committee determined that it should attempt to collapse present Rule 5.1 into Rule 5, so that there will be one single Appellate Rule that includes all varieties of appeals by permission, present and perhaps future. It is hoped that the product will be available for consideration by the Standing Committee at the same time as Rule 23(f).

One modest drafting change was suggested. The most recent draft refers to appeal from an order "granting or denying a request for class action certification." Deletion of "a request for" was suggested on the ground that it might be redundant, or alternatively might effect an unwise restriction by failing to provide for appeal in the particularly sensitive situation in which a trial court has acted on its own motion to grant or deny class certification. The deletion was approved unanimously.

As revised, new subdivision (f) was approved unanimously.

Benefits and Burdens of Class Action

The next portion of the minimum changes draft to be discussed was (b)(3) subparagraph (F). This draft simplifies the draft that emerged from the November meeting. The November meeting generated a subparagraph (G): "whether the public interest in -- and the private benefits of -- the probable relief to individual class members justify the burdens of the litigation[.]" The minimum changes draft renumbers this factor as subparagraph F, and eliminates any explicit reference to the public interest: "whether the probable relief to individual class members justifies the costs and burdens of class litigation." In this form, the factor emphasizes the importance of the relief to individual class members -- even a significant aggregate sum, when divided among a large number of plaintiffs, may provide such trivial benefit that the justification for class litigation must be on grounds other than the benefits to individual class members.

The origin of the probable relief factor lies in concern that Rule 23(b)(3) is an aggregation device that, separate from the special concerns reflected in (b)(1) and (b)(2) class actions, should focus on the individual claims being aggregated. The traditional focus and justification for individual private litigation is individual remedial benefit. Most private wrongs go without redress. Class treatment can provide meaningful redress for wrongs that otherwise would not be righted, and the value of the individual relief can be important. But class actions should not stray far from this source of legitimacy. Public enforcement concerns should enter primarily when Congress creates explicit private enforcement procedures. As the note to one of the drafts articulated this view, "we should not establish a roving Rule 23 commission that authorizes class counsel to enforce the law against private wrongdoers." Focus should hold steady on the objective cash value and subjective intrinsic value of the relief available to actual class members.

The "corrective justice" and "deterrent" elements of small-claims class actions were noted repeatedly as a supplement to the focus on private remedies. It was urged that consideration of the value of probable relief to individual class members does not foreclose consideration of these elements as well. But it also was urged that indeed this factor should focus only on the value of private relief. Any other view would put courts in the position of weighing the public importance of different statutory policies, and perhaps the relative importance of "minor" or "technical" violations as compared to flagrant or intentional violations.

Discussion immediately turned to the two central elements of the formulation. How is a court to predict the probable relief? And what are the costs and benefits invoked?

One suggestion was that attention should focus in part on a determination whether the motivating force of the class action is a desire for attorney fees.

"Probable relief" in the (b)(3) context is damages. The example that was used in much of the ensuing discussion was an overcharge of a 2¢ a month imposed by a telephone company for 12 months on 2,000,000 customers. The aggregate damages of $480,000 are not trivial. But it is not clear that such a class should be certified.

Discussion also wove around the question whether assessment of "probable relief" includes a prediction whether the class claim will prevail on the merits. In the November discussion, the probable relief factor was held separate from consideration of the merits. The calculation was to be made on the assumption that the class position would prevail on the merits. If direct consideration of the probable outcome on the merits is eliminated, however, it is possible to incorporate a prediction of the outcome on the merits in measuring the "probable relief." Language reflecting that possibility is included in the note that accompanies the draft that eliminates the more direct references to outcome on the merits.

Consideration of the substantive merits of the underlying claims through this factor, not as an independent matter, led to the oft-discussed fear that consideration of the merits would lead to expanded discovery surrounding the certification decision. The comparison to preliminary injunction proceedings was noted -- they may entail much or little discovery -- but found not helpful because of the special factors that affect preliminary injunction decisions. A preliminary injunction decision may be converted to trial on the merits when circumstances permit full information to be assembled and presented before the need to restrain. It may rest on a small fraction of the information needed for trial on the merits. The driving force is the need to preserve the capacity to grant effective relief on the merits, not the calculus of class certification.

It also was asked whether the present rule that certification decisions must be made without reference to the merits is, in practice, a fiction. Explicit recognition of what many feel is a common practice, left unspoken because consideration of the merits is supposed to be forbidden, might lead to wiser reliance on the probable merits.

One effort to bring this role of the merits to a point was made by asking whether the rule should refer to the probable value of the "requested" or "demanded" relief, so as to focus only on the relief, not the merits. This suggestion was quickly rejected.

Alternatives to considering the merits at the certification stage were suggested. One was to require particularized pleading of the elements of each claim offered for class treatment.

Cases with multiple claims were discussed. If one version of a class claim would afford substantial relief, that should be sufficient at least for initial certification. Recognizing that the question of class definition is interdependent with the questions posed by multiple claims, it was understood that the probable relief on all claims suitable to a single class could appropriately be considered and weighed against the costs and burdens entailed by class treatment. At least conceptually, it may be that certification is proper as to some class claims but not another claim that would add greater costs and burdens than the probable relief on that claim.

The problem of weighing returned, with the question whether individual claims averaging a few hundred dollars would justify class treatment. It was noted that the median individual recovery ranges reported by the Federal Judicial Center study ran from something more than $300 to something more than $500. What is to be weighed against the predicted recovery? "Every possible argument will be made." Class proponents will argue public enforcement values.

John Frank addressed the Committee, urging that trivial claims class actions are a major problem, providing token recoveries for class members and big rewards for attorneys. "This Committee is not the avenging angel of social policy." Congress can create enforcement remedies, some administrative, some judicial, pursued by public or private enforcers.

Further Committee discussion suggested, first, that class actions are not filed on claims that, as pleaded at the outset, would yield only trivial relief. The Federal Judicial Center Study, covering two years in four districts, found 9 cases out of 150 certified classes in which the individual recoveries were less than $100; only 3 of them involved individual recoveries less than $25, with the lowest figure $16. But it was responded that very small claim cases do in fact exist. At least in some parts of the country, very small claims classes are filed in state courts and removed. These cases require enormous administrative work. And they breed cynicism about the courts.

The question of claim size also led to the question whether the initial certification decision should be subject to review as progress in the case provides clearer evidence of the probable relief. Initially plausible demands for significant relief may become increasingly implausible as a case progresses. It was agreed that if there is quick and undemanding certification, the certification decision should be open to reconsideration and subclassing or decertification when it appears that the probable relief fails to justify the remaining costs and burdens of class treatment.

A motion to adhere to the language of the "minimum change" draft passed by vote of 9 to 3. The question whether subparagraph (F) should include consideration of the merits in assessing the probable value of individual relief was discussed further during the later deliberations that voted to discard the explicit consideration of probable merits that was adopted by the November draft.

Need For Class Action

The November 1995 draft added a requirement to subdivision (b)(3) that a class action be "necessary" as well as superior for the fair and efficient adjudication of the controversy. For the reasons noted in the introduction, this concept has been difficult to explain. The draft considered at this meeting suggested replacement of the "necessary" finding by adding a new subparagraph (A) and rewording subparagraph (B). Proposed subparagraph (A) would add as a factor in determining superiority "the need for class certification to accomplish effective enforcement of individual claims." Proposed subparagraph (B) would refer to "the practical ability of individual class members to pursue their claims without class certification and their interests in maintaining or defending separate actions."

The first question was whether factor A is antithetical to factor F as just approved. Factor A suggests that class certification is necessary if claims are too small to support individual enforcement. Factor F suggests that class certification is undesirable if claims are too small. The answer was that the two provisions are complementary. Factor A cuts in two directions. If individual class member claims are so substantial as to support individual litigation, certification may be inappropriate. If class member claims are too small to support individual litigation, certification may be needed to provide meaningful individual relief. But if the individual relief that can be afforded by a class action does not justify the costs and burdens of class litigation, certification should be denied.

The relationship between (A) and (B) also was questioned; in many ways, they seem redundant of each other. The emphasis on the need for class certification for effective enforcement, however, can go beyond the practical ability of individual class members to pursue their claims without certification. Separate actions will not be brought by all members of a class who seem practically able to do so, whether because individual actions in fact are not practicable or because of inertia. Even if separate actions are brought, they may not prove as effective as a class action that pools resources to mount a more effective showing. Class actions also may prove more "effective" for reasons that are more questionable, such as pressure to settle even weak claims that are aggregated into the class. These values of class actions were defended as the heart of (b)(3), the touchstone purpose of aggregation. But it was noted that small-claims (b)(3) class actions have fared quite well since 1966 without any explicit element like proposed factor (A).

The distinction between practical individual enforcement and efficient class enforcement in some ways reflects the distinction between opt-in and opt-out classes. Even with individually substantial claims, there is little reason to believe that the number of participating class members will be the same if the class is certified only for those who opt in as if the class is certified for all but those who opt out. (b)(3) exerts a pressure toward compulsory joinder by requiring an election to opt out of the class. Factors (A) and (B), together with factor (C), allow explicit consideration of the desirability of this inertial pressure to remain in a class for group litigation.

A motion to delete proposed factor (A) passed, 8 to 5. A motion to separate proposed factor (B) into two parts passed unanimously. As restructured, factors (A) and (B) would read: "(A) the practical ability of individual class members to pursue their claims without class certification; (B) class members' interests in maintaining or defending separate actions;".

The discussion noted that the practical ability to pursue individual actions remains a two-edged factor. It weighs in favor of class certification, all else remaining equal, if individual actions are not practicable. It weighs against class certification, all else remaining equal, if individual actions are practicable.

Another drafting change from present factor (B) also was noted. The 1966 rule refers to the interest "in individually controlling" separate actions. The proposed language refers to the interest in maintaining or defending separate actions. This language better reflects the full range of alternatives that must be considered. An alternative to a proposed class action may be a different class action, or a number of different class actions. Other alternatives may include intervention in pending actions, actions initially framed by voluntary joinder, consolidation of individual actions -- including consolidation for pretrial purposes by the Judicial Panel on Multidistrict Litigation or transfers from separate districts for consolidated trial in a single court or limited number of courts, and stand-alone individual actions. Individual members of a proposed class may not "control" many of these alternatives in any meaningful sense, but the alternatives must be considered nonetheless.

Melvin Weiss then addressed the Committee. He has been litigating class actions from a time before adoption of the 1966 amendments. Plaintiff class lawyers were taught then that they were to play the role of private attorney general. That role is confirmed by the adoption of (b)(3) classes. The size of individual class member recoveries was not thought important. The need for private-attorney-general classes is growing. Government enforcement resources are shrinking absolutely, and are shrinking even more in relation to the level of conduct that needs to be corrected. Telemarketing fraud abounds. 900 telephone numbers are an illustration. Suppose most members of a class are hit with $10 or $20 charges for calls to a 900 number, with only a few whose bills run much higher. The government may eventually put a stop to a particular operation, but that provides no redress for the victims. Class-action lawyers do that. It is hard work. It is risky work. Of course class counsel deserve to be paid. If the Committee wants to say that a $2 individual recovery is trivial, it should say so. The matter should not be left to open-ended discretion and open hostility to class enforcement. In one action, the class won $60,000,000 of free long-distance telephone services; this is a "coupon" settlement, but provides a real benefit to class members. Class-action attorneys protect victims. Some even are forced to borrow to finance a class action. These social services should be recognized and appreciated. It would be ironic to cut back on class actions at a time when the rest of the world is admiring American experience and seeking to emulate it.

Peter Lockwood addressed the Committee, observing that factors (A) and (F) do not provide any standards. (A) seems to say the porridge is too hot, (F) that the porridge is too cold, and the whole rule seems to say that courts should seek a nice serving temperature. It is difficult to suppose that a Committee Note could say that a $200 individual recovery is sufficient to justify a class action. This proposal is dangerously close to the limits of the Enabling Act, trespassing on substantive grounds. The purpose of Rule 23 is to enforce small claims that are legally justified. There cannot be any effective appellate review of trial-court application of these discretionary factors. Anecdotal views of frivolous suits, settled by supine defendants, do not justify an unguided discretion to reject class certification. Factor (F) should be reconsidered.

Beverly Moore observed that factor (F) allows refusal to certify a class if individual claims are small, even though aggregate class relief would be substantial and the costs of administration are low. But certification should remain available if in fact efficient administration is possible. If a defendant has a continuing relationship with class members, for example, it may be possible to effect individual notice at very low cost by including it with a regular monthly mailing. Distribution of individual recoveries may be accomplished in a similar manner. Note should be made of this possibility.

Committee discussions returned to the relationships between factor (A), the practical ability of class members to pursue individual actions, and factor (F), the value of the probable relief to individual members. It was noted that factor (F) involves balancing the complexity of the litigation and the costs of administration in relation to individual benefits. Even the 24¢ individual recovery might qualify for class treatment if it is possible to resolve the merits and administer the remedy at low cost. The practical ability factor encourages certification of small-claims classes, just as the probable individual relief factor at times will limit certification of small-claims classes. If it is apparent at the time of certification that the individual value of the probable class relief is small, the certification decision must weigh the costs and burdens of a class proceeding. There is no specific dollar threshold. Individual recoveries of $50 in a "laydown" or summary judgment case may easily justify certification. Claims for $200 or $300 may not justify certification in a setting that requires resolution of very complex fact issues or difficult and uncertain law issues. This approach means that an initial decision to grant certification, relying on substantial apparent value or apparent ease of resolution and administration of the remedy, remains constantly open to reconsideration and decertification if the probable relief diminishes or the burdens of resolution and administration increase.

Prediction of the Merits

The November 1995 draft added a requirement that in certifying a (b)(3) class the court make a finding on the probable outcome on the merits. Two alternatives were carried forward. One would require only a showing that the class claims, issues, or defenses are not insubstantial on the merits. The other would adopt a balancing test, requiring a finding that the prospect of success on the merits is sufficient to justify the costs and burdens imposed by certification. Either required finding would be bolstered by a separate factor requiring consideration of the probable success on the merits of the class claims, issues, or defenses. Many observers, representing both plaintiff and defendant interests, reacted to these alternatives with the concerns noted during the first parts of this meeting. These concerns were addressed in the most recent draft by limiting the requirement to cases in which an evaluation of the probable merits is requested by a party opposing class certification.

It was urged that some form of explicit consideration of the probable merits should be retained as part of a (b)(3) certification decision. A preliminary injunction decision requires consideration of the probable merits in addition to the impact on the parties of granting or denying injunctive relief. The public interest often is considered as well. There is a substantial body of learning surrounding this practice in the preliminary injunction setting that can illuminate the class-action setting. It is appropriate to require a forecast of the ultimate judgment before unleashing a class action. There is much at stake; in some cases, the very existence of a defendant is in jeopardy. The prospect that defendants may not want preliminary inquiry into the merits of a plaintiff class claim can be met by requiring the proponent of certification to make a demonstration on the merits, but allowing the opponent of certification to waive the requirement.

Further support for required consideration of the merits was found by John Frank in recent cases, such as In re Rhone-Poulenc Rorer Inc., 7th Cir.1995, 51 F.3d 1293, which emphasized the fact that plaintiffs had lost 12 of the 13 individual actions that had been pursued to judgment at the time of the class certification. The coercive settlement pressure arising from certification even in face of such litigation results also was emphasized by the court. He urged that it is a false terror to be concerned that stock market disaster will follow a finding of sufficient probable success to warrant certification. We should find a way to junk bad cases early.

Discussion of the Rhone-Poulenc decision led to the observation that the defendants had just now offered $600,000,000 to settle all of the pending individual actions all around the country. This offer shows that the class claims were far from weak. Courts may go too fast about the task if consideration of the probable merits is approved.

Discovery concerns continued to be expressed. Consideration of the merits will lead to merits discovery as part of the certification process, and it will be difficult to limit discovery in ways that do not defeat the desire to avoid the burdens that would flow from actual certification.

Beyond the difficulties engendered by probable success predictions, the Federal Judicial Center study shows that ample protection is provided by motions to dismiss or for summary judgment. Consideration of factor (F), the individual value of probable class relief, will further aid in avoiding trivial actions. If there is any need for added protection, it can be met by making it clear that a court can act on Rule 12 and 56 motions before deciding whether to certify a class.

Without formal motion, it was concluded that the Committee had decided by acquiescence to delete the November draft provisions requiring a finding of probable merit and including probable success on the merits as a factor pertinent to the (b)(3) certification decision.

Attention then turned to the alternative of incorporating consideration of the probable outcome on the merits in the factor (F) balancing of the individual value of probable class relief against the costs and burdens of class litigation. The Committee materials included the suggestion that this result might be achieved by including in the Committee Note to factor (F) language something like this: "In an appropriate case, assessment of the probable relief to individual class members can go beyond consideration of the relief likely to be awarded should the class win a complete victory. The probability of class success also can be considered if there are strong reasons to doubt success. It is appropriate to consider the probability of success only if the appraisal can be made without extended proceedings and without prejudicing subsequent proceedings. This factor should not become the occasion for extensive discovery that otherwise would not be justified at this stage of the litigation. Neither should reliance on this factor be expressed in terms that threaten to increase the influence that a certification decision inevitably has on other pretrial proceedings, trial, or settlement."

Support was expressed for this approach, with the reservation that the draft focused only on the negative. It should be integrated with the statement, agreed upon earlier, that certification may be justified for small claims when there is a very strong prospect of success. Further support was found in the continuing concern that aggregation of large numbers of individually weak claims can create a coercive pressure to settle. Certification often is a major event, even a critical event.

Consideration of the merits in this fashion also was supported on the ground that the certification decision in a (b)(3) proceeding must look ahead to the ways in which the case probably will be tried. The predominance of common issues and the superiority of class treatment depend heavily on the trial that will follow.

This "commentary-in-the-Note" strategy was opposed on the ground that it would whittle down the trial judge's discretion. Even without any discussion in the Note, lawyers and judges will seize on the idea that the value of probable relief depends not only on the amount that will be awarded upon success on the merits, but also upon the probability of success. Factor (F) can be used in this way, and can be found to support departure from the Eisen rule that forbids consideration of probable merits at the certification stage.

Opposition also was expressed on the ground that the initial discussion of factor (F) had assumed that it focused solely on the amount of probable relief, not the probability of defeat on the merits. The problems persist whatever the level of emphasis in the text of the Rule or the Note. Consideration of the merits will entail discovery on the merits, and an expression evaluating the probable merits for certification purposes will carry forward to affect all subsequent stages of the litigation. Even if the Note were to say that this process should not justify any discovery on the merits, nefarious results would remain.

Consideration of the merits, moreover, suggests that certification can be denied because of doubts on the merits even though the case cannot be dismissed under Rule 12 or resolved by summary judgment. Courts in fact require particularized pleading of class claims at a level that supports vigorous use of Rule 12.

It also was suggested that the proposed Note language is not a "soft" compromise of a difficult debate. The Committee should decide what it wants to do, and be explicit in the text of the Rule.

Sheila Birnbaum urged that the suggested Note is a balanced attempt to go beyond the limits of Rules 12 and 56, in a way that focuses on the extraordinary case. There should not be discovery, but the merits should be open to consideration with factor (F).

Beverly Moore suggested that every defense lawyer will want to get into the merits at the certification stage in every case. The Draft Note reflects empirically invalid assumptions that there are many frivolous cases and coercive settlements. That is not so.

Peter Lockwood observed that the draft Note fragment can only address cases that cannot be resolved by summary judgment. He asked how is a court to determine that a case that is strong enough to go to trial on a Rule 56 measure still is not strong enough to certify.

Robert Heim, who had initially supported consideration of the merits, but has moved away from the November 1995 draft proposals, supported the proposed Note on factor (F). The concern with discovery is overstated; there is substantial discovery on certification issues now. And there are cases that are very weak. Judges have felt hamstrung by the Eisen prohibition of merits review. The draft authorizes a "preliminary peek."

Alfred Cortese also supported the proposed note. Some claims justifiably earn certification under (b)(3) because they have merit but cannot practicably be enforced individually. Others should be weeded out.

The proposition that the draft Note would merely open a small door for consideration of the merits was doubted. Once the door is open, legions will march through.

A motion to reject the draft Note discussion of incorporation of the merits in the factor (F) determination was adopted, 8 votes to 5.

A motion was made to say nothing about consideration of the merits in conjunction with the factor (F) determination. It was suggested that the Note has to say something, because in the face of silence many courts will read factor (F) to support consideration of the probable result on the merits. "Probable relief" intrinsically includes the probability of any relief. The motion to say nothing was adopted, 7 votes to 6.

Settlement Classes

The November draft included in subdivision (b)(3) a new factor (H) that included as a matter pertinent to the predominance and superiority findings:

(H) the opportunity to settle on a class basis claims that could not be litigated on a class basis or could not be litigated by [or against?] a class as comprehensive as the settlement class * * *

Discussion began with the question whether this factor should be added. It was recalled that the November meeting discussed settlement classes without reaching any conclusions. There are a wide variety of settlement classes. It seemed to be the consensus in November that not enough is known to support intelligent rulemaking with respect to futures classes. The use of settlement classes under subdivision (b)(1) also seems too complicated for wise rulemaking. But for (b)(3) classes, the Third Circuit decision in the General Motors pickup truck litigation has stirred the question whether a class can be certified only on the hypothesis that certification of that class is appropriate for litigation. Many believe that the Third Circuit opinion permits application of the subdivision (a) prerequisites and the subdivision (b)(3) factors in a way that permits certification of a class for settlement purposes even though the same class would not be certified for trial. Others are uncertain. Settlement classes have been found useful by many courts. The practice has evolved from initial hesitancy to regular adoption as a routine practice. They have worked not only in the exotic cases that attract widespread attention, but also in smaller-scale cases such as a class of 1,200 homeowners seeking post-hurricane insurance benefits. The class probably could not have been certified for trial because there were many individual questions. A class that could not be certified for litigation because of choice-of-law problems, general problems of manageability, the need to explore many individual issues, or the like, may profitably be certified for settlement. Subdivision (H) is the law everywhere, with the possible exception of the Third Circuit. But if Rule 23 remains silent, other courts may be troubled by the uncertainties engendered by some readings of the Third Circuit opinion. On the other hand, it may be argued that courts are in the business of trying cases, not mediating settlements. To certify for settlement a class that the court would not take to litigation is to take courts into the claims-administration business. Just what is properly the stuff of judicial business remains open to dispute.

The first response was that settlement classes are extremely important, for plaintiffs and defendants alike, but that it may not be appropriate to adopt a rule that does not provide a list of factors to help the trial judge. Many settlements, moreover, are important because they provide a means of dealing with future claimants. In some situations settlement may not be possible unless all claimants, present and future, are included. In others, failure to provide for future claimants may mean that by the time future claims ripen there will be no assets left to respond in judgment. Futures classes would be left in the wilderness by this draft.

The next response was an observation by John Frank that settlement classes have been the most offensive part of the current class-action process. They offer a bribe to plaintiffs' counsel to take a dive and sell res judicata. As a moral matter, do we want this in the judicial system? If so, settlement classes should at most be allowed only if the same class would be certified for litigation. And it should be made clear that all requirements of the rule apply to futures classes. There also should be provision for increased judicial scrutiny of any proposed settlement. Professor Jack Coffey's views on this subject are sound. The often-decried "coupon" remedies all have been settlement classes.

The choice was put as a minimalist choice between doing nothing or taking a modest first step. Factor (H) does not speak to the futures settlements now pending on appeal in the Third and Fifth Circuits. It only says that the fact that a case cannot be tried as a class need not defeat certification for settlement.

Another option was offered, suggesting that perhaps subdivision (e) should be amended to include the list of factors for reviewing settlements recommended by Judge Schwarzer in his Cornell Law Review article. Subdivision (e) also might provide that closer scrutiny is required if a class is certified at the same time as a proposed settlement is presented. The Committee has never explored this prospect beyond preliminary observations. Nor has it considered the question whether independent counsel might be appointed to assist in evaluation of a proposed settlement.

Opposition to factor (H) was expressed on the ground that it might encourage judges to certify classes simply in the hope that a settlement would clear the docket. It is unsavory to certify a class that cannot ultimately be tried. How can we receive and certify a class that would not be tried? A related fear was that the factor would encourage certification of litigation classes in hopes that the certification would spur settlement.

Support for settlement classes was expressed on the ground that settlement can avoid choice-of-law problems that defeat certification of a broad class. Article III requirements and personal jurisdiction standards still must be met. A settlement class can make all the difference in resolving massive disputes. The pending silicone gel breast implant cases and the Georgine asbestos settlements come to mind. These settlement classes also can avoid problems of individual causation that would defeat any attempt at class-based litigation. Certification of a (b)(3) settlement class permits dissatisfied class members to opt out.

The view was suggested that cases that rest on a settlement reached before certification are so different that they should be addressed in a separate rule, perhaps as a new Rule 23.3.

It was suggested that perhaps settlement classes should be put in subdivision (e) by a provision allowing the court to waive the requirements of (b)(3) for purposes of settlement. The response was that the proposal is not that the requirements of (b)(3) be waived, but that these requirements be applied with recognition of the differences presented by the settlement context.

Article III and personal jurisdiction questions were addressed briefly. There is a live controversy between individual class members and the party opposing the class; the only question is how many of these live controversies can be resolved by class treatment. Personal jurisdiction concerns are mollified by the facts of notice and opportunity to opt out. In federal courts, moreover, all class members ordinarily will have sufficient contact with the United States to satisfy all due process requirements.

The opportunity to opt out of a (b)(3) class was again stressed as an important factor in the settlement-class equation. Class members will opt out if the settlement represents a bargain to sell res judicata on terms favorable to the defendant. If class members choose not to opt out, having notice of the class and the settlement, they are not hurt. If Rule 23(b)(3) is to be used for mass torts, the choice well may lie between permitting settlement classes and adopting the creative devices that have been used by some courts to substitute for litigated resolution of the required elements of individual claims. The Fifth Circuit decision in In re Fibreboard deals with the difficulties of these devices.

Further support for settlement classes was expressed with the view that most settlement classes "are not fixes. There are legitimate uses." Clients are better off, particularly when the defendants have insurance. Settlement also has the advantage of treating alike people who, although similarly situated, would be treated differently in separate actions. Choice-of-law, differences in local courts and procedure, problems of proving individual causation, and the like ensure disparate treatment if class disposition is not available.

Thomas Willging reminded the Committee of the information provided by the Federal Judicial Center study. Of 150 certified classes in the study, 60 were certified only for settlement. 30 of these 60 had consent to a settlement at the time of certification. 25, "mostly (b)(3) classes," did not, and indeed in 8 of these 25 there was opposition to certification. All of the 25 had at least 2 months between the motion and certification.

A motion was made that Rule (b)(3) should not speak in any way to settlement classes. The motion was defeated by vote of 5 for and 8 against.

Turning to the question of what should be said about settlement classes, the suggestion was that a means should be found to say that the court should apply all the prerequisites of subdivision (a) and the requirements of (b)(3) in light of the knowledge that the case was being certified for settlement, not trial. An alternative suggestion was that subdivision (e) be amended to provide that a trial court may, if the parties consent, certify a settlement class even though a class action might not be superior or manageable for litigation.

The next suggestion was that a new subdivision (b)(4) be adopted, providing that if the parties consent a settlement class can be certified even though the (b)(3) requirements are not met. This suggestion met the response that (b)(3) is the right location if settlement bears on application of the predominance and superiority requirements.

Further discussion of the (b)(4) alternative generated several draft proposals. One would have added a new clause in subdivision (b)(3), at the end of the first sentence: "provided, however, that if certification is requested by the parties to a proposed settlement for settlement purposes only, the settlement may be considered in making these findings of predominance and superiority." It was concluded, however, that the prerequisites of subdivision (a) and the requirements of (b)(3) could more clearly be invoked by adoption of a specific settlement class provision as a new subdivision (b)(4). After various drafting alternatives were considered, discussion focused on a draft reading:

(4) the parties to a settlement request certification under subdivision (b)(3) for purposes of the settlement, even though the requirements of subdivision (b)(3) might not be met for purposes of trial.

As a separate paragraph of subdivision (b), paragraph (4) is controlled directly by subdivision (a). Subdivision (a) also is invoked by the first paragraph of subdivision (b), which repeats the requirement that the prerequisites of subdivision (a) must be satisfied. In addition, the provision for "certification under subdivision (b)(3)" means that the predominance and superiority requirements of subdivision (b)(3) must be satisfied, following consideration of the pertinent factors described in (b)(3).

The phrase allowing certification even though the requirements of subdivision (b)(3) might not be met for purposes of trial is intended to make it clear that the prerequisites of (a) and the requirements of (b)(3) must be applied from the perspective of settlement, not trial.

A suggestion to delete the words "for purposes of trial" was rejected as inconsistent with the need to make clear the differences between settlement classes and litigation classes.

The description of "parties to a settlement" is intended to require that there be a complete settlement agreement at the time class certification is requested. It was argued that provision should be made for a "conditional" settlement class certification, to be made in hopes that a settlement might be reached but acknowledging that the class must be decertified if settlement is not reached. This argument was rejected on at least two grounds. The first was that no prudent lawyer would suggest certification of a settlement class unless agreement had already been reached; if there seem to be cases in which certification is ordered before a settlement is presented before approval, it is either because of bad lawyering or because the parties have chosen not to present an agreement actually reached. The second was that there are undue risks that certification of a settlement class before agreement is reached may lead to coercive pressures to settle, reinforced by the threat of taking an untriable class to trial.

A motion to adopt the proposed subdivision (b)(4) was approved unanimously.

A later motion to reconsider proposed (b)(4) to add "proposed," so that it would recognize a request for certification by the parties to "a proposed settlement." It was objected that this change would encourage certifications that could coerce settlement, based in part on the fear that the certification might be carried forward to trial of an unmanageable class. Certification for settlement purposes should not be available merely because the parties "have an idea about a settlement." The motion failed with 2 supporting votes and 11 opposing votes.

Subdivision (e)

The earlier discussions of subdivision (e) were revived with a suggestion that the special master provision in (e)(3) of the November draft should be adopted. The biggest problem with settlements is that they sidestep the adversary process, depriving the court of the reliable information needed to evaluate a settlement. The idea of the draft provision is to ensure independent review. There is evidence that some state-court judges are simply rubber-stamping class settlements. Some means of independent investigation should be required at least for settlement classes. Adversary process is provided only if there are objectors.

It was objected that this seemingly benign provision could have unintended adverse consequences. There is a problem, but this solution may make things worse. If someone else is appointed to investigate the settlement, responsibility may transfer from the judge to the adjunct. The parties, indeed, may agree on the master, who may provide a less probing inquiry than the court would provide. It is better to leave the responsibility squarely on the judge, who will respond with careful inquiry.

It was suggested that instead of incorporation in subdivision (e), the use of special masters might be noted in the Note to the settlement class provisions of new subdivision (b)(4).

Sheila Birnbaum observed that substantial protection is provided by the requirement of notice of settlement. The parties want to ensure that the notice is sufficiently strong to protect the settlement judgment against collateral attack. At the stage of settlement, it is the defendant who pays for the notice; cost is not an obstacle to effective notice.

The key is adequate class representation. Special masters, or for that matter the class guardians who were suggested in earlier discussion, are no better assurance than direct supervision of the named class representatives. The problem, moreover, arises with other class actions. Classes certified for litigation under subdivisions (b)(1), (2), or (3) may settle after certification. The certification itself may result from stipulation.

John Frank spoke in favor of proposed (e)(3) as "better than a band-aid." It would provide some added protection against the fear of class sell-out settlements.

H. Thomas Wells, Jr., suggested that present subdivision (e) settlement procedure is adequate. If there are problems, they arise from inadequate implementation of the procedure.

It is possible to appoint a guardian ad litem for the class, and appointments have been made when the need arises. Settlement classes can come into being quickly, usually after little discovery. They are "packaged." It is hard for a judge to be an independent examiner. There ought to be an independent voice. But the "guardian" label should be avoided, because many collateral consequences are likely to flow from the label.

Adoption of the draft paragraph (e)(3) was opposed on the ground that courts now have power to rely on masters or magistrate judges, or to appoint guardians or other independent representatives to investigate a settlement. It may be appropriate to comment on these matters in the Note to new subdivision (b)(4), but there is no need for an independent provision.

A motion to add proposed paragraph (e)(3) failed, 5 for and 8 against.

It was observed that hearings are held on subdivision (e) approval motions, and provide the best means of review. There is no explicit hearing requirement in subdivision (e), however. It was moved that an explicit hearing requirement be added. The rule would read: "A class action shall not be dismissed or compromised without hearing and the approval of the court, after notice of the proposed dismissal or compromise shall be has been given * * *." The motion passed with 9 supporting votes.

Maturity

It was moved that subdivision (b)(3) factor C be amended as proposed in the drafts, adding "maturity" of "related" litigation "involving class members." The reasons for adding the maturity factor are those discussed in November, and reflected in the draft Note. The motion carried unanimously.

Subdivision (c)(1)

Subdivision (c)(1) now requires that the determination whether to certify a class must be made "as soon as practicable" after commencement of the action. The draft completely revises (c)(1). The question whether the "as soon as practicable" requirement should be deleted flowed into the question whether it is desirable to propose every possible improvement in Rule 23 at one time. The proposals already adopted will require extensive consideration and will draw much comment during the succeeding steps of the Enabling Act process. There is much to be said for not making the process more complicated than necessary to advance the most important changes. On the other hand, it is not likely that Rule 23 will be revisited for at least another ten years. For the last many months, it has been tacitly assumed that if a few substantial changes are proposed, the many other changes in the draft would fall by the way. We must be careful about the number of changes proposed.

A motion was made to revise subdivision (c)(1) to require determination whether to certify a class "when practicable" after commencement of the action. Substitution of the full draft revision was suggested as an alternative, but put aside because the changes were more stylistic than substantive. The motion was adopted by consensus. It was pointed out that the substitution of "when practicable" would serve the same function as the proposal to add a new subdivision (d)(1) expressly permitting decision of motions to dismiss or for summary judgment before the certification question is addressed. The Note to revised (c)(1) can point out that the revision removes any support for the minority view that the "as soon as practicable" requirement defeats pre-certification action on such motions.

Subdivision (b)(2)

The draft would revise subdivision (b)(2) to resolve the ambiguity that has led some courts to rule that it does not authorize certification of a defendant class. The motion failed by 2 votes for and 11 votes against.

Subdivision (c)(2): (b)(3) Class Notice

The November draft includes at lines 156 to 161 a provision that would authorize sampling notice in a (b)(3) class if the cost of individual notice is excessive in relation to the generally small value of individual members' claims. A motion to adopt this provision was resisted on the ground that it is inconsistent with the new (b)(3) factor (F) that allows refusal to certify a class when the probable value of individual relief does not justify the costs and burdens of class litigation. It was responded that to the contrary, this notice provision will implement the purposes of factor (F) by reducing the costs and burdens of certification, making it feasible to enforce claims that otherwise might not justify class litigation. Some concerns were expressed about the requirements of due process. The motion failed for want of a second.

It was agreed that the proposed revisions of Rule 23 agreed upon at this meeting should be submitted to the Standing Committee with a recommendation for publication for public comment.

New Business

The American College of Trial Lawyers Federal Rules of Civil Procedure Committee has recommended that the Committee take up the question whether the scope of discovery authorized by Rule 26(b)(1) should be restricted. The recommendation is supported by a detailed chronology of past Committee consideration of the many problems that surround the scope and practice of discovery. This topic will be on the agenda for the fall meeting. Earlier discussion of the proposal to amend Rule 26(c) emphasized the early and recent concerns that have tied the scope of discovery to protective-order practice. The Committee has continually sought to sidestep the fundamental question by attempting more modest approaches. The 1993 adoption of mandatory disclosure in Rule 26(a) is the most recent example. The time has come to consider the central questions once again. And thanks are due to the American College of Trial Lawyers for the careful supporting work they have provided.

Standing Committee Self-Study

The most recent draft Self-Study prepared by the Standing Committee self-study subcommittee was included in the agenda, along with a set of questions framed by the Reporter for this Committee. Professor Coquillette, as Reporter of the Standing Committee, suggested that the several advisory committees need not be concerned that the self-study will stimulate a response that must be anticipated by advisory committee deliberations and advice. This Committee took no action with respect to the draft self-study.

Admiralty Rules

Proposals to amend Supplemental Admiralty Rules B, C, and E were added to the agenda at the last minute. It was concluded that better advance preparation will be required to support informed consideration of these proposals. They are carried forward to the fall agenda.

Next Meeting

It was agreed that the next meeting of the Committee will be held on October 14 and 15.

Judge Higginbotham, as chair, closed the meeting by noting deep appreciation and thanks to John Rabiej and Mark Shapiro for their continuing and excellent support of the Committee. He also expressed thanks to all Committee members for sustained, diligent, and successful work.

Respectfully submitted,

Edward H. Cooper, Reporter