The Advisory Committee on Civil Rules met at the University of Pennsylvania Law School on February 16 and 17, 1995. The meeting included many participants who were invited by Professor Stephen B. Burbank as host, and sponsored by the University of Pennsylvania Law School. Committee members who attended included Judge Patrick E. Higginbotham, Chair, and Judge David S. Doty, Carol J. Hansen Fines, Esq., Francis H. Fox, Esq., Assistant Attorney General Frank W. Hunger, Mark O. Kasanin, Esq., Judge David F. Levi, Judge Paul V. Niemeyer, 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. Judge Alicemarie Stotler attended as chair of the Standing Committee on Rules of Pratice and Procedure, as well as Judge William O. Bertelsman as Liaison Member from that committee and Professor Daniel R. Coquillette as Reporter of that committee. Judge Jane A. Restani attended as liaison representative from the Bankruptcy Rules Advisory Committee. Peter G. McCabe, John K. Rabiej, Robert P. Deyling, and Mark D. Shapiro represented the Administrative Office. Judge William W. Schwarzer, William Eldridge, Thomas E. Willging, Laural L. Hooper, and Robert J. Niemic attended from the Federal Judicial Center. Invited participants present were Judge Edward R. Becker, Daniel Berger, Esq., Professor Stephen B. Burbank, host, Elizabeth Joan Cabraser, Esq., Professor Samuel Estreicher, Robert C. Heim, Esq., Phillip D. Parker, Esq., Judge Lowell A. Reed, Jr., Sol Schreiber, Esq., Henry Thumann, Esq., Melvyn Weiss, Esq., and Profesor Stephen C. Yeazell. The observers included Alfred Cortese, Esq., Fred Shoup, Esq., and Professor A. Leo Levin.

The meeting began with welcoming remarks by Dean Colin Diver and Professor Burbank.

Judge Higginbotham introduced the purpose of the meeting as a continuation of the Committee's efforts to gather information about the operation of Civil Rule 23 and the possible opportunities for amending the rule. He noted that the Federal Judicial Center has undertaken a sophisticated and very much welcome effort to gather rigorous empirical data, and observed that it also is important to hear from as many sources of practical experience as possible.

Judge Higginbotham then turned to an outline of the "Contract With America" legislative agenda, and described the more direct ways in which bills growing out of the Contract would -- in current form -- affect judicial procedure. H.R. 10, the Common Sense Legal Reforms Act of 1995, contains many direct procedural provisions, many of them dealing with topics that are outside the reach of the Rules Enabling Act process. Title I includes provisions for "loser-pays" attorney-fee awards in diversity litigation, amendment of Evidence Rule 702, amendment of Civil Rule 11, and pre-filing notice requirements for civil litigation. Title II, amending the securities laws, contains many procedural provisions that have been studied by a subcommittee as noted below. Senator Heflin has reintroduced a bill that would require that a majority of the members of all rulemaking committees be practicing lawyers. Senator Kohl has again introduced a bill that would require that all protective discovery orders be based on hearings and findings relating to impact on public health and safety; it was noted that the proposed amendment of Civil Rule 26(c) now on the agenda of the Judicial Conference was framed after careful study of this bill in an attempt to respond to the underlying concerns in a more effective manner.

Senator Grassley has introduced a bill that would provide fee-shifting in diversity cases, and that would enact an offer-of-judgment statute. The bill is similar to the proposal made by Judge Schwarzer that prompted Committee consideration of Rule 68. The Committee has continued to hold the topic on the agenda; the Federal Judicial Center has not yet completed its study of actual practice under present Rule 68. Rule 68 was one of the issues discussed at an Institute of Judicial Administration meeting in 1993, where among other matters game theory was used to suggest behavior patterns that are confirmed by trial lawyer diagnoses of the probable impact of the proposed amendment. In the face of continuing Congressional concern, it may prove important to move Rule 68 back to a more central place in Committee deliberations.

Another Senate bill, S. 300, includes another array of provisions that would substantially affect procedure. Other bills include provisions designed to control or reduce litigation brought by prisoners. The rather limited present provisions for requiring exhaustion of prison remedies would be expanded substantially.

Professor Rowe commented on the diversity fee-shifting provision in H.R. 10. As intended, and as apparently drafted, fee-shifting would apply only if the action were first filed in federal court, and would not apply if the action were brought to federal court by removal. In any state that follows the "American Rule," the result would be pro-plaintiff: a plaintiff who has any significant fear of losing can avoid fee-shifting by filing in state court, while a plaintiff who believes that recovery is certain can win attorney fees simply by electing to file in federal court. In response to a question, he noted that apparently the center of attention has been on individual plaintiff litigation, not litigation between large business firms that may react quite differently to the prospect of fee shifting. Some House Committee members seem concerned about the deterrent impact of fee shifting on plaintiffs, but for the moment it is difficult to measure the extent of this concern. Fees are not defined as an element of "costs," so there is no Rule 68 consequence; there is no indication that the supporters have given any thought to the possibility of integrating this provision with offer-of-judgment provisions. One of the participants observed that the insurance industry is studying creation of policies to indemnify plaintiffs against fee liability; it was suggested that since policies that indemnify defendants have long been accepted, such insurance would not be found contrary to public policy. Another participant suggested that the bill was designed to deter "frivolous" litigation, and asked whether there is any understanding of the actual impact of such a rule on risk-averse litigants, either plaintiffs or defendants. It was responded that there seems to be some awareness of the problem, but that again it is difficult to get much sense of the depth of understanding. It also was asked whether any thought has been given to requiring attorneys to file periodic statements as fees accumulate, so each side will know what its exposure is; the simple answer was "no." It was suggested that the bill may not be particularly pro-plaintiff -- that there are not many "sure-fire" claims in the world of litigation, nor many plaintiffs so confident as to believe they have one. Finally, it was pointed out that the rule has the strange character that it is overwhelming for plaintiffs who are, although poorly, able to respond to a fee award in at least some measure, while it has no effect against an impecunious plaintiff who is unable to respond at all. This effect is often encountered in England.

Further discussion noted that Rule 11 is specifically targeted in a number of bills that seek to undo many provisions of the 1993 amendment, although at least most bills seem to preserve the "safe harbor" provision. Much of the concern with "abusive" litigation is focused on product-liability actions, arising not so much from claims that are unfounded under current law as from dissatisfaction with current law. Since product liability rules also are a topic of close congressional attention, the focus could change. At any event, it may prove important to begin gathering information about the actual impact of the 1993 amendments. On an anecdotal level, we know -- or think we know -- that the level of "satellite" Rule 11 litigation has dropped dramatically. But it may be responded that this is because Rule 11 has been gutted, not because there is any reduction in abusive litigation. One of the most important questions will be to study the operation of the safe harbor provision. It is possible that the provision is working well -- that service of Rule 11 motions before filing has the desired effect of causing frivolous assertions to be dropped, and might even prove more effective than the earlier practice because it encourages cost-free abandonment rather than dig-in defensiveness. Both the Federal Judicial Center and the American Judicature Society would like to do studies of the impact of the 1993 amendment if funding can be found. It was suggested that many grass-roots efforts may be having an impact on frivolous and abusive practices as well, growing out of recent concerns for civility in litigation, Civil Justice Reform Act plans, and the like.

Discussion by the participants suggested a variety of views on the ways in which the Advisory Committee might respond to legislative proposals that affect rules of procedure. One view was that the Committee should not be unduly reserved, that Congress truly wants neutral advice on troubling policy issues. Even on this view, the when, where, and how questions remain difficult. Another suggestion was that other groups, such as the American Bar Association Litigation Section and the Association of the Bar of the City of New York, have found it effective to create position papers that are made available to all participants in the legislative process. The Committee was reminded that many of these legislative proposals are -- or are closely tied to -- substantive matters that the Committee cannot comment on. The Committee is limited in its role and what it can say.

Thomas Willging then presented a preliminary phase of the Federal Judicial Center study of class actions, based on analysis of the class actions in the most recent FJC "time study." The study is being conducted chiefly by Willging, Laural Hooper, and Robert Niemic, with the guidance of William Eldridge.

The first lesson learned in the FJC study was that the Administrative Office data on class action filings are hopelessly incomplete. The preliminary report based on analysis of those findings, presented to the Civil Rules Committee at the meeting in October, 1994, has been retracted. This sad lesson was learned while beginning the intensive study of all class actions terminated in the Eastern District of Pennsylvania from July 1, 1992 to June 30, 1994. They found many class actions that had not been reported to the Administrative Office: Administrative Office figures showed 38 terminated actions, while an additional 99 were found. These numbers are conservative counts -- if, for example, ten actions were brought against one defendant, they were counted as one action. That means that 72% of the filings were missed by the Administrative Office figures. All prior studies and reports of Rule 23 actions based on Administrative Office figures accordingly are suspect. And there are no reliable national data on class-action filings. It is difficult to guess at the causes or nature of the underreporting. One possibility, for example, is that there is a greater tendency not to report such actions as prisoner filings that simply include a boilerplate reference to action "on behalf of all other persons similarly situated." The Administrative Office recognizes the problem, and efforts are under way to correct the data gathering. Even prompt corrections, however, will mean that it still will take several years to accumulata data that can support studies of trends over time.

There was substantial general discussion of the difficulty of making a complete count of class action activity, even by such means as computer searches of clerk's records for the word "class." Many participants believed that there are class actions, particularly in the "civil rights" fields, that never come to the surface.

The Time Study data are quite different. They involve a purely random national sample of all cases filed during selected brief periods between November 1987 and January 1990. A total of 8,320 cases were included; 51 of these were class actions. They give a "small but clear picture," but there is no way to know whether the picture is representative. A few of the more interesting aspects of these cases were noted briefly. 24% of the cases arose under the securities laws. Civil rights cases -- prisoner, employment, and "other" -- together accounted for 36% of the total. Only 2 of the 51 asserted diversity jurisdiction, a pattern paralleled in the data gathered in the Northern District of California and the Eastern District of Pennsylvania; the restrictive jurisdiction doctrines have had the expected impact. Only half of the cases had "class activity" after the Rule 23 assertions in the complaint. There were no defendant classes. There was relatively little debate about which type of class, (b)(1), (b)(2), or (b)(3), should be certified. Motions under Rule 12(b)(6) and Rule 56 were frequently made and decided before a decision whether to certify a class. Cases that were certified had a much higher rate of in-court settlement. There was a wide range in the ratio between class recovery and attorney fees; although the data are sparse, there tended to be an inverse relationship between the amount recovered and the ratio, with fees falling to a smaller proportion of recovery as recovery increased. Cases filed as class actions took more judge time, and those certified as class actions took much more judge time than the "average" civil action.

During discussion, it was stated that the ongoing study is noting whether settlement is announced simultaneously with the motion for class certification. Identity of counsel also is noted, to determine whether class actions commonly are brought by the same repeat counsel, or instead are often brought by counsel with little class-action experience. One comment was that there is lots of litigation over the distinction between (b)(1), (2), and (3) classes in mass torts, even though there is not so much dispute in other subject areas. It was observed that in districts with local rules requiring that a class certification motion be made within a defined time after filing, there is a strong pressure against pre-certification disposition of Rule 12(b)(6) and 56 motions even though there is no requirement that there be a prompt ruling once the certification motion is made. It also was observed that there are cases like Exxon Valdez and the Chicago drug antitrust cases where class actions are tried simultaneously with large numbers of individual actions.

Judge Scirica and Professor Rowe then presented the report of a subcommittee on pending securities litigation legislation. The subcommittee also included Committee members Doty, Vinson, and Wittmann. Judge Scirica served as chair, and Professor Rowe as reporter. From the perspective of the Advisory Committee, the central question posed by these bills is whether securities litigation is so unique that it needs special procedural rules, displacing the authority of the trial judge to work out the best answer under the more general and flexible authority of current procedure. The Committee cannot undertake to advise Congress on the substantive provisions of the bills, and must recognize that the procedural provisions are closely affected by the substantive provisions.

The rapid evolution of the legislative process complicates the task of determining whether the Committee can provide any helpful advice to Congress on the procedural aspects of the securities litigation reform bills, and if so how the advice might be provided. Much of the attention so far has focused on H.R. 10. Earlier versions of the bill required that class representatives have specified minimum shareholdings; that has been dropped. Another approach to reform, primarily substantive, has been to make it harder to win on the merits -- that approach too is evolving, as shown by initial elimination of recovery for recklessness, followed by substitution of gradually expanding definitions of recklessness. The shifting approach to "fraud on the market" theories of liability provides another substantive illustration. Loser-pays attorney fee provisions also are undergoing change, gradually increasing the prospects that a losing party may avoid fee liability.

One directly procedural approach is to adopt heightened pleading requirements, demanding very detailed pleading of scienter, limiting the plaintiff to one amendment of the complaint, and staying discovery during the pleading stage (subject to exceptions).

The statement and testimony of Chairman Levitt of the SEC was praised as a very good identification of abuses and possible solutions. The central point of the statement is that private rights of action are essential means of policing the private securities markets, a vital supplement to SEC enforcement. There are a number of specific SEC recommendations that will be seriously considered by Congress. One recommendation is that Congress should ask the Advisory Committee to study Rule 9(b) pleading standards, and to address the Circuit split on application of the specificity standard for pleading fraud. The SEC strongly favors creation of safe-harbor rules for forward-looking statements; liability for recklessness; and private liability for aiding and abetting violations. The SEC also is undertaking to develop a program for filing amicus curiae briefs on dispositive motions in private litigation, and also on fee applications. The SEC likes joint and several liability on a proportionate basis, but only for cases of recklessness rather than actual intent. The SEC endorses legislative proposals to limit races to the courthouse, and to prohibit disproportionate recoveries by representative plaintiffs. It is skeptical about proposals to increase oversight of class counsel by guardians or plaintiff steering committees.

Provisions that have disappeared, at least for the moment, include those that would include the power to adopt rules governing court procedure in the SEC's rulemaking authority for safe-harbor protections, and provisions for appointing a guardian ad litem for a plaintiff class. In addition to the specific pleading provisions, provisions for a plaintiff class steering committee would establish a unique and complicated substitute for the ordinary class representatives recognized under Rule 23. The purpose of the steering committee proposal seems to be to establish a "real client" for class counsel; because it is difficult to know how this scheme would work, it is equally difficult to know what alternative means might be found as a general revision of Rule 23 to accomplish the same ends.

Discussion of the Rule 23 aspects of the securities litigation bills included a comment on the practice adopted by some judges of soliciting bids by competing firms to become class counsel. This procedure was said to add an undesirable layer of complexity to getting class actions initiated. The next observation was that there are no problems in class actions in the securities field of any particular importance. With a possible exception for attorney fees, the institution is not out of order and does not need fixing. It also was suggested that opt-in classes might prove useful, and even that a blend of opt-in and opt-out features might be used in a single case, relying on opt-out for smaller investors and opt-in for larger investors who might realistically bring independent actions. A rejoinder was that institutional investors like things as they stand: by failing to become representative parties they can avoid the burdens of discovery, remain in the background, and still participate in the recovery. Again, it was suggested that if some version of the securities litigation legislation should be adopted, the Advisory Committee should consider adopting a new Rule 23 version in some form that incorporates the legislative procedures so as to avoid the confusion that could arise from attempts to integrate the statutory procedures with the procedures that still are governed by the rules.

A quick summary was provided of other procedural provisions in the securities litigation portion of H.R. 10, including: disclosure of settlement terms; security for payment of costs (including attorney and expert fees) in class actions; the time for seeking attorney fees, set differently than Rule 54(d)(2)(B); discovery sanctions; a limitation on dismissal or withdrawal that clearly would affect Rule 41, and might affect such matters as the opt-out feature of Rule 23(b)(3); jury interrogatories on demand as to the scienter of any defendant; and, again, the specific pleading requirements. It was observed that dismissal for failure to meet specific pleading requirements would be followed by a new action, brought by different counsel, improving on the lessons learned from the first action, and so on until a plaintiff managed to meet the requirements and go on with the litigation.

In conclusion, it was repeated that the focus of the Committee must be on what, if anything, can be done in the rulemaking process to respond to the concerns reflected in these legislative proposals. Congress believes that there are serious procedural problems, and it is important for the Committee to seek to learn as much as can be about the nature of Congressional concerns and the underlying realities. Coordinated and sympathetic study and response are called for.

Friday morning began with a presentation by Professors Cooper and Rowe on the Rule 23 draft that has been developed by the Committee over the last several years. Much of the discussion explored the likely workings of the draft if it were adopted in its present form, and the assumptions that seem to underlie its provisions. Several of the participants observed that their own experience reflected the lesson of the FJC analysis of the time study data -- there is seldom much dispute about the choice between a (b)(1), (2), or (3) class. At times, however, the choice is vigorously litigated, particularly in the emerging uses of class actions to manage mass tort claims. The litigation in these cases may be very much concerned with the impact of the category of the class chosen on notice requirements and the opportunity to opt out of the class. This phenomenon was thought compatible with the observation that class action practice has matured with respect to many fields, in which lawyers and courts generally know whether a class should be certified and what type of class is appropriate, while practice is less certain in other fields.

It was noted that the distinctions between (b)(1), (2), and (3) classes need not be dissolved to pursue changes in notice and opt-out requirements. Specific provisions addressing notice in (b)(1) and (2) classes can be added, and the demanding provisions for notice in (b)(3) classes can be tamed,directly. In like fashion, the lack of any provision for opting out of (b)(1) or (2) classes, the requirement that members be allowed to opt out of (b)(3) classes, and the lack of any opt-in alternative, can be addressed directly. Some participants believed that it is better to maintain the now "traditional" division among different forms of classes, in part because the different classes have markedly different histories and purposes. The (b)(1) class has an ancient lineage that helps to legitimate class practice in general; (b)(2) classes reflect the proud civil rights heritage of the 1960's; and (b)(3) classes represent both the dramatic expansion of remedies for small claims that could not profitably be pursued in individual actions and the growing efforts to aggregate claims that are (or would be) brought in individual actions.

There was some moderate but inconclusive discussion of the value of opt-in classes. Such classes could address some of the difficulties encountered with particular forms of class actions, most obviously defendant classes. They also could be useful in mass tort cases, not only to assuage concern about individual control of individual claims but also to obviate concerns about the inconvenience of litigation in a distant forum and the difficulty of working through choice-of-law problems. A plaintiff who opts into a class that is certified to apply a particular body of law in a particular forum has not been coerced on either score.

Notice also was discussed. There was some sentiment in favor of adopting more pointed provisions detailing the content of notices. One suggestion was that it might help to draft an illustrative notice to be included in the Appendix of Forms. No particularized suggestions as to content were made.

Substantial concern was expressed about the impact of the "willing" representative requirement on class actions. It was noted that defendant class actions are useful in suing large accounting firms that are organized as partnerships and in suing securities underwriters who are organized in ways that do not allow for entity treatment. Another illustration of useful defendant classes is provided by actions involving multiple insurers, commonly involving state law claims. This was one of the settings in which it was urged that Rule 23 is a model for state practice, and that it is appropriate to consider state uses in considering amendments. At the same time, it was recognized that the position of the representative defendant can be very complicated. Defending on behalf of others carries fiduciary obligations to the class, and forecloses the opportunity to settle freely. Because the stakes are increased, good-faith representation may seem to require a proportional increase in defense effort. Many actions may generate conflicts of interest between the representative defendant and other members of the class, perhaps obvious and perhaps subtle. The attempt to respond to these problems by requiring a willing representative may not be satisfactory, but alternative responses will require careful thought.

Another portion of the draft Rule 23(a)(4) that drew comment was the reference to fiduciary duty. No one doubts that those who seek to represent a class bear fiduciary duties to the class from the moment they assert representative status. This almost casual reference to fiduciary duty, tied only to recognition of the power to relieve representatives or class counsel from their assignments, does not clarify anything. In part for that very reason, it may cause confusion. Some courts, for example, have thought it appropriate in some circumstances to approve a greater recovery for class representatives than for other class members, reflecting the work (and perhaps risks) undertaken by the representatives on behalf of the class. There is no reason why a brief reference to fiduciary duty in the text of the rule should of itself change this result, but it would provide a text to anchor new arguments. This revision was held up as a model of the well-intentioned changes that should not be undertaken without a better focused purpose and more clearly expressed implementation of that purpose.

The focus of the draft amendments on issues classes was discussed briefly. It was recognized that the changes do no more than underscore options that are available under the rule as it stands. And some skepticism was expressed about the desirability of certifying classes for a single issue -- resolution of one issue of liability in a dispersed tort, for example, would simply leave all remaining issues to be resolved in the ordinary course.

This discussion initiated discussion of a topic that recurred repeatedly throughout the day -- whether the problems of mass tort actions are so distinctive that a separate rule should be developed. One advantage might be the opportunity to address the problem of "futures" claimants that seem to be unique to this setting, involving people who have been exposed to an injury-causing agency but who have not yet experienced the injurious consequences, or do not know of the consequences, or do not even know of the exposure. Doubts were raised in response. A specific mass torts rule may seem so laden with substantive overtones as to raise legitimate doubts about the wisdom of invoking regular rulemaking procedures. And experience is in a stage so embryonic as to provide very little foundation for drafting a rule with any real hope of avoiding dramatic unintended consequences.

The proposal for permissive appeal from orders granting or denying class certification also was discussed. Many participants believed that the opportunity for appeal, controlled in the discretion of the court of appeals, is highly desirable. The decision on class certification can have overwhelming importance. A defendant may feel forced to settle by certification, while a plaintiff may feel forced to abandon the claim by denial of certification. Some, on the other hand, expressed doubts. It was suggested that defendants resist certification only when they believe they are liable; they should not be given the opportunity to prolong the litigation and add to the plaintiff's burdens by appealing a certification order.

The next part of the program was a panel discussion led by Judge Schwarzer as to settlement classes, Judge Reed as to "futures" classes, and Judge Becker as to mandatory classes.

Addressing settlement classes, Judge Schwarzer began with the observation that settlement classes seem susceptible to abuse, and certainly have given rise to perceptions of abuse. At the same time, there is no careful empirical evidence on these problems.

A settlement class is one in which class certification is addressed as part and parcel of a settlement: certification is sought at the same time as the parties announce their settlement and seek approval of it through the class action procedure. This phenomenon is developing in mass tort cases. It does not seem likely that those who drafted current Rule 23 ever thought of this. To the contrary, they most likely expected that class definition and notice would be determined by adversary contest, providing information to the court and protecting the interests of class members. In the settlement class, the judge stands alone, dependent for information on one-time adversaries who have joined in a nonadversary request. The judge still is obliged both to decide on class certification and, if certification is granted, on approval of the settlement. How can the judge properly discharge these responsibilities?

This introduction was followed by a list of questions addressed to settlement classes: (1) Should settlement classes be allowed at all? They do have value, but is it enough, on balance, to legitimate the device? (2) If settlement classes are permitted, should Rule 23 be amended to address related problems? The only opportunity the judge has to address the requirements of subdivisions (a), (b), and (c) is provided by the subdivision (e) proceeding for approval of the settlement. How is the judge to get behind the parties' agreement on these issues? (3) So, is it clear that Rule 23(e) should require compliance with (a) and (b) for settlement classes, and provide notice under (c)? What is the burden, what the benefit? (4) If Rule 23 is not adequate, how should it be changed? There are no standards for approval in (e). Appellate decisions discuss factors to be considered, but it does not seem likely that they constitute a comprehensive, guiding body of law. Matters to be considered if amendment is undertaken include: [A] Revision should be trans-substantive, not only for mass torts; the fundamental issues likely are the same across all categories of actions. [B] Changes should be neutral, not targeting any substantive ethical rules and principles. [C] Methods should be devised to ensure that the judge gets adequate information; it would be a start to require findings. [D] So too, methods should be devised to ensure that the judge can consider all affected interests. (5) If procedures can be devised to protect all interests, can a settlement class be made mandatory, to ensure a global settlement? Substantial opt-outs weaken the purpose. Under the present rule, availability of a mandatory class turns on (b)(1) and an assessment of the assets of the defendant as a potentially "limited fund" in relation to the claims; this may be upside down of a more desirable system that seeks to preserve the defendant as an ongoing, contributing social institution.

Judge Reed then addressed futures classes, again seeking to raise questions rather than provide answers. He invited consideration of his opinion at 157 F.R.D. 246 as one illustration of the problems and tentative solutions. A "futures class" involves persons who have been exposed to a product or property, who have not or may not have manifested disability or actual loss, but are aware of a potential for injury. They do not get anything on settlement, but remain eligible for relief in the future. It is very difficult to define such a "late-maturing class" for certification. Indeed, the class may include fact dilemmas beyond the person who knows of exposure but does not know of injury -- there are those who may have forgotten about exposure, or may not even know of exposure.

Futures classes pose difficult questions of standing and "case or controversy" requirements. Is there an "injury-in-fact" simply because of exposure? Judge Reed said yes, but the answer may be shaped by the mature science relating to asbestos.

Due process notice questions also are difficult. Notice is critical to the opportunity to opt out under Rule 23(c)(2), and the opportunity to opt out is critical to an assertion of nationwide personal jurisdiction. See 1993 Westlaw 472812 (October 28). The Manual for Complex Litigation talks about the rule that you do not have to have actual notice. Opportunities for direct and collateral attack are based on notice. The notice must say in bold that the persons addressed are in the class, even if not sick. And it must list the benefits from being in the class.

There are ethical issues facing counsel who represent both present and futures classes: do they need separate counsel? What can the court do about this question when a settlement is presented?

Resolution of futures claims commonly requires a structure for monitoring the class members and for assertion of claims as they mature. These must be communicated to the class.

Attention must be given to notifying the class so that objectors can become informed and make their views known. It is necessary to address such questions as the right of objectors to intervene or demand discovery.

Judge Becker addressed mandatory classes. They are classes not certified under (b)(3). The mandatory classes come with a lot of baggage, of legal background. There are not many problems with (b)(1)(A) or (b)(2) classes. The problem arises from (b)(1)(B) limited fund classes. If there is a truly limited fund, class treatment again makes sense. "Punitive damages overkill" arising from multiple demands for punitive damages in independent actions is a particular problem. There may be a due process violation at some point in a sequence of multiple punitive awards; a mandatory class seems a good resolution. Of course this becomes involved with the substantive law. It worked in Agent Orange in the Second Circuit, but was denied in the Skywalk litigation in the Eighth Circuit and Dalkon Shield litigation in the Ninth Circuit.

The decision in Phillips Petroleum v. Shutts may be a problem, perhaps implying that mandatory classes may violate due process, or at least may violate due process unless there is consent or contact with the forum. The rationale of the Shutts decision must be determined: is it that plaintiffs must retain some control of litigation that disposes of their claims? The "Ticor" case in the Third Circuit involved certification under (b)(1) and (2); later litigation was allowed to proceed on the theory that there were substantial damage claims, giving a right to opt out. The Supreme Court managed to avoid decision of these issues when they came up by way of the Ninth Circuit.

The policy issues surrounding mandatory classes include: (1) efficiency -- this is cheaper than processing multiple individual actions. (2) Fairness -- it may be cheap, but is it fair? How far does fairness depend on the distance of the forum from class members? (3) Autonomy, individual control of litigation affecting individual claims, is important. Is a lawsuit the business of the parties, or the public? (4) Federalism -- particularly, what should be made of state class actions, if claimants are allowed to opt out of any federal class?

One approach would be to make (b)(3) classes mandatory, across the board or in some measure. Another would be to discard any bright-line rule based on the nature of the class, and allow a flexible approach in which opting out may be allowed -- or denied -- in any form of class; the draft Rule 23 already prepared by the Committee illustrates this approach. A decision must be made whether to "push the edge of the envelope" in the mass tort area, where we trench on substantive law and constitutional issues. And, at the very beginning, it must be determined whether it is worth undertaking any revision of Rule 23 when it is not possible to frame firm answers to many of the important questions.

Discussion of the panel's questions ensued. The first question raised was whether there should be a specific provision, perhaps in Rule 23(e), for settlement classes. Judge Schwarzer responded that the rules should be trans-substantive. Mass torts should not be specifically addressed. The idea of a settlement class may be contrary to Rule 23, which contemplates adversary presentation on the issues of certification, notice, and opting out. Perhaps subdivision (e) could be restructured to illustrate the matters the court should consider when confronted with a proposed class and class settlement.

It was observed that there is a large body of caselaw on approving settlement that courts know and follow. And it was responded that the findings made through these procedures cannot be treated with great seriousness. The rejoinder was made that defendants are not interested in unsupported findings that subdivision (a) and (b) requirements have been met, leaving them without protection against future litigation. When defendants contest class definition, they seek a narrow class to reduce the scope and expense of the litigation; but when they get to settlement, they want a broad class to protect against future litigation. This participant tells class representatives at the beginning that they will be stuck with the litigated class definition -- that he will not undertake to settle for a broader class.

It was asked how the rule can be structured so there is a true case-or-controversy on the Rule 23(e) presentation? What will work, as a practical matter? If the rule cannot make it happen, it can create a process. It is not uncommon to have a case filed with settlement and class certification presented at the same time. As a matter of economic reality, these are defendant-driven settlements. Defendants want global settlement and mandatory classes, but do not want to set up a class action framework that will invite claims until they know the claims can be settled. All parties are under enormous pressures to settle -- emanating in part from the court. Defendants must have total peace.

Another participant praised Judge Schwarzer's presentation. It is troubling that lawyers for present claimants also represented the futures class in the Georgine settlement. Judge Weinstein has twice appointed separate counsel for future claimants. Maybe there should be opt-out rights too. No one knows who these future claimants are, how many they are, what their injuries will be: present victim counsel will try to get the most that can be got for the present victim clients. The independent representation theme was echoed by another participant, who suggested that it might be helpful to provide in Rule 23 for independent counsel for subclasses.

In similar vein, it was stated that there is an adequacy of representation problem. Class counsel is trying to resolve too many claims; the court should appoint counsel for at least one subclass to decide "whether the pot is big enough, and whether it is being divided fairly." It is important to distinguish between creating the fund and dividing it. Rule 23(e) should not be drafted in a way that interferes with this. And it would be good to put the fund immediately under the control of the plaintiff class or subclasses, so that they can reap the earnings while division is being accomplished.

It was suggested by another participant that the key lies in defining what the judicial role is: a rule cannot specify on a trans-substantive basis what must be considered. Most settlements occur after there has been some adversary contest on such matters as class certification. An enumeration of factors in Rule 23(e) would help; most circuits have enumerated such factors. The Manual for Complex Litigation is a big help, specifying a process that begins with preliminary approval, and so on. This process seems to be universally used, even though it is not in Rule 23. In a related vein, it was observed that Rule 23 should not itself spell out a formalized settlement administration procedure -- that the expense and delay would be overwhelming. Another observation was that the sophistication of the trial judge is an important factor in ensuring sound results.

Still another stated that the level of scrutiny of a settlement class varied markedly depending on whether there has been a prior contested certification: if so, there is a framework. If not, there is more scrutiny of the issues.

One of the judges suggested that it should be made clear in Rule 23 that if the certification question first arises at settlement, the lawyers must show the factors that support certification. And a practitioner reverted to the earlier theme in response: there is a change in behavior upon settlement, which deprives the court of the information sources it needs.

It was asked whether interpleader might be used once a fund is set. One reply was that interpleader might be turned into a claims resolution facility of some value. Another was that Rules 22, 23, and 24 work well together. Again, the Complex Litigation Manual shows how they interlock. Rule 22 has been used by insurers for bankrupt clients; Rule 24 has been used for intervention by a class. It is always possible to comply with the forms.

Another participant noted surprise at the concern expressed about supervision of settlement -- in his experience, there is a lot of effective supervision.

The possibility of using special masters to participate in the process of approving settlements was raised, drawing from the Committee's current Rule 23 draft. One participant noted that he had played a master-like role in evaluating confidential information and making recommendations that did not reveal the information to the judge. The possibility of more active investigation on behalf of the court was noted, but not pursued. It was suggested that the need to provide representation for the unrepresented is greater than the need to have the court substitute in some more open-ended and not adversarial role.

Then it was suggested that it is premature to deal with these settlement questions in Rule 23, that settlement classes should be dealt with in the Manual for Complex Litigation. Recent asbestos settlement classes involved negotiated prefiling settlements that worked because all the lawyers involved were very experienced and had large portfolios of cases. There seemed to be no real conflicts of interest. The lawyers feared that if they filed first, an activist judge would have forced them to be recalcitrant or to be bound in a shotgun marriage. They need to know the settlement terms with certainty before they can approach class certification and actual settlement. The problem of future classes is at times referred to as "unk-unks": unknown numbers of claimants who will have claims of unknown size. Severity of injury commonly covers a wide range, from mildly irritating symptoms to death. The uncertainty is extreme. Both plaintiffs and defendants are frightened by the prospect of bankruptcy as an alternative claims-administration system. The plaintiff lawyers in these asbestos futures settlements were the best lawyers in the country on the subject; they knew and cared about what they were doing. The framework of settlement is shaped by the reality of the transaction costs encountered in past methods of resolving asbestos claims. Historically, $1 has been spent on defense for every $1 paid to claimants; the result is that claimants, after paying their own lawyers, have received one-third of the money devoted to the litigation. It is in the defense interest to maintain this ratio if they cannot achieve global settlement. These vast wasted costs make it desirable to attempt a different resolution, to run the risks involved in establishing private diagnostic systems, monitoring, and claims processing facilities. The inability to predict which person will suffer which consequences of what severity forces this mode. But it is probably impossible to attempt to generalize in a rule about the circumstances that make this possible and desirable.

A provisional summary of the issues raised for the Committee by this discussion was then attempted: The Manual for Complex Litigation must be considered as a supplemental device, an alternative to amending Rule 23. That does not answer all of the questions whether Rule 23 should be changed in some ways, or whether other rules should. Congressional remedies may not be possible, as demonstrated by the asbestos experience.

Whether Rule 23 changes are needed at all remains uncertain. The mass tort phenomenon seems to be driving the process. If that is so, it must be asked whether asbestos and breast implant litigation are an isolated phenomenon -- and perhaps, when more is known, may be quite different from each other. The breast implant defendants are haunted by what happened in asbestos, perhaps failing to see the fundamental differences. In the implant litigation there is a definable universe of claimants. The manufacture of implants is a discrete event, occurring over a relatively few years. Traditional Rule 23 approaches can work better here than in asbestos. So is all of this discussion an attempt to design a system for asbestos? And isn't that foolish, in part because too late? The science of asbestos is relatively mature, particularly in comparison to the science as to silicone gel breast implants. And there has been ample experience with the ways in which asbestos cases actually try out; there is much more limited experience with actual trial of breast implant cases.

Taking a broader perspective, it can be said that what is happening in mass cases is not that courts have been finding a way to resolve cases. Instead, the effort to find ways to resolve vast numbers of related cases has developed a procedure that produces a mass courts are not equipped to handle. The result is a mass, not a case. The parties, driven by self-interest, then create a claims processing facility that resembles legislatively created administrative systems. Even as administrative and executive claims agencies seem to be moving ever closer to judicial models, these arrangements move courts away from judicial procedure and toward an administrative role foreign to their history, tradition, and special attributes. "The circle closes."

It was asked whether a blanket prohibition on punitive damages would help process these cases. An answer was offered that it might make things worse, creating an incentive to litigate causation in each individual case.

More open-ended discussion raised a variety of additional issues. One was whether a choice-of-law statute might help; it was observed that in the school asbestos litigation, the Third Circuit found that the law of most states was quite similar. Another was whether Congress might be able to dispense with jury trial in favor of an alternative system of administration: what form might the administration take? One system might be to pick up on the forms being developed by settlements -- to establish defendant-funded systems that allow individual future claims to mature, resolve questions of actual injury, identify the products that caused injury, and provide compensation according to a grid. These speculations were cut short by the suggestion that reliance on action by Congress is a long-shot. The rulemaking process must tend to its own responsibilities on the assumption that Congress will not act.

It was suggested that a court can certify a class sua sponte for purposes of case management, appoint masters to design a claims facility, and so on. If defendants perceive a crisis is in the making, they can anticipate this.

It was observed that the in terrorem effect of aggregating marginal cases may be unfair. Perhaps aggregation should not be allowed until litigation arising from a particular source of multiple injuries has matured. When it seems clear that there are large numbers of well-founded claims, aggregation may become appropriate. Another participant seconded this observation with the wry comment that some lawyers believe that aggregation deals with claims of de minimis strength and value. Yet another found it horrid to force people into consolidated proceedings, forcing settlement.

The discussion turned to substantive issues with the note that Congress has never addressed the aggregation of small claims, and may address the perceived problems of aggregation -- including aggregation of large individual claims -- by addressing substantive law more than procedure. Current interest in federal products liability legislation is an indication of this possibility. But if substantive law is the problem, it was asked, why should the rulemaking committees write procedural rules to address the problem?

The American Law Institute Complex Litigation project was noted next. It focuses on bringing litigation into one court, and seems to assume that federal courts will inherit most of the work, and will be able to do it well. This may be an overly sanguine assumption; most lawyers and courts may not be up to the task. In response, it was asked what is the measure of comparison? If we look to state courts, the view is mixed. Many states are reluctant to consolidate multiple actions even on a county-by-county basis. And if settlement is thought so horrid, we must ask whether the courts are ready to try them as an alternative. Settlement is a contract. Settlement can work, despite questions about fairness. Consolidation can be viewed as a default rule reserved for cases that fail to settle, admitting that case-by-case trial is the nominal mode but recognizing that it is not possible.

Careful consolidation was supported as a device that can accomplish a lot. Courts can develop ways to try cases by samples so carefully selected that parties will go along. Additional incentives can be offered to join, as accelerated trials and the promise of reduced discovery burdens. It may be that plaintiffs will be willing to waive punitive damages for other advantages -- the parties agree not as settlement of the claims but as establishing a procedural framework for resolving the claims. The "Ahearn" settlement, before turning to the futures class, resolved more than 50,000 cases and present claims pending in many different courts; this sort of result can be achieved even without consolidation. It was defendants who set all of this in motion. In the early days of the Bendectin litigation, after an approved settlement was reversed, trial was consolidated on a basis that was mandatory for the Ohio cases but optional for all others; the finding of no causation effectively resolved many cases. A response was that there is a big risk in having a consolidated trial, just as in having a class trial, before the science is mature. (And, after all these years, there still may be some question whether Bendectin will be found a teratogen.)

The other side of consolidation is that plaintiffs may find it forces trial, by leading defendants not to settle. And at the same time, it may build up expectations of those who have marginal claims or no claim at all. Trial level percolation has a value; a few trials make people realistic, and should occur before consolidation is undertaken. Massive repetition of massive discovery need not be a problem, even without consolidation; after the first few cases have unearthed all of the unearthable liability information, it can be shared -- protective orders should not be a problem. (The pending Rule 26(c) amendments would make doubly certain of this.)

But there is a lot of pressure on all parties to settle. Perhaps the Manual on Complex Litigation should teach judges to provide lawyers the means to reach agreement before mass consolidation occurs.

Consolidation and class actions were described as "apples and oranges"; consolidation is not the same as class certification. The Cimeno litigation had the consent of the parties to consolidate.

The final portion of the meeting was devoted to brief summary reactions from the invited participants who had been able to remain to the very end.

Professor Yeazell said that Rule 23 should be revised. Notice should be "delinked" from the (b)(1), (2), and (3) classifications. It is far better that revisions be made by the Advisory Committee and the rest of the Enabling Act process, not by Congress.

Mr. Berger expressed equivocal feelings about amending Rule 23. Amendments might make the rule worse, not better. If any changes are made, they should be incremental. The only hope for achieving a rational product is to follow the Enabling Act procedure. Attention should focus on areas where law and practice are not yet well settled. Courts could use guidance in such emerging areas as toxic torts. It is unclear whether (b)(1) needs attention. (b)(2) is being used in mass torts for compensatory relief. (b)(3) has no workable definition of predominance. The Committee Notes should not be cryptic; they should be designed to help bench and bar. Policies should be stated, recognizing counterpolicies. The chief policy of Rule 23 is accomplishing effective relief.

Ms. Cabraser believes that Rule 23 should be amended. The draft of Rule 23(a) is good, apart from the requirement that a class representative be "willing." The draft is benign in clearly setting out the purpose of Rule 23 and the changes. The only part of the draft that is not helpful is the provision for permissive appeals in subdivision (f). Subdivision (e) would be much improved by incorporating guidelines based on Judge Schwarzer's suggestions. The use of masters or magistrate judges acting as masters could be improved by using the master before the settlement proposal is made to the court for approval. Involvement of a master in helping the parties formulate the proposal could be one of the factors considered in evaluating the settlement; once a proposal comes to the point of submission for approval, it is too difficult to make changes in the proposal. The coverage of settlement approval procedures in the Manual for Complex Litigation should be expanded, or perhaps a separate manual could be prepared for this topic.

Professor Hazard advised that the significance of the (b)(1), (2) and (3) categories should be reduced, but in a way that preserves the old learning. Topics that deserve attention include notice, opt-out, defendant classes, and predominance. There may not be much more that can be said about predominance, although it may be possible to mark the distinctive nature of claims that cannot economically stand alone -- such claims should almost automatically be certifiable. Adequacy of representation is more important than the adequacy of the formal representative. Remember that the representatives of the income and principal beneficiaries in the Mullane case were not members of either class, but strangers appointed to represent their interests.

Mr. Heim suggested that some changes in Rule 23 would be desirable to restore adversarial balance, and reduce implicit economic terrorism. Rule 12(b)(6) and 56 motions should precede certification, reducing the force for prompt certification. Draft Rule 23(f) strikes the right balance on appeal rights. The opportunity to appeal grant or denial of class certification may impede pressure for settlement, but that is a good thing. Rule 23(c)(2) notice provisions can easily be improved. "Settlement classes" may deserve special treatment if there has not been earlier consideration of certification.

Judge Becker advised that draft Rule 23(f), creating a permissive appeal opportunity, is good. Early Rule 12(b)(6) and 56 motions are good. Some actions have class allegations as mere throw-aways; this should not be enough to invoke all of Rule 26(e). The draft Rule 23 cleans up a lot of things; much of it is good. Perhaps the dual requirements of typicality and commonality should be abandoned -- the requirement of typicality adds little to the requirement of commonality. More has to be done on notice. There should be some better provision to care for the class member who does not get notice of the need to file an individual claim to participate in a recovery; the most workable procedure may be to impose an obligation on class counsel to make inquiries when a known significant claimant fails to file a claim.

Judge Reed expressed approval of Judge Schwarzer's criteria for approving settlement. He is very much concerned about the settlement class, which increases ethical burdens on lawyers. The draft Rule 23(f) appeal provision is a good idea, but care must be taken to be sure that it is appropriately limited.

Judge Schwarzer advised that expectations for improving Rule 23 should be set at a reasonably low level. The real problem of mass torts is federalism: federal proposals do not reach state courts. It is not clear that there is any pressing need to amend Rule 23, although the draft is elegant. There is some argument for amending subdivision (e). He would drop the 23(f) appeal proposal; entry of final judgments under Rule 54(b) and permissive interlocutory appeals seem adequate to bring important certification questions to the courts of appeals. Committee notes can provide a lot of guidance.

Mr. Weiss stated that Rule 23 is resilient, and grows to meet new social needs. It is still growing. Any significant change will stir up years of uncertainty and doubt. The Committee should wait ten years, and then revisit the questions of mass torts.

Mr. Schreiber likes Judge Schwarzer's Rule 23(e) suggestions. He observed that some problems arise when the Judicial Panel on Multidistrict Litigation sends complex toxic tort cases to tyro judges -- although there is an understandable need to increase the pool of judges experienced in handling complex consolidated proceedings, it may be wise to arrange the growth of experience in more sensible stages. Some of the more practical problems that should be considered include the need for multiple counsel to represent subclasses or class members with conflicting interests; the practice of "claims made" settlement funds; calculation of counsel fees, whether by lodestar or fraction of the amount recovered for the class; and mass torts -- including the question of relying on statistical projection to resolve individual damages issues, as Judge Real has just done in a case involving 10,000 claimants.

Mr. Thumann urged that at most minor amendments should be made to Rule 23. The Committee draft improves the language of the rule, but does not make any significant change. Why bother? The change he would most favor would be adoption of the draft Rule 23(f) appeal procedure, which provides an opportunity to get out from the in terrorem effect of an improvidently certified class.

These summaries left almost no time for comment. It was suggested that eliminating the typicality requirement might not cause much loss, but also noted that Rule 23 has emerged from a long background of interest-group litigation.

Respectfully submitted,

Edward H. Cooper, Reporter