Emory Law Journal

Morphing Case Boundaries in Multidistrict Litigation Settlements
Margaret S. Thomas Assistant Professor of Law, Louisiana State University, Paul M. Hebert Law Center. I thank the LSU Law Center for providing research support for this project. I am grateful for insightful critiques of an earlier draft of this Article from Robert Bone, Edward Cantu, John Devlin, Gordon Hylton, Liz McCuskey, Allan Stein, and the participants in the 2013 Junior Faculty Federal Courts Workshop at Brooklyn Law School and the 2013 Marquette Law School Junior Faculty Workshop. I thank Jaime Dodge and the editors of the Emory Law Journal for inviting my participation in this colloquy.I also wish to acknowledge the helpful contribution of my research assistants: Jacob Gower, Catherine Sens, Brian Lindsey, and Charmaine Borne, and the past LSU Law Center students in my Federal Complex Litigation course whose fruitful discussion of Vioxx provided the inspiration for this project. Any remaining errors are entirely my own.

Abstract

The boundaries of federal multidistrict litigation (MDL) are blurring, as district courts seek innovative ways to facilitate global settlements to resolve multijurisdictional, multidimensional, national mass torts. The techniques emerging from the district courts have mostly evaded appellate review and received little scholarly attention, but they raise important challenges to traditional understandings of the nature of MDL and complex litigation. This Article argues that factually similar cases proceeding in multiple court systems in mass tort disputes create a “federalism problem” for global settlements: global settlements typically benefit from oversight by a single judge, but often there is no single judge who can exercise control over all the parties who might participate in such a settlement. This Article identifies a trend emerging in MDL settlements that attempts to solve the federalism problem by extending the MDL court’s authority. In the settlement phase, some MDL judges have begun experimenting with the exercise of power over state litigants (and even individuals who made private claims but never filed suit in any court), in order to facilitate global settlements. In this situation, the “case” appears to encompass the national mass tort settlement itself. This Article concludes that the aggregative trend toward transjurisdictional settlement authority in MDL has no basis in the MDL statute. The emerging practice submerges the federalism problem into the settlement agreement without regard to the inherent limitations on the federal court’s structural power, but the federalism problem remains unsolved.

Introduction

Over a decade ago, Professor Erichson observed that complex litigation finds ways to aggregate itself. 1Howard M. Erichson, Informal Aggregation: Procedural and Ethical Implications of Coordination Among Counsel in Related Lawsuits, 50 Duke L.J. 381, 469 (2000). His observation remains true: innovative forms of aggregation in multidistrict litigation (MDL) are developing to achieve global peace in complex litigation. 2See infra Part I. Among the most important of these innovations are transjurisdictional global settlements under the authority of the MDL court, 3For clarity, this Response refers to the federal court presiding over the MDL as the “MDL court.” Some federal judges use the term “MDL transferee court” to refer to this same court (as cases are transferred to this court from somewhere else). See, e.g., Eldon E. Fallon, Common Benefit Fees in Multidistrict Litigation, 74 La. L. Rev. 371, 372–73 (2014) (using the term “MDL transferee court”). This Response treats these terms as synonymous. affecting claimants outside the MDL, including private claimants who never filed lawsuits.

The important disaggregative innovations identified by Professor Dodge in her article pose a challenge to the conventional understanding of complex litigation. 4See Jaime Dodge, Disaggregative Mechanisms: Mass Claims Resolution Without Class Actions, 63 Emory L.J. 1253 (2014). The private claims settlement facilities she identifies as postdispute “disaggregative mechanisms” promise to expedite payment to persons harmed by mass torts by resolving individual claims privately, outside the court systems. 5Id. at 1276–77. As she points out, they are defendant-designed private alternatives to litigation, and they seek to allow defendants to get out of mass tort liability quickly and cheaply. 6See id. at 1257–58. However, early experiments with such devices suggest these private claims resolution mechanisms have difficulty resolving mass torts without intervention from public courts. 7See id. at 1262 (“[W]hile parties elect private disaggregative mechanisms because they yield better outcomes than the default litigation system, in many disputes a public disaggregative mechanism would offer a superior option.”).

MDL and aggregate settlements remain the primary path to global peace in complex litigation. This Response thus focuses on how MDL helps foster global settlements, the forces that drive these settlements toward aggregation, and the structural difficulties posed by mass tort litigation spanning multiple court systems.

The British Petroleum (BP) oil spill litigation provides a useful illustration of the shortcomings of private claims resolution systems. Recent scholarly scrutiny of the results obtained in the private claims facility created by BP after the Gulf Coast oil spill disaster has raised questions about whether private alternatives to litigation provide any real benefits over public litigation. 8See generally Samuel Issacharoff & D. Theodore Rave, The BP Oil Spill Settlement and the Paradox of Public Litigation, 74 La. L. Rev. 397 (2014) (comparing the Gulf Coast Claims Facility to the class action settlement that replaced it, finding that the class action results in greater payments to claimants). The Gulf Coast Claims Facility (GCCF) was supposed to resolve all private claims against BP for the oil spill privately, without resort to any courts. 9Id. at 400 (“Ambitiously, the GCCF set out to expeditiously resolve all of the private oil spill related claims against BP outside of the court system.”). In theory, this informal, private system of paying claims should have offered cost savings to both claimants and BP, and resulted in more efficient compensation. 10Id. at 398 (“This [GCCF] settlement . . . at least in theory, should have been the best of all worlds.”). It failed to do either. The task of crafting a global settlement of private claims ultimately fell upon the federal court system because the private system was unable to achieve global peace. Along the way, the terms of the court-enforced settlement were heavily litigated (and are in fact still subject to dispute). 11See discussion infra note 24.

Despite the GCCF paying out over $6.2 billion to more than 220,000 GCCF claimants in 18 months, 12Issacharoff & Rave, supra note 8, at 400. the vaunted benefits of the GCCF’s private resolution never fully materialized—even for BP, as the defendant that designed it. Thousands of claimants still opted to file lawsuits instead of pursuing claims in the GCCF. 13Id. These suits ultimately were consolidated in a very expensive, lengthy MDL in federal court. 14Id. at 400–01. The document discovery alone involved 90 million pages of documents. 15Id. at 401. The cost and complexity of the MDL caused BP to abandon its disaggregated approach seeking individual settlements in the GCCF and work instead toward a court-supervised mass settlement with the Plaintiffs’ Steering Committee (PSC) in the MDL, ultimately resulting in two traditional aggregate settlements through the formal class action device. 16See id.; see also Dodge, supra note 4, at 1312 (discussing class settlement). The settlements then spawned multiple appeals to the Fifth Circuit. 17In re Deepwater Horizon, 744 F.3d 370, 373–74 (5th Cir. 2014) (affirming the district court’s interpretation of the settlement agreement, and recounting the byzantine procedural history). Interpretation of the BP settlement was first appealed to the Fifth Circuit in 2013, remanded with instructions for the district court to reconsider certain data in calculating “Business and Economic Loss,” then returned to the Fifth Circuit in a second appeal. Id. While the interpretation of the settlement agreement was pending before one panel of the Fifth Circuit, a different panel of the same court also considered a separate appeal of the certification of the class by the district court. Id. at 374; see also In re Deepwater Horizon, 739 F.3d 790, 795 (5th Cir. 2014) (affirming class certification). The GCCF was an abject failure if its goal was for the defendant to avoid the transaction costs of aggregate settlements: these transaction costs have been estimated to have cost BP $600 million. 18See Issacharoff & Rave, supra note 8, at 402 (observing the BP class action settlements included a $600 million reserve fund to cover fees to private counsel, costs of class notice, and discovery in the MDL).

Despite these transaction costs, the highly contested, court-administered aggregate settlements in the Deepwater Horizon case appear to have generated better results for the plaintiffs than the GCCF—though BP has vigorously disputed the interpretation of the settlement agreement yielding those results. 19See discussion supra note 17 (recounting the litigation regarding the interpretation of the settlement agreement). New research by Professors Issacharoff and Rave suggests the GCCF claimants generally received significantly smaller payments than those participating in the aggregate settlements. 20Issacharoff & Rave, supra note 8, at 402, 406 fig.1 (depicting a comparison of claim valuations in the GCCF and class settlement); id. at 413 (“We have been unable to find any significant category of recovery in which claimants did better under the GCCF than under the scheduled payments of the class settlement.”). The aggregate public litigation thus generated higher compensation than private claims resolution, despite higher transaction costs. 21See id. at 402. They observe this creates a paradox that challenges the conventional economic theory that lowered transaction costs should benefit both parties. 22Id. at 403. In fact, their analysis shows public litigation produced better results for injured parties, even though it was much more complicated and expensive. 23See id. at 402. Quite simply, aggregate litigation in federal court appears to have been worth the trouble and cost to bring. 24The ultimate success of the court-administered settlement is not yet final, though the Fifth Circuit upheld the district court’s interpretation of the settlement agreement. At the time of this writing, BP has sought en banc review, and may yet seek review by the U.S. Supreme Court. On March 17, 2014, BP filed a petition to have the entire Fifth Circuit review the terms of the settlement agreement en banc, after losing before a three-judge panel of that court. See Petition for Rehearing En Banc for Appellants BP Exploration & Production Inc., BP America Production Co., and BP P.L.C. at 5–6, In re Deepwater Horizon, 744 F.3d 370 (5th Cir. 2014) (No. 13-30315), available at https://www.thestateofthegulf.com/media/67350/2014-03-17-Petition-for-Rehearing-En-Banc-for-Appellants-BP-Exploration-.pdf. This suggests that dysfunctions within private claims facilities identified by Professor Dodge appear to be real, 25See Dodge, supra note 4, at 1300–01, 1305. and the benefits may be illusory.

The GCCF experience shows that such private resolution devices are likely to be inadequate to achieve global peace without intervention from the courts and some form of aggregate settlement—at least in “elastic” cases with vast numbers of undefined plaintiffs. 26See Francis E. McGovern, Settlement of Mass Torts in a Federal System, 36 Wake Forest L. Rev. 871, 888 (2001) (defining elasticity in mass torts to refer, at least in part, to “the undefined nature and number of plaintiffs”). Elasticity creates special difficulties in resolving mass tort litigation. Professor McGovern has aptly observed that claims resolution facilities that are especially efficient tend to then result in additional claims being filed:In litigation terms, if the supply is the number of cases processed by the system and cost is the transaction cost of that processing, then an elastic mass tort would have an increase in filings, whereas an inelastic mass tort would not. Aircraft crash cases are, for example, inelastic; asbestos cases are highly elastic. Since only 10–20% of all actionable torts result in litigation, there is a universe that remains unfiled. If there is a claims resolution facility that processes claims quickly and at low costs, one can anticipate a much higher filing rate than one would otherwise expect in the tort system. These larger numbers of claims can create a major dilution of benefits, particularly if they are accompanied by large numbers of false positives that cannot be eliminated.Francis E. McGovern, The What and Why of Claims Resolution Facilities, 57 Stan. L. Rev. 1361, 1383 n.66 (2005) [hereinafter McGovern, The What and Why]. Instead of “radically upending the traditional view that aggregation was the only way to resolve mass claims,” 27Dodge, supra note 4, at 1316. flaws in the GCCF demonstrated the necessity of both public litigation and aggregation to resolve elastic mass torts. Disaggregative private claims facilities, like the one used by BP, thus have to be understood in the context of their relationship to MDL, and the elusive quest for global peace in complex litigation.

This Response identifies a trend toward transjurisdictional aggregation in MDL mass torts settlements that runs counter to the disaggregative trend identified by Professor Dodge and other scholars. 28See id. at 1257 & n.14 (discussing scholarship focusing on smaller class actions and disaggregation). MDL courts are finding ways in the settlement phase to collectively resolve staggering numbers of claims in nationwide, transjurisdictional mass settlements. These mass settlements are becoming more aggregated, not less, even when Rule 23 class certification is unavailable. Some of the same problems with class litigation that are producing the disaggregative trend Professor Dodge identifies in the private sector appear to be producing the opposite trend in MDL. 29It may be useful to clarify what is meant by “aggregation” and “disaggregation” for purposes of this article. True, formal aggregation traditionally involved the joinder of parties or certification of class actions. See, e.g., Erichson, supra note 1, at 409. There are, however, myriad other ways suits can be partially aggregated for group resolution, whether formally in court or informally through coordination. See generally Erichson, supra note 1. This contrasts with disaggregation, which emphasizes the separateness of parties, claims, defenses, and remedies. See Dodge, supra note 4, at 1257. A new form of MDL settlement is emerging that is increasingly aggregative, and sometimes even operating transjurisdictionally to control all claims related to a mass tort, regardless of whether they are part of the MDL.

The relationship between the BP MDL and the private claims resolution facility illustrates the aggregative pressure exerted by MDL settlements. Thousands of people and businesses affected by the oil spill brought separate state and federal lawsuits, and thousands more settled their individual claims through the GCCF. 30See In re Oil Spill by the Oil Rig “Deepwater Horizon” in the Gulf of Mex. on Apr. 20, 2010 (Oil Spill I), MDL No. 2179, 2011 WL 6817982, at *3, *5 (E.D. La. Dec. 28, 2011); see also Dodge, supra note 4, at 1256 & n.8 (discussing the GCCF). The MDL court ultimately not only supervised the global settlement of the claims filed in federal court, but it also exercised control of the GCCF settlements and even some claims filed in state court. 31See In re Deepwater Horizon, 739 F.3d 790, 796 (5th Cir. 2014) (noting that BP worked with the PSC “to transfer claims from the GCCF to a program supervised directly by the district court”); In re Deepwater Horizon, 732 F.3d 326, 329 (5th Cir. 2013) (“In March 2012, the district court granted the parties’ request to implement a process to transfer claims from the GCCF to a court-supervised program that the parties agreed to in principle.”); Oil Spill I, 2011 WL 6817982, at *1, *6; Issacharoff & Rave, supra note 8, at 400–02. It initially assessed 4%–6% of the gross amount of all settlements to create a common benefit fund to pay for work by attorneys that benefitted in some general sense all plaintiffs—applying the assessment not only to claims that were part of the MDL when settled, but also some state court cases and GCCF claims. 32See Oil Spill I, 2011 WL 6817982, at *1, *3, *5. The court later modified this to exempt GCCF claimants but not state court claimants. See In re Oil Spill by the Oil Rig “Deepwater Horizon” in the Gulf of Mex. on Apr. 20, 2010 (Oil Spill II), MDL No. 2179, 2012 WL 161194, at *2 (E.D. La. Jan. 18, 2012). The non-MDL claimants in state court (or the GCCF) had not necessarily opted into any global settlement or consented to be part of the MDL—nor even filed individual actions in or been removed to federal court. They had merely brought claims that were part of the same mass tort. When these claimants had negotiated their own individual, separate settlements with BP, a fraction of their individual settlements flowed into the MDL for redistribution by the federal court as part of the aggregate settlement. The private disaggregative device of the GCCF thus appears to have been subsumed by the MDL court’s conception of its power over the “litigation,” as the court exercised control over it.

The parties who settled individually outside the MDL mass settlement thus suffered a double blow: they generally received less money from BP, and some of them still had to pay for common benefit work performed in the MDL. 33See, e.g., Oil Spill II, 2012 WL 161194, at *1 (ordering defendants to withhold 4% of gross settlements, judgments, or other payments to government entities to deposit into a common benefit fund); id. at *2 (ordering defendants to withhold 6% of same for any other plaintiff or claimant); Issacharoff & Rave, supra note 8, at 413. They gambled on higher payments for less cost, and they lost on both counts. This was because BP apparently agreed to pay a premium to get out of the class actions. 34See Issacharoff & Rave, supra note 8, at 404–12. Professors Issacharoff and Rave have demonstrated through the BP settlements that defendants sometimes will pay the higher transaction costs associated with aggregate litigation because they “want peace, and they are often willing to pay for it.” 35Id. at 413. They argue there are thresholds of settlement participation that represent enhanced value to defendants, even when the participation obtained is not total. Id. at 415–16. Indeed, MDL defendants often appear quite willing to pay for that peace through mass settlements, even without the utility of Rule 23.

In the past, the prospect of global peace was once likely to be maximized through aggregated settlements under Rule 23. 36The controversial aggregate settlement of personal injury claims belonging to anyone exposed to Agent Orange in Vietnam was one of the earliest cases to demonstrate the power of Rule 23 to impose global peace. See In re “Agent Orange” Prod. Liab. Litig., 818 F.2d 145, 169 (2d Cir. 1987) (affirming the adequacy of notice to “[a]nyone who believed that he or she had suffered injury as a result of exposure to Agent Orange in Vietnam”); see also Anne Bloom, From Justice to Global Peace: A (Brief) Genealogy of the Class Action Crisis, 39 Loy. L.A. L. Rev. 719, 731–32 (2006) (reviewing the terms of the Agent Orange class action settlement); id. at 734 (observing the impact the Agent Orange class action had on reshaping mass tort litigation); id. at 735 (reporting that, after Agent Orange, federal courts began approving aggregate settlements in other mass tort contexts). Rule 23 was also used to craft a global settlement in the Dalkon Shield product liability litigation, another early example of aggregate resolution of personal injury claims. In re A.H. Robins Co., 880 F.2d 709 (4th Cir. 1989); Bloom, supra, at 735. The floodgates then opened and other cases followed. See Bloom, supra, at 735–36. Indeed, during this period, class treatment of settlements was sometimes prompted by defendants seeking global peace. Id. at 746–47. The prospect of certifying nationwide classes or settlement-only classes in mass torts has waned; 37After the Court’s decisions in Amchem Prods., Inc. v. Windsor, 521 U.S. 591 (1997), and Ortiz v. Fibreboard Corp., 527 U.S. 815 (1999), attempts to use the class action device under Rule 23 to aggregate nationwide mass tort claims to facilitate global settlements have been increasingly doomed because federal courts often decline to certify such classes. See Bloom, supra note 36, at 747; Samuel Issacharoff, Private Claims, Aggregate Rights, 2008 Sup. Ct. Rev. 183, 208 (“As a result, class actions seemed to drop out of the available set of tools for attempting to settle most mass torts . . . .”); see also Jeremy Hays, The Quasi–Class Action Model for Limiting Attorneys’ Fees in Multidistrict Litigation, 67 N.Y.U. Ann. Surv. Am. L. 589, 601 (2012) (discussing the waning utility of class actions under Rule 23 and the imperfect nature of MDL as a replacement). This demise of class actions was long heralded by scholars. See Charles Silver, Comparing Class Actions and Consolidations, 10 Rev. Litig. 495, 500 (1991) (observing in the early 1990s that it was already true that narrow interpretation of class action rules made it difficult for district judges “to craft tort class actions that survive review”). however, MDL judges have responded with creative solutions to wrangle as many claimants as possible into mass settlements. As reliance on Rule 23 has diminished, MDL has ascended as the most important federal procedural device to aggregate (and settle) mass torts. 38Fallon, supra note 3, at 372–73 (observing the increasing significance of MDL). Indeed, by one estimate, as much as 15% of all civil litigation in the federal courts is MDL. 39Id. at 373 (citing John G. Heyburn II & Francis E. McGovern, Evaluating and Improving the MDL Process, Litigation, Spring 2012, at 26).

While MDL may involve judge-created ad hoc aggregative devices that are powerful, flexible, and effective in settling mass torts, it paradoxically does so in a way that nominally emphasizes the disaggregated nature of individual suits filed by separate plaintiffs, transferred into the MDL. Increasingly at the settlement stage, MDL also emphasizes the unitary nature of “the litigation,” treating a mass settlement as a proxy for a “case.”

Though MDL’s unifying power is emerging as a way to effectuate global settlements, MDL nevertheless is defined by statute as a procedural device composed of individual lawsuits, with separate claimants. 40See 28 U.S.C. § 1407(a) (2012). It thus has an inherently disaggregative quality. Traditionally, all individual claimants have to personally opt in to any proposed MDL mass settlement, legitimizing it through personal consent. 41See McGovern, The What and Why, supra note 26, at 1377 (discussing consent of the settlement beneficiaries as a vehicle for achieving legitimacy of the claims resolution facility). The separate nature of the cases transferred to the MDL potentially also creates barriers to efficiency through limits on federal jurisdiction: additional separate cases are nearly always filed in state court, and some of them may be nonremovable. 42Cf. Manual for Complex Litigation (Fourth) § 20.31 (2004) (discussing federal jurisdictional boundaries in MDL litigation). When mass torts spawn cases in multiple state courts and in federal court, aggregative mechanisms can founder. 43Erichson, supra note 1, at 415 (“[A] state court action cannot be consolidated with a federal court action unless the state court action is first removed to federal court, which in many cases cannot be accomplished. Due to these restrictions, consolidation has limited utility as a method of aggregating dispersed cases.” (footnote omitted)).

Modern mass tort “litigation” thus often has several courts with power over various plaintiffs who comprise the litigation’s components. This means that there may be no single court that is well positioned to facilitate and administer the global settlement of the “litigation.” Parallel proceedings create a dilemma for parties in the march toward global settlements: who will oversee the global settlement, if no single court has power over all the settling parties?

This dilemma has become “the federalism problem” of complex litigation, and the problem is magnified in the post–class action era. The federalism problem is fundamental to understanding the radical form of transjurisdictional aggregation emerging in MDL. The emerging breed of MDL settlement responds to the federalism problem by redefining the MDL court as a transjurisdictional settlement manager.

The emergence of transjurisdictional MDL mass settlements has lacked attention from scholars and appellate courts. The trend has evolved slowly, beneath the surface of aggregate litigation. While scholarly and appellate attention has been focused on Rule 23, these transjurisdictional MDL settlements have quietly asserted control over billions of dollars worth of claims in state and federal courts, aggregating on a massive scale at the settlement stage using ad hoc mechanisms, usually with the consent of the parties. The emerging transjurisdictional MDL settlement trend has the potential to subsume not only federal cases but also state cases that could not be brought in federal court, and sometimes absorbs private claims not brought in any court, and even controls other independently created, private claims facilities. 44Professor Dodge identifies such private claims facilities as emerging disaggregative mechanisms. See Dodge, supra note 4, at 1272.

This Response attempts to fill that void by identifying the emergence of these transjurisdictional mass settlements, where federal MDL judges operating to effect global settlements, sometimes beyond the federal court system. Part I thus traces the evolution of the new form of MDL settlement. Part II situates that evolution within Congress’s limitation on MDL power and locates the federalism problem submerged within this emerging form of aggregation.

This Response suggests that the innovations in MDL courts regarding global settlements reflect the emergence of a new understanding of the “case” in complex litigationone that purports to transcend individual parties and their claims and encompasses entire transjurisdictional disputes over mass torts. It argues that this new breed of MDL functions only through acquiescence of other courtsparticularly state judges. Instead of solving the federalism problem in mass tort litigation, it buries it in acquiescence and consent. 45This Article does not address the limits on MDL subject matter jurisdiction at the settlement stage, or how broadly the U.S. Constitution might allow the MDL courts to define the “constitutional case” (in the sense of the litigation unit). Those questions are the subject of a forthcoming follow-up article by this author.

I. The Evolving Nature of Multidistrict Litigation Settlements

Although the MDL procedure emerged from the federal courts’ experience in the 1960s with antitrust cases involving electrical equipment manufacturers, Congress envisioned the procedure having utility in a wide range of subject areasincluding airplane crashes, intellectual property, products liability, and securities suits. 46H.R. Rep. No. 1130, at 2–3 (1968), reprinted in 1968 U.S.C.C.A.N. 1898, 1899–1900; see also In re Phenylpropanolamine (PPA) Prods. Liab. Litig., 460 F.3d 1217, 1230 (9th Cir. 2006) (discussing and quoting the legislative history of the MDL statute). That vision has largely been fulfilled, as MDL has been widely utilized in a broad range of subject matters. 47See Judicial Panel on Multidistrict Litig., Calendar Year Statistics of the United States Judicial Panel on Multidistrict Litigation: January Through December 2012, at 12 (2013), available at http://www.jpml.uscourts.gov/sites/jpml/files/JPML_Calendar_Year_Statistics-2012.pdf (noting that as of December 31, 2012, pending MDLs included cases related to products liability, sales practices, securities, air and other disasters, antitrust, contracts, employment practices, and intellectual property). Mass torts, 48Although “mass torts” is a flexible term, it is generally understood to encompass situations where many people are injured either from a single accident or event, or use of or exposure to the same product, and each has a claim for individual damages. See L. Elizabeth Chamblee, Unsettling Efficiency: When Non-class Aggregation of Mass Torts Creates Second-Class Settlements, 65 La. L. Rev. 157, 164–65 (2004) (discussing the definition, and four categories: mass accidents, dispersed/personal injury mass torts, property damage mass torts, and economic loss torts). however, have been by far the most difficult kinds of cases to organize in MDLs: with many people injured in different places, cases multiply in state and federal courts, often defying efforts to consolidate them due to the lack of removability of some of the state cases. 49See, e.g., William W. Schwarzer et al., Judicial Federalism in Action: Coordination of Litigation in State and Federal Courts, 78 Va. L. Rev. 1689, 1691–95 (1992). The federalism problem has thus been a fundamental, structural feature of elastic mass tort litigation, where there are multiple court systems with power over different parties in the same mass tort dispute.

In federal court, consolidation is usually straightforward. Once the Judicial Panel on Multidistrict Litigation (JPML) orders the creation of an MDL, § 1407 provides specific powers to district courts on the receiving end of a transfer from the JPML to conduct “coordinated or consolidated pretrial proceedings.” 5028 U.S.C. § 1407(a) (2012). The meaning of “coordinated or consolidated” proceedings has received scant attention from MDL courts, even when it impacts their decision-making. Cf. In re Equity Funding Corp. of Am. Sec. Litig., 375 F. Supp. 1378, 1384 (J.P.M.L. 1973) (“We have repeatedly declined to attempt to determine in what way and to what extent the litigation should be coordinated or consolidated. From the very beginning we have left that determination to the discretion of the transferee judge.” (emphasis added)). The structure of MDL thus requires a single federal transferee judge to oversee hundreds, or thousands, or tens of thousands of factually related individual cases filed in many federal district courts across the United States. 51See 28 U.S.C. § 1407(a)–(b). Meanwhile, parties litigating in nonremovable state court actions move forward in parallel proceedings separately from the MDL.

Although MDL was created by federal statute in 1968, 52Id. § 1407. it remains one of the least studied types of federal litigation, receiving scant attention from scholars or the Supreme Court, even as it has become the federal procedural “work horse” in resolving mass torts. 53See, e.g., Deborah R. Hensler, The Role of Multi-districting in Mass Tort Litigation: An Empirical Investigation, 31 Seton Hall L. Rev. 883, 906 (2001) (concluding that MDL in mass tort cases facilitates “the growth of mass tort litigation” despite contractions in availability of class certification). Perhaps this may be due to the relative success of the MDL model: large MDLs generate very little appellate precedent, 54Howard M. Erichson, A Typology of Aggregate Settlements, 80 Notre Dame L. Rev. 1769, 1770 (2005) (“Because settlements in non-class actions need no court approval, they rarely generate reported decisions. In addition, confidentiality agreements frequently prevent publication of settlement terms. For both of these reasons, aggregate settlements tend to fly under the radar of most observers.” (footnote omitted)). particularly compared to class actions. The model encourages (and often successfully achieves) settlement, 55See Hensler, supra note 53, at 894 (“[C]ollecting and transferring like claims to a single judge often encourages settlement of these claims, which may be the real goal of the parties requesting multi-districting.”). without the procedural opportunities to challenge aggregation that are present in Rule 23 for class actions, as there is no class certification process present in the MDL, and likely no opportunity for appeal.

MDL is sometimes classified as a form of aggregate litigation, 56See, e.g., Dodge, supra note 4, at 1256. but it exists as a hybrid that is both aggregate and disaggregate at the same time. According to § 1407, it consists of separately filed lawsuits transferred for consolidated, pretrial proceedings. 5728 U.S.C. § 1407(a). In practice, it may use bellwether trials of selected individual cases (illustrating the disaggregated essence of those cases), 58Eldon E. Fallon et al., Bellwether Trials in Multidistrict Litigation, 82 Tul. L. Rev. 2323, 2344 (2008). then translate the results into rational settlement valuations for a mass settlement of remaining claims (operating in aggregated fashion), 59E.g., In re Vioxx Prods. Liab. Litig. (Vioxx III), 802 F. Supp. 2d 740, 760 (E.D. La. 2011); Fallon et al., supra note 58, at 2338. which requires individual claimants to opt into the settlement (returning their disaggregated identity), 60E.g., Vioxx III, 802 F. Supp. 2d at 760–61; In re Vioxx Prods. Liab. Litig., MDL No. 1657, 2007 WL 3354112, at *1 (E.D. La. Nov. 9, 2007). while overseeing compensation of the group of plaintiffs’ attorneys who brokered the settlement (much like class counsel in aggregated litigation). 61See generally Vioxx III, 802 F. Supp. 2d 740 (allocating common benefit attorneys’ fees). Judges may order individual discovery over some issues, while relying on collective discovery for others. 62Barbara J. Rothstein et al., A Model Mass Tort: The PPA Experience, 54 Drake L. Rev. 621, 625–27 (2006) (discussing strategies MDL courts use to manage discovery). They may create a streamlined, unitary process for filing claims, while insisting those claims specify facts that would be relevant to individual defenses. 63See, e.g., In re Phenylpropanolamine (PPA) Prods. Liab. Litig., 460 F.3d 1217, 1223–24 (9th Cir. 2006) (discussing use of case management orders); see also McGovern, The What and Why, supra note 26, at 1388 (“Just as litigation raises different issues than do global resolutions, the type of information that parties seeking a global resolution may need is typically different from the individualistic, case-by-case approach of common law adjudication. Often there is even opposition to the collection of global data to be used in individual trials; a focus on the entire case can be maintained in the context of settlement without adversely impacting the litigation of individual cases.”).

MDL, like traditional class actions, generally operates at the settlement stage using a claims resolution facility created by a settlement agreement. 64McGovern, The What and Why, supra note 26, at 1380–81 (observing that class actions under Rule 23(b)(1)(B) and (b)(3), “state class actions, bankruptcy, [MDL], and mass settlements all reach closure with a claims resolution facility”). These facilities entail the defendants’ funding of assets from which to pay plaintiffs participating in the settlement, and some process to resolve large numbers of claims against that fund. 65See id. at 1361–62. Despite this resemblance to Rule 23 at the settlement stage, MDL operates outside the formal constraints of Rule 23. It requires no class certification and imposes none of the complex procedural protections for absent class members built into Rule 23. 66See, e.g., Fed. R. Civ. P. 23(c)(2)(B) (describing the notice required for members of a class under 23(b)(3)); Fed. R. Civ. P. 23(d)(1)(B) (describing the court’s role to protect class members); Fed. R. Civ. P. 23(e)(2) (requiring court approval of class settlements); Fed. R. Civ. P. 23(e)(5) (permitting class members to object to settlement terms); Fed. R. Civ. P. 23(g) (requiring the court to appoint and supervise class counsel); Fed. R. Civ. P. 23(h) (giving the court power over attorneys’ fees for class counsel); see also Elizabeth Chamblee Burch, Aggregation, Community, and the Line Between, 58 U. Kan. L. Rev. 889, 898 (2010) (observing that MDL operates as a “procedural no man’s land—somewhere in between individual litigation and class action litigation, but without the protections of either”).

This Part demonstrates that MDL judges’ understanding of the power of the MDL court to “coordinate and consolidate” at the settlement stage is evolving rapidly. In particular, some judges on the vanguard are reconceptualizing the relationship of the MDL to nonfederal cases pending outside the federal court. A controversial new form of MDL settlement is being formed by pressure to pave new paths to global peace.

A. Limited Power over Individual Cases Comprising Multidistrict Litigation

In the 1970s, when MDLs were a new phenomenon, a few early published cases began to sketch the scope of the MDL court’s power. Most of these cases reflected a narrow, formalistic view of that power. 67During this early period in MDL history, “business was slow.” Fallon, supra note 3, at 372. The scope of the MDL court’s power has long been analogized to the power in non-MDL cases that exists under Federal Rule of Civil Procedure 42. This rule permits a district court to order consolidation of separate actions involving “a common question of law or fact.” 68Fed. R. Civ. P. 42(a)(2); see, e.g., In re Equity Funding Corp. of Am. Sec. Litig., 416 F. Supp. 161, 175 (C.D. Cal. 1976). Under this view, upon transfer into the MDL, such consolidation before the MDL court simply creates an administrative, procedural “wrapper” around individual cases, with each case retaining its individual identity. 69See, e.g., In re Equity Funding, 416 F. Supp. at 176 (“[T]he effect of such pretrial consolidation is not and cannot be to ‘merge the suits into a single cause, or change the rights of the parties, or make those who are parties in one suit parties in another.’” (quoting Johnson v. Manhattan Ry., 289 U.S. 479, 496–97 (1933))). Under this view, any aggregative effect is at best partial and limited: the MDL judge was to oversee consolidated proceedings, but the MDL itself was still composed of individual cases, or sometimes a collection of both individual cases and class actions. 70Fallon, supra note 3, at 375 (describing the modern trend of combining class actions with individual actions to form MDLs).

In one early MDL, this concept was challenged when the plaintiffs jointly filed an amended “Unified and Consolidated Complaint” against all of the defendants, replacing their individual complaints. 71In re Equity Funding, 416 F. Supp. at 170–71. The district court understood the consolidation of pleading to be an administrative procedure to expedite proceedings, not a redefinition of the parties or nature of the particular cases. 72Id. at 176.

This early understanding defined the MDL in terms of an administrative procedural wrapper containing individual cases that were transferred into the MDL. This understanding created a strict limit on the power of the MDL court: it extended only to those cases inside the wrapper. 73See id. at 175. This strict limitation created important practical problems, as the use of common benefit funds became the norm in MDLs and defendants sought to settle all related cases in state and federal court in “global settlements.”

As early as 1973, an MDL pending in the Northern District of California involving an Alaskan air crash forced the Ninth Circuit to confront the outer boundary of this definition of the MDL in the settlement context. 74544 F.2d 992, 997–98 (9th Cir. 1976).Hartland v. Alaska Airlines involved more than thirty actions filed in state courts in Washington and Alaska, 75Id. at 994. as well as eight more filed in federal court that were transferred to the MDL. 76Id. Although the alignment of the parties and claims was convoluted, on the surface, the Ninth Circuit understood the federal district judge in the MDL to have purported to have power to “approve” the airline defendant’s settlement with two claimants who had asserted claims that were not part of the MDL. 77Id. at 994, 996–99. One of these claimants had brought suit in state court in Alaska; the other had filed no lawsuit in any court. 78Id. at 994. The district judge nevertheless required both claimants’ counsel to deposit 5% of their settlements into the federal court for a common benefit fund. 79Id. at 996–99. The Ninth Circuit issued a writ of mandamus, reversing this order for lack of jurisdiction over these settlements: these claims were not part of the MDL over which the federal judge exercised power. 80Id. at 1001–02. In other words, they could not be aggregated with the federal proceeding.

The Ninth Circuit’s decision contains little explanation, perhaps because the fundamental jurisdictional premises appear so self-evident. 81See id. at 1001. Both claims had been presented to the airline outside federal court. The case filed in Alaska state court had never been removed to federal court, and there would have been no diversity to support removal. 82See id. at 997. The other claim had been privately presented to the airline, without any case ever having been filed, so no court had asserted jurisdiction over it. 83Id. at 994, 996.

Nevertheless, the jurisdictional question was a good bit more complicated than the opinion’s analysis suggested. The procedural history in the case shows that the district judge had at least a colorable basis to justify exerting control over the settlement: one of two plaintiffs had also separately brought a Federal Tort Claims Act (FTCA) suit in federal court, which was part of the MDL. 84Id. at 997–98. That plaintiff’s settlement with the airline not only released claims against that airline, but also all other claims against other defendants, including the FTCA claim pending in federal court, in the MDL. 85Id. at 999. The MDL court’s exercise of control over the release of claims in the MDL was likely within its inherent power, although the Ninth Circuit did not acknowledge this. Indeed, the circuit’s decision to reverse the district court largely ignored the complexity of the question of jurisdiction over a settling claimant litigating in multiple jurisdictions. With scant explanation, the Ninth Circuit instead emphasized the separateness of the actions, holding the district court exceeded its jurisdiction in attempting to reach the settlement proceeds in the state litigation. 86Id. at 1001.

Hartland is nevertheless important because it illustrates that as early as 1973, the problem of settlements in related cases pending outside the MDL had fully emerged. Moreover, the case reveals how thorny the problem is, when the same plaintiff proceeds in multiple forums against different defendants (some of whom are part of the MDL, and some of whom are not) and then reaches a comprehensive settlement involving both the state and federal action. Nevertheless, the opinion features a very narrow understanding of the scope of the MDL court’s power.

The Hartland model of MDL provided stringent limits on the capacity of MDL to serve as an aggregative device. The formalistic view of the MDL as a procedural wrapper containing individual cases ensured the integrity of each plaintiff’s claim as a separate litigation unit. The Hartland model appears to have been relatively stable, with little reported precedent, until the early 1990s.

In 1992, the Fourth Circuit reached a similarly narrow result in In re Showa Denko K.K. L-Tryptophan Products Liability Litigation–II (Showa Denko). 87953 F.2d 162, 165–66 (4th Cir. 1992). This MDL involved 470 federal cases, with approximately 300 defendants, in a pharmaceutical product liability suit. 88Id. at 164. A district court created a common benefit fund for discovery by requiring each plaintiff in the United States with a claim to pay $1,000 into the fund, as well as 0.5% of any settlement or jury award. 89Id. The district court’s reasoning was straightforward: “whether a plaintiff settles or litigates, he or she benefits from the discovery.” 90Id. The PSC contended that the district court’s motivation was simply to compensate the PSC for its expenses in conducting discovery to be shared by all. 91Id. at 164–65. The order appeared to be targeted at the practical problem of free riders (plaintiffs litigating outside the MDL who benefitted from the work of the PSC to reach a settlement, but contributed nothing to the cost of that work). 92The free rider problem is fundamental to MDL case management. The PSC and various other court-appointed plaintiff committees in MDLs typically handle discovery, argue motions, appear in court, conduct hearings, and sometimes even conduct bellwether trialswork that theoretically benefits all plaintiffs in the MDL, but this work is not compensated while the MDL is pending. Fallon, supra note 3, at 373–74. Common benefit funds have thus long been used in this context to ensure this work is eventually compensated out of the proceeds of settlements or trial judgments in favor of MDL plaintiffs. Id. at 375–76.

The district court’s order was innovative in the way it avoided the mistake of the district court in Hartland: instead of ordering state plaintiffs to contribute to the federal common benefit fund, the order instead purported to bind the federal defendant to certify payment of the assessment in any settlements reached anywhere. 93Showa Denko, 953 F.2d at 164. In other words, the court exercised control over the party before it in the MDL to control the settlement of cases outside the MDL. The MDL court’s order reached cases pending in other federal districts that had not been transferred to the MDL, as well as 683 cases in state court, and 180 potential claimants who had not yet even filed suit. 94Id.

The Fourth Circuit reversed the MDL court’s order, emphasizing the jurisdictional limits of federal courts. 95Id. at 165–66. Relying in part on the groundwork laid by the Ninth Circuit in Hartland, the Showa Denko opinion observed that § 1407 does not expand the jurisdiction of the federal courts. 96Id. Thus, the MDL “court’s jurisdiction . . . is limited to cases and controversies between persons who are properly parties to the cases transferred [to the MDL], and any attempt without service of process to reach others who are unrelated is beyond the court’s power.” 97Id. The Fourth Circuit observed that the district court’s attempt to levy assessments against settlements outside the MDL had “the very real potential of interfering with discovery proceedings in state court.” 98Id. at 166.

In the Fourth and Ninth Circuits, the scope of the MDL court’s power was clear: the MDL court’s power extends only to the individual federal cases transferred into the MDL, and no further. 99 The Second Circuit was an earlier outlier in its approach to the federalism problem. It endorsed the use of the All Writs Act to remove state litigation to federal court in Agent Orange cases when a group of plaintiffs claimed to have discovered their own injuries after the settlement of a federal class action had concluded. In re “Agent Orange” Prod. Liab. Litig., 996 F.2d 1425, 1430–31 (2d Cir. 1993). In other words, it purported to force aggregation in the federal forum. This aggressive interference with separate state litigation was subsequently forbidden by the U.S. Supreme Court in Syngenta Crop Protection, Inc. v. Henson, 537 U.S. 28 (2002). See Chamblee, supra note 48, at 179–80 (discussing the Second Circuit’s use of the All Writs Act in the Agent Orange case and the subsequent reversal by the Supreme Court). The fundamental message in these circuits left MDL judges with a significant practical problem in complex cases with state and federal components: how does one broker a settlement bringing global peace in a complex national mass tort litigation if no single judge has the power to oversee that global settlement, and how will the PSC get compensated for its work in achieving such a settlement?

This question has crystallized the federalism problem of MDL settlements, making the path to global peace in such litigation treacherous. Where litigation proliferates in multiple court systems, each overseeing only part of the total inventory of cases, bringing all of the cases under the umbrella of a global settlement overseen by the federal court is virtually impossible under the Hartland/Showa Denko view of the MDL court’s power.

The federalism problem has grown more acute since the 1990s as it has become increasingly difficult to certify the type of national class actions that would give a federal court supervisory power over such settlements under Rule 23. 100See, e.g., Silver, supra note 37, at 500 (observing in the early 1990s that it was already true that narrow interpretation of class action rules made it difficult for district judges “to craft tort class actions that survive review”); see also Hays, supra note 37, at 601 (discussing the “waning” utility of class actions under Rule 23 and the imperfect nature of MDL as a replacement). MDL courts started pushing back against the stable, historical model of MDL power at the same time that the Supreme Court began constricting the availability of class settlements under Rule 23.

Part II.B examines the emergence of a recharacterization of the role of MDL judges that pushes back against narrow understandings of the scope of the MDL court’s power. MDL began using innovative strategies to affect (and sometimes even capture control over) parties outside the MDL. These strategies respond directly to the federalism problem.

B. Coordination Between State and Federal Courts as an Alternative to Formal Aggregation

MDLs often have thousands of parties and cases that cannot, as a practical matter, be individually adjudicated. The need for global settlements and common benefit funds is often acute—for both the parties and the court. Global settlements risk becoming unmanageable if no single court has the power to enforce them, and they risk failure if not enough plaintiffs in various court systems are included. 101See, e.g., Elizabeth Chamblee Burch, Litigating Together: Social, Moral, and Legal Obligations, 91 B.U. L. Rev. 87, 99 (2011) (“When defendants decide to settle, they want finality. They thus want to sweep as many plaintiffs as possible under the settlement rug.”). Moreover, they require some mechanism to arrange compensation for common benefit work leading to the settlement—in whatever court system that work took place. 102See id. at 123 (noting the problem of “[fights] over common resources—such as settlement funds”).

To solve this problem, MDL courts have long explored methods of coordinating state and federal litigation. 103See Manual for Complex Litigation (Fourth) § 20.31 (2004) (“State and federal judges, faced with the lack of a comprehensive statutory scheme, have undertaken innovative efforts to coordinate parallel or related litigation so as to reduce the costs, delays, and duplication of effort that often stem from such dispersed litigation.” (footnote omitted)); Schwarzer et al., supra note 49, at 1700–01 (discussing early efforts to coordinate cases in the 1970s). Such coordination commenced as early as 1972 in airplane crash litigation, and continued in many other major cases in the 1970s and 1980s. 104Schwarzer et al., supra note 49, at 1700–01.

This coordination has generally been well received by commentators. The Manual for Complex Litigation now recommends that MDL judges “communicate personally with state court judges who have a significant number of cases in order to discuss mutual concerns and suggestions,” and share “pretrial orders and proposed schedules.” 105Manual for Complex Litigation (Fourth) § 20.312. Similarly, the Conference of Chief Justices directed the National Center for State Courts to “take all available and reasonable steps to promote communication between state and federal courts for the purpose of establishing best practices for the management of like-kind litigation that spans multiple state jurisdictions and federal districts.” 106Catherine R. Borden & Emery G. Lee III, Beyond Transfer: Coordination of Complex Litigation in State and Federal Courts in the Twenty-First Century, 31 Rev. Litig. 997, 1008 (2012) (quoting Conference of Chief Justices, Resolution 2: Directing the National Center for State Courts to Promote Communication and Best Practices for the Management of Like-Kind Litigation That Spans Multiple State Jurisdictions and Federal Districts (Jan. 26, 2011), available at http://ccj.ncsc.org/~/media/Microsites/Files/CCJ/Resolutions/01262011-Directing-NCSC-Promote-Communication-Litigation-State-Jurisdictions.ashx).

Communication, sharing, and coordination are only practically useful if a measure of uniformity ensues. Coordination thus likely involves some persuasion and influence by the MDL judges over their state court colleagues, urging them to follow the MDL court’s lead. 107See, e.g., Schwarzer et al., supra note 49, at 1702–03 (discussing a state court’s decision to “track” the federal court’s case management plan in asbestos litigation); id. at 1704–05 (same, in separate asbestos litigation); id. at 1743–44 (discussing the risk of federal judges exerting too much influence over state colleagues). It also creates risks of one court voluntarily ceding control to another. 108See, e.g., id. at 1744 (“In light of federal courts’ greater resources, this tendency [to control the litigation] is understandable, but judges should take care that dominance be avoided if possible.”). Scholars and judges have both warned that interjurisdictional judicial cooperation risks turning into a “collective decision” on the merits of the litigation, which threatens judicial independence. 109Borden & Lee, supra note 106, 1019–20 (discussing criticism by Professor McGovern and Judge Schwarzer).

Cooperation often occurs through ex parte communications among judges. 110Id. at 1014 (noting that a majority of judges aware of ongoing state proceedings communicated with their state counterparties, and of those who communicated, 94% did so directly). Indeed, MDL pretrial orders sometimes reflect an implicit expectation that the state judges will agree to enforce the federal orders in the state actions. For example, in a major products liability case involving medical devices, when the federal court ordered a deduction from settlements for work and costs of the PSC, the order was not just limited to the federal cases pending in the MDL. The judge extended the order to “orthopedic bone screw cases that are finally disposed of in state courts, to the extent the parties agree, or if ordered by the presiding judge or an authorized judge of that state court.” 111Pretrial Order No. 402, In re Orthopedic Bone Screw Prods. Liab. Litig., MDL No. 1014, 1996 WL 900349, at *3 (E.D. Pa. June 17, 1996) (emphasis added). In other words, the parties and state court were invited, but not required, to follow the order—though the implication indicated an expectation that the invitation would be accepted. This type of expectation may be a reflection of joint coordination and personal understandings between the federal and state judges prior to the issuance of such orders.

Hearings at which judges jointly preside are another technique to facilitate such cooperation. 112Manual for Complex Litigation (Fourth) § 20.313 (2004). This technique is rarely used in practice. 113Borden & Lee, supra note 106, at 1021 (discussing a survey of MDL transferee judges finding only 13% of the respondents held joint hearings).

The In re Diet Drugs MDL is a key example of open collaboration between the federal judge and the state judge overseeing related cases. 114See, e.g., Pretrial Order No. 467, In re Diet Drugs (Phentermine/Fenfluramine/Dexfenfluramine) Prods. Liab. Litig., MDL No. 1203, 1999 WL 124414, at *1 (E.D. Pa. Feb. 10, 1999) [hereinafter In re Diet Drugs] (“[T]he court has conferred with the [California state court judge who] . . . . is the state-wide coordinating judicial officer duly appointed by the Supreme Court of California to administer the diet drug cases filed in California state courts.”); Manual for Complex Litigation (Fourth) § 20.313. There were joint state/federal hearings, followed by the MDL court’s coordination order that observed that the California state judge was “expected to approve in substantially similar language” the same order. 115In re Diet Drugs, 1999 WL 124414, at *1. The order purported to require the defendant “[u]pon approval” of the state court judge, to deduct 6% of any amount paid to the plaintiffs, even in the state court cases, and deposit that sum into the MDL’s common benefit fund. 116Id. at *4. The common benefit fund in the In re Diet Drugs case was ultimately controlled by the state judge, rather than the federal MDL judge. 117Manual for Complex Litigation (Fourth) § 20.312. More typically the MDL court assumes control of the common benefit fund, but it might allow attorneys litigating state cases who have no cases in the MDL to make claims for work benefitting the national litigation. 118Id. § 20.312 & n.705.

At its core, the movement toward cooperative interjurisdictional coordination purports to respect the traditional limits on the MDL court’s power by asking the state courts to work together with the federal court and, ultimately, jointly enforce the MDL pretrial orders. While unobjectionable in terms of the formal power of the federal court generally, this kind of interjurisdictional judicial coordination nevertheless has sometimes come under fire for creating delays and encouraging strategic maneuvering by both plaintiffs and defendants. 119Id. § 20.31; Francis E. McGovern, Rethinking Cooperation Among Judges in Mass Tort Litigation, 44 UCLA L. Rev. 1851, 1858 (1997).

This coordination may be vulnerable to the interpersonal dealings of the various judges in working together. 120But see Borden & Lee, supra note 106, at 1025 (finding that few MDL transferee judges perceived any reluctance among state judges to coordinate with the MDL proceeding); id. at 1027 (discussing a survey of state judges finding that they did not perceive relationships with federal judges to be problematic). The charisma and persuasive ability of the federal MDL judge to influence state colleagues may be a decisive factor in the success—leaving the success of coordination potentially vulnerable to variances in personality and experience. 121See Schwarzer et al., supra note 49, at 1736–37. It could collapse if either court decides to act unilaterally or independently in a way that harms administration of the cases in the other court. 122Manual for Complex Litigation (Fourth) § 20.311.

While few state or federal judges report problems due to interpersonal relationships, 123See supra note 120 and accompanying text. reluctance to cooperate does occur. Indeed, one such instance of state refusal to cede control of complex litigation to a federal MDL resulted in a dispute that ultimately required resolution by the U.S. Supreme Court in 2011 in Smith v. Bayer Corp. 124131 S. Ct. 2368 (2011). The MDL involved thousands of lawsuits over the drug Baycol. 125In re Baycol Prods. Litig., 218 F.R.D. 197, 201 (D. Minn. 2003). The federal judge “engaged in extensive efforts to coordinate its proceedings with state courts handling Baycol cases.” 126In re Baycol Prods. Litig., 593 F.3d 716, 720 (8th Cir. 2010), rev’d sub nom. Smith, 131 S. Ct. 2368. Despite these efforts, the federal court issued an injunction to force a state court to follow the MDL’s lead: when the federal court denied class certification to a putative class of West Virginia purchasers of Baycol in the MDL, 127Smith, 131 S. Ct. at 2374. it then issued an injunction to bar the state judge in West Virginia from independently considering a motion for certification of a similar class under state law in a parallel proceeding. 128Id. The Supreme Court unanimously held the Anti-Injunction Act prevented the federal judge overseeing the MDL from issuing such an injunction. 129Id. at 2375, 2382.

Smith demonstrated MDL courts have no formal power to coerce state judicial compliance in parallel proceedings. 130See Borden & Lee, supra note 106, at 998 (discussing the implications of Smith for interjurisdictional coordination). Cooperation must be welcomed by all jurists overseeing proceedings outside the MDL, or it simply does not work. When it works, such cooperation helps parties and judges engage in informal aggregation through coordination, 131Professor Erichson coined the term “informal aggregation” to refer to coordinated efforts that treat “the litigation as single, integrated whole,” while still maintaining separate lawsuits with independent claims. See Erichson, supra note 1, at 383. navigating through the federalism problem to achieve global settlements in mass tort actions—indeed, some experienced MDL judges believe that such settlements are impossible without coordination. 132See Schwarzer et al., supra note 49, at 1714 (discussing cases in which defendants would not have settled had they been forced to go to trial in state cases).

Despite the utility of coordination, the federalism problem persists, even with excellent coordination—it merely appears in a different form: multiple courts, and multiple judges, maintain control over different pieces of the litigation involved in the settlements. This means there is no single authority in charge of implementing, interpreting, and enforcing the settlement terms. Unless the state courts cede power to the central authority of the federal MDL, decision-making is structurally diffuse, leading right back to the federalism problem. As long as these separate court systems do not engage in centralized decision-making, ceding power to the MDL, the structural problem persists. 133Centralized decision-making has been identified as a line federal and state judges should avoid crossing. Borden & Lee, supra note 106, at 1019–20.

Parties understandably often want one judge overseeing the entire global settlement to offer a single, authoritative voice regarding the settlement agreement’s interpretation and implementation. For this reason, federal MDL judges face some pressure to facilitate settlement by asserting control over such settlements to bring about global peace. 134See, e.g., Schwarzer et al., supra note 49, at 1719–20 (observing that as early as 1980, in an MDL concerning fire in an MGM Hotel, the federal judge met with state plaintiffs to assess the value of their claims and facilitate a global settlement, because that was the only kind of settlement defendants would accept).

Part I.C demonstrates that some MDL courts have increasingly been answering this challenge by moving well beyond cooperative coordination. MDL courts have begun facilitating global settlements and administering common benefit funds in portions of “litigation” filed in state courts, and sometimes never filed in any court at all. In some situations, they become the global settlement facilitator over state and federal claims in mass tort litigation, without regard to the court system in which those claims were brought (or even whether claims were brought in court at all). They also become the global administrator over payments for common benefit work undertaken in any court system, assessing fees against settlements in any forum, and handing out money for work performed in any forum. The MDL judges in this context have become universal “settlement facilitators” overseeing transjurisdictional, multidimensional “litigation,” transcending the limits of individual lawsuits. 135Judith Resnik, From “Cases” to “Litigation,Law & Contemp. Probs., Summer 1991, at 50–52. In the context of global settlements, the trappings of voluntary cooperation appear to sometimes yield in favor of expedience and the need for global peace, particularly when those orders involve opt-in settlements in which the state parties want the MDL court’s oversight of the settlement.

C. Trailblazing New Paths to Global Peace Through Transjurisdictional Settlement Administration

An important series of twenty-first century MDLs reflect an emerging understanding of the federal courts’ power to influence parties, attorneys, and cases pending in state court, using mechanisms of both aggregation and disaggregation simultaneously. This new understanding has emerged almost exclusively in the context of global settlements (i.e., settlements that dispose not only of cases pending in an MDL, but also related, nonremovable cases pending around the country in state courts). It has made appearances in massive pharmaceutical and medical device products liability cases, involving thousands of plaintiffs suing individually in many different courts.

This new approach puts the MDL court in the role of a global settlement administrator, overseeing the resolution of claims filed in any court (state or federal), and sometimes even private claims tendered to the defendant without ever having been filed in any court. In other words, the federal court takes a measure of control over settlement of the mass tort itself, without regard to whether the federal court would have had jurisdiction in each individual claim subject to the settlement. Individual claimants opt in to the resulting settlement, accepting its structure and oversight by the MDL court. 136Such claimants have the choice whether to sign onto the settlement. They may instead decline to settle and continue to litigate individually. See, e.g., Fallon, supra note 3, at 378–79 (“The [MDL settlement] agreement is usually an opt-in agreement; when a claimant opts into the agreement, the claimant and the claimant’s primary counsel agree to be bound by the terms of the agreement, including the payment of common benefit fees.”).

This remarkable approach did not emerge fully formed from any single case. It has evolved iteratively, using innovations that have been expanded and built upon by experienced MDL judges learning from the experiences of colleagues. Understanding its current force requires a look back at a series of cases that have collectively reconceptualized the function of the MDL court at the settlement phase. Taken together, these decisions have implicitly redefined the “case” of the MDL—and at the same time created a new space that is at once powerfully aggregated and yet formally disaggregated.

1. In re Zyprexa’s Invention of the “Quasi–Class Action”

In 2006, District Judge Weinstein famously likened the “many individual related cases” within an MDL to a “quasi-class action.” 137In re Zyprexa Prods. Liab. Litig., 424 F. Supp. 2d 488, 490 (E.D.N.Y. 2006). He was referring to a diffuse collection of individual cases not qualifying for formal class treatment under Rule 23, yet still having some of the characteristics and classwide litigation. 138See id. at 491 (“While the settlement in the instant action is in the nature of a private agreement between individual plaintiffs and the defendant, it has many of the characteristics of a class action and may be properly characterized as a quasi-class action subject to general equitable powers of the court.”). He coined the term in relation to a global settlement negotiated in a major pharmaceutical products liability MDL in In re Zyprexa Products Liability Litigation. 139Id. at 490–91. The concept was the seed of the current revolution in MDL courts’ thinking about the scope of their power. Zyprexa itself, however, was fairly modest in its use of the MDL court’s settlement power.

In Zyprexa, Judge Weinstein capped legal fees for most settlements at 20% and further ordered that PSC common benefit work would be paid out of a general global settlement fund. 140Id. Claimants to the fund had to opt in by choosing to participate in the settlement. Though the global settlement fund paid claims in both state and federal court, the MDL court order limited the order’s application to “all settling actions in this multidistrict litigation.” 141Id. at 496 (emphasis added).

Judge Weinstein was active in coordinating with state court judges to encourage them to follow his lead, but the order itself imposed no obligation on parties outside of the MDL. He instead sent a letter to state court judges “suggesting coordination and cooperation.” 142Id. at 491. Despite the “quasi–class action” moniker, Zyprexa stayed well within the traditional boundaries of the federal courts and the conventional understanding of MDL as a mere procedural wrapper around individual suits that had been filed in federal court. The global settlement likely worked because it was attractive to individual plaintiffs who might otherwise bear the risk, delay, and cost of continuing to litigate alone.

The emergence of the quasi–class action analogy in MDL was thus a modest innovation in its original context. Although likening the role of the MDL court in settlement to the role of a court overseeing a certified class, the quasi–class action analogy was merely descriptive, rather than a justification to move off of the well-trodden path of cooperative coordination to solve the federalism problem using the aggregative toolbox of Rule 23. 143The analogy received the stamp of approval in a concurrence in a Second Circuit opinion. See In re Zyprexa Prods. Liab. Litig., 594 F.3d 113, 129–30 (2d. Cir. 2010) (Kaplan, J. concurring). Judge Kaplan agreed with Judge Weinstein about the “substantial similarities” between MDLs and class actions, and the need for common benefit funds of the type utilized by Judge Weinstein in the Zyprexa settlement. See id. This quickly changed in subsequent cases.

2. In re Guidant’s Extension of MDL Power over State Court Settlements

An unpublished order in 2008 in the In re Guidant Corp. Implantable Defibrillators Products Liability Litigation broke new ground with the quasi–class action analogy. 144MDL No. 05-1708 (DWF/AJB), 2008 WL 682174 (D. Minn. March 7, 2008). Relying on Zyprexa’s definition of the MDL as a quasi–class action for settlement purposes, 145Id. at *6. the court considered a complex settlement agreement in this products liability case, with a total settlement value of $240 million.

The case involved 8,550 plaintiffs who brought individual actions for injuries allegedly caused by defective defibrillators and pacemakers. 146Id. at *1, *10. The settlement agreement specified that the MDL court would decide the amount of common benefit payments to counsel who had engaged in work resulting in a common benefit. 147Id. at *4. The court understood the settlement covered not only claims in the MDL, but also state cases, and even claims that had not yet been filed in any court. 148See id. at *3.

The MDL court ordered that $10 million of the $240 million global settlement was to be set aside for common costs. 149Id. at *4. Additionally, over the objection of state court plaintiffs, 150Id. at *11. the MDL court ordered that $34.5 million be set aside for common-benefit attorneys’ fees. 151Id. at *16. The net result was that a portion of the settlement money from the state cases and unfiled, privately presented claims flowed into the federal MDL common benefit fund upon settlement. The court went so far as to find that the state plaintiffs benefitted from the work of the attorneys in the MDL in negotiating the settlement, even in “state court cases where a significant amount of work was done on behalf of individual Plaintiffs with little to no sharing of work product from the MDL attorneys.” 152Id. at *13.

The court ultimately ordered state court plaintiffs, and even claimants who had never filed a case in any court, to give up about 15% of their individual settlements to the common benefit fund in the MDL for distribution to the MDL’s PSC. The state court plaintiffs and other claimants had to choose either to participate in the global settlement (and pay into the MDL common benefit fund) or walk away from the settlement, and proceed alone. 153Id. In other words, by choosing to participate in the settlement, these state court plaintiffs were contractually agreeing to the authority of the federal court to redistribute part of their settlement for common benefit work. They were thus effectively opting into a partially aggregative device in the same moment they were exercising their right to settle their individual lawsuit.

Relying on Judge Weinstein’s order in Zyprexa, the MDL court in Guidant pushed the quasi–class action analogy much further than Judge Weinstein had. Like Zyprexa, Guidant presumptively capped attorney contingency fees in all cases that were part of the settlement; however, the cap applied whether the settled cases were filed in state or federal court. 154Id. at *18–19 (capping attorneys’ fees at 20%). The individual retainer agreements between attorneys and their clients in the individual cases in both state and federal court were generally nullified by this order. 155See id. at *19. The PSC’s work on behalf of the litigation trumped those individual retainer agreements.

Guidant is a path-making decision in its expansion of the MDL court’s power to encompass not only the federal litigants, but also anyone with a claim to be settled. The opt-in settlement agreement defined the scope of the MDL court’s power, under this approach. This solution was pragmatic in that it revealed a way to achieve global settlements and avoid the federalism problem through a settlement agreement. The MDL court emerged as a transjurisdictional global settlement administrator, overseeing both the global settlement and common benefit fund, with power over all the settling parties, including those who were never part of the MDL.

3. Pushing the Boundaries of the Quasi–Class Action to New Frontiers in the In re Vioxx Global Settlement Agreement

In re Vioxx Products Liability Litigation followed Guidant’s lead, blending the unilateral assertion of settlement authority in an MDL where there had also been state–federal coordination. 156See In re Vioxx Prods. Liab. Litig. (Vioxx I), 574 F. Supp. 2d 606 (E.D. La. 2008); Vioxx III, 802 F. Supp. 2d 740, 760 (E.D. La. 2011) (discussing a meeting between MDL judge, and judges from Texas, New Jersey, and California in New Orleans with representatives from Merck and the PSC); Transcript of Status Conference Before the Honorable Eldon E. Fallon, United States District Judge at 22–23, Vioxx I, 574 F. Supp. 2d 606 (MDL No. 1657-L), available at http://vioxx.laed.uscourts.gov/Transcripts/01182008trans.pdf (discussing the federal judge’s conversations with state judges in New Jersey and California concerning settlement administration). It also expanded the concept and offered a defense of the new reconceptualization of federal judicial power in MDLs. The global settlement agreement was again the crux of the exercise of this power.

The Vioxx settlement agreement was reached between the defendant Merck and a Negotiating Plaintiffs’ Committee (NPC) of attorneys for claims that the pharmaceutical drug caused strokes and heart attacks—in other words, it was not a settlement between any particular plaintiff (or even group of plaintiffs) and the defendant to settle any plaintiff’s specific claims, but rather an agreement among a select group of plaintiffs’ attorneys to settle the litigation. 157The NPC included “all counsel appointed to the Executive Committee of the Plaintiffs’ Steering Committee” in the MDL, and also representatives of coordinated state proceedings in New Jersey, California, and Texas. See Settlement Agreement Between Merck & Co., Inc. and the Counsel Listed on the Signature Pages Hereto 1 (Nov. 9, 2007) [hereinafter Vioxx Settlement Agreement], available at http://www.officialvioxxsettlement.com/documents/Master%20Settlement%20Agreement%20-%20new.pdf; see also Vioxx I, 574 F. Supp. at 609 (announcing the settlement agreement between NPC and Merck); Adam Liptak, In Vioxx Settlement, Testing a Legal Ideal: A Lawyer’s Loyalty, N.Y. Times, Jan. 22, 2008, A12.

It was a novel settlement agreement in the way it ensured that nearly all the plaintiffs who had claims (whether in the MDL or not) would choose to participate in the global settlement. It obligated plaintiffs’ attorneys to certify that all clients they represented had opted into the settlement; any plaintiffs who rejected the settlement would have to find other counsel to represent them. 158See Vioxx Settlement Agreement, supra note 157, at 2; see also Issacharoff, supra note 37, at 218 (discussing the Vioxx settlement terms). The plaintiffs thus retained the individual ability to accept or reject this settlement agreement, based on the autonomy of their separately filed actions. At the same time, their attorneys acted in concert to exert pressure to ensure maximum participation by the individual plaintiffs. Aggregative mechanisms to control the plaintiffs’ attorneys thus contrasted with the disaggregative autonomy of the individual plaintiffs.

While the novel mechanism for achieving maximum plaintiff opt-ins has received significant scholarly attention, 159See, e.g., Issacharoff, supra note 37, at 218 (discussing the innovative Vioxx settlement mechanism). the innovative role of the MDL court in overseeing settlement with nonfederal claimants has been largely overlooked by commentators. The global settlement agreement purported to resolve claims filed in state and federal courts, 160 Vioxx Settlement Agreement, supra note 157, at 1. which at that time included 26,000 active lawsuits, with 47,000 plaintiffs, and another 14,500 claimants who had sought compensation from Merck without filing suit. 161Id. The agreement appointed District Judge Fallon, the federal MDL judge, to preside over the global settlement (i.e., of both state and federal claims). 162Vioxx I, 574 F. Supp. 2d at 609 (“The Settlement Agreement expressly contemplates that this Court shall oversee various aspects of the administration of settlement proceedings . . . .”); see also Vioxx Settlement Agreement, supra note 157, § 6.1.1, at 29; Amendment to Settlement Agreement § 1.2.10 (Jan. 17, 2008), available at http://www.officialvioxxsettlement.com/documents/Amendments%20to%20Master%20Settlement%20Agreement.pdf. Crucially, the settlement agreement obligated each settling party to agree to the authority of District Judge Fallon to act in this capacity. 163Vioxx Settlement Agreement, supra note 157, §§ 8.1.1–.2, at 33. It went so far as to purport that this federal judge would “sit as a binding arbitration panel and whose decision shall be final, binding and Non-Appealable.” 164Id. § 8.1.2., at 33.

Despite the attempt to hang the trappings of private dispute resolution on the court’s oversight of the MDL, the settlement agreement invoked the full power of the federal court. 165See, e.g., id. § 10.1.2, at 37 (“The submission of fraudulent Claims will violate the criminal laws of the United States, and subject those responsible to criminal prosecution in the federal courts.”). The agreement authorized District Judge Fallon to create and administer a common benefit fund, created from a levy of 8% of the recovered amounts from every claimant who settled under the agreement. 166Id. § 9.2.1, at 35. The agreement provided that he would consult with (but not necessarily defer to) the state court judges who had presided over the cases not in the MDL. 167Id. § 9.2.3, at 36. His subsequent enforcement of the agreement made clear that he was acting in his capacity as a U.S. District Judge, exercising the power of a court, not the power of a private arbitration forum, by invoking the court’s “equitable authority.” 168See Vioxx I, 574 F. Supp. 2d 606, 612 (E.D. La. 2008) (discussing “this Court’s equitable authority over the global settlement” as well as its “supervisory authority” over attorneys’ fees).

In this role as transjurisdictional claims administrator, the Vioxx MDL court oversaw a $4.85 billion settlement fund. 169Id. at 609. As in Guidant and Zyprexa, the court asserted power to limit attorneys’ fees owed by settling plaintiffs by relying on the emerging quasi–class action analogy. 170Id. at 612 (“[T]he Court finds that the Vioxx global settlement may properly be analyzed as occurring in a quasi-class action, giving the Court equitable authority to review contingent fee contracts for reasonableness.”). District Judge Fallon understood the settlement agreement itself to contemplate this power. 171See Fallon, supra note 3, at 378 (discussing section 9.2 of the settlement agreement).

To his credit, District Judge Fallon recognized that the need for global peace in mass tort litigation pending simultaneously in state and federal courts would make settlements like this one increasingly common, 172Vioxx I, 574 F. Supp. 2d at 613. and he thus carefully explained the sources of authority he invoked. He pointed to three different sources of authority to support the decision to override the fee agreements of all settling plaintiffs, including those who were not part of the MDL: (1) the contractual terms of the settlement itself, 173Id. at 614 (“The Settlement Agreement expressly grants this Court the authority to oversee various aspects of the global settlement administration.”). (2) the court’s equitable authority over the administration of the global settlement, 174Id. at 611–12 (invoking the quasi–class action analogy). and (3) the court’s inherent authority to exercise ethical supervision over the parties. 175Id. at 612; see also In re Vioxx Prods. Liab. Litig. (Vioxx II), 650 F. Supp. 2d 549, 553–54 (E.D. La. 2009) (reconsidering and reaffirming the sources of authority discussed in Vioxx I). The discussion of each of these sources of power, however, implicitly assumed that this power extended automatically to all parties to the settlement. This implicit assumption reflects a fundamental shift in the nature of MDL.

Some of the settling parties never had cases in the MDL because their state court cases were not removable, or they had made private claims to Merck without ever filing a lawsuit in any forum. 176See generally Vioxx I, 574 F. Supp. 2d at 611–14. The scope of the MDL court’s order crystallizes when it is read in conjunction with the settlement agreement, which unambiguously included state and federal claims, as well as private claims tendered directly to the defendant, and its appointment of Judge Fallon to preside over the entire settlement. See discussion supra notes 160–61 (observing the settlement agreement’s inclusion of 14,500 private claimants who never filed suit and its resolution of claims filed in state and federal court); discussion supra note 162 (discussing the settlement agreement’s appointment of the MDL judge to preside over the global settlement); discussion supra note 163 (discussing the settlement agreement’s requirement that parties agree to the authority of the MDL judge to preside over the settlement). This means the only possible source of authority over them would have been their choice to opt into the settlement agreement itself. In this sense, the settlement became the federal controversy over which the MDL court was now presiding.

The Vioxx court ultimately capped fees for all claimants enrolled in the global settlement at 32%, without regard to whether those claims had ever been part of the MDL. 177Vioxx I, 574 F. Supp. 2d at 617. This nullified the retainer agreements of attorneys whose clients had agreed to higher contingent fees. 178See id. (acknowledging that 33%–40% was a standard contingent fee). From this 32% cap, additional sums for common benefit work would eventually be deducted. 179Id. at 617–18. Later, the court fixed the value of the common benefit work at over $315 million (6.5% of the $4.85 billion settlement). 180Vioxx III, 802 F. Supp. 2d 740, 758 (E.D. La. 2011).

Within 31 months, over $4.35 billion was distributed to 32,886 claimants, resulting in an “efficiency [that was] unprecedented in mass tort settlements of this size.” 181Id. at 762. Although District Judge Fallon credited the attorneys on both sides, the plan and lien administrators, the pro se curator, and the special masters for the efficiency, 182Id. the efficiency unquestionably also flowed directly from the innovative control over the global settlement exercised by the federal district judge, which allowed one voice to provide clear guidance on difficult questions regarding the settlement agreement’s interpretation and application. Over 32,000 separate claims were resolved by an aggregative settlement agreement that emphasized the disaggregated nature of the claims being settled. A hybrid had fully emerged—a third path in between disaggregated individual litigation and full aggregation under Rule 23.

Vioxx illustrates the possible benefits to parties when the federalism problem with regard to disaggregated litigation ceases to plague the dispute: defendants get global peace, plaintiffs obtain compensation for the harm they suffered, lawyers are rewarded for their work bringing the case to closure, and it all happens rather quickly in a public forum. These benefits arguably derived from the aggregate qualities of this settlement procedure in the MDL, qualities created through ad hoc mechanisms that are unbounded by Rule 23.

This efficiency was not without its detractors in the litigation. A small group of attorneys challenged the MDL court’s authority to enforce the fee cap. 183Vioxx II, 650 F. Supp. 2d 549, 551 (E.D. La. 2009). Their argument emphasized the sua sponte quality of the order, where no claimant had challenged the reasonableness of the fees. 184Id. at 556. These attorneys claimed there was no “Article III ‘case’ or ‘controversy’” and thus no subject matter jurisdiction over the issue. 185Id. Judge Fallon rejected the argument because the parties themselves invoked the power of the district court to preside over the settlement agreement when they opted into it, and he relied on Guidant and Zyprexa to justify the power of an MDL court to cap contingent fees even in the absence of a challenge to the reasonableness of those fees by claimants. 186Id. at 557–58.

The Vioxx settlement has been criticized on multiple grounds by scholars. For example, Professors Baker and Silver have argued that it interfered with the contractual relationship between attorneys and their clients. 187See Lynn A. Baker & Charles Silver, Fiduciaries and Fees: Preliminary Thoughts, 79 Fordham L. Rev. 1833, 1847–48 (2011). Professors Erichson and Zipursky have also argued that the enforcement of the settlement itself “pushed the envelope in legal ethics” by obligating plaintiffs’ attorneys to recommend it to clients. Howard M. Erichson & Benjamin C. Zipursky, Consent Versus Closure, 96 Cornell L. Rev. 265, 267 (2011). Professors Silver and Miller have critiqued the quasi–class action concept. 188Charles Silver & Geoffrey P. Miller, The Quasi–Class Action Method of Managing Multi-District Litigations: Problems and a Proposal, 63 Vand. L. Rev. 107, 109–11 (2010). Attorney Jeremy Grabill, a former law clerk to Judge Fallon, has added to the chorus critiquing the application of the “quasi–class action” idea in the Vioxx MDL, arguing it “added to the confusion that now exists concerning the proper role for courts to play.” Jeremy T. Grabill, Judicial Review of Private Mass Tort Settlements, 42 Seton Hall L. Rev. 123, 123, 178 (2012). Professor Issacharoff, however, has observed that whatever the force of these criticisms, the settlement was effective in helping plaintiffs to recover compensation and received an “overwhelmingly positive response” from the plaintiffs themselves. 189Issacharoff, supra note 37, at 219. He also lauds the manner that it used a “private arrangement to overcome the disfunctionality of the formal procedural system.” 190 Id.

Professor Issacharoff’s observation suggests that the Vioxx settlement agreement may be a public model that parallels defendant-created, postdispute, private claims resolution vehicles—with the important difference that this one was designed collaboratively by the court, the PSC, and the defendant and has some highly aggregative characteristics. 191See Dodge, supra note 4, at 1275–79 (describing postdispute disaggregative mechanisms); id. at 1306–07 (analyzing the dysfunctional features of defendant-designed claims resolution facilities). The Vioxx agreement also effectively solved concerns about fractional participation through controversial attorney “arm-twisting,” threatening to freeze out plaintiff-side attorneys who failed to get all their clients to participate in the settlement. 192See id. at 1310 (discussing problems of fractional enforcement).

Although Judge Fallon did not discuss the definition of the “case” before the MDL court with respect to the state court plaintiffs, or the claimants who never sued anywhere, the Vioxx decision reflects an implicit new understanding of the scope of the controversy over which the MDL court had authority: the authority in the settlement phase encompassed the entire mass tort dispute, extending to all claimants opting into a global settlement administered by the MDL court, regardless of where their individual lawsuits or claims originated (and enormous pressure may be brought to bear on attorneys to ensure all their clients opt in to maximize the number of claimants under the MDL court’s power).

This understanding stands in stark contrast to the narrower understanding of the MDL court’s power in earlier cases. MDL is no longer merely a procedural wrapper around a collection of individual lawsuits filed in or removed to federal court. The twenty-first century MDL court’s power becomes transjurisdictional by asserting control over a global settlement fund created by a private mass settlement agreement. The boundary of the MDL court’s power became defined by the mass settlement, not just the individual claims that had once comprised the MDL, and yet it still purported to preserve the autonomy of those individual claimants by requiring the claimant’s signature on the settlement agreement to trigger the global power.

Despite the aggregative qualities of the settlement, the Vioxx litigation also reflects the disaggregative potential of MDL. The MDL court conducted six individual bellwether trials—trials of separate cases, each decided by a different jury. 193Fallon et al., supra note 58, at 2335. The defendants won all but one of them in federal court. 194Id. The one federal victory for the plaintiff then resulted in a settlement with that plaintiff while it was on appeal. 195Id. at 2336. Additionally, approximately thirteen related cases were tried in state court in six different states around the same time period. 196Id. at 2335. While these were individual trials that had no binding effect on the rest of the claims in the MDL, the results nevertheless spurred resolution for everyone else who ultimately participated in the global settlement. 197See id. at 2338 (“[B]ellwether trials can precipitate and inform settlement negotiations by indicating future trends, that is, by providing guidance on how similar claims may fare before subsequent juries.”). Despite the defendant’s success in trying the cases, Merck nevertheless then entered global settlement negotiations and agreed to fund a $4.85 billion global claims settlement facility. 198See id. at 2337.

4. Expanding MDL Power in the Deepwater Horizon Litigation in the Absence of a Global Settlement Agreement

The Vioxx innovations regarding the scope of the MDL court’s power found immediate application in the Deepwater Horizon litigation, based on the largest oil spill in U.S. history. 199See Oil Spill I, MDL No. 2179, 2011 WL 6817982, at *1, *2 n.1 (E.D. La. Dec. 28, 2011); Vernon Valentine Palmer, The Great Spill in the Gulf . . . and a Sea of Pure Economic Loss: Reflections on the Boundaries of Civil Liability, 116 Penn St. L. Rev. 105, 105 (2011) (discussing the magnitude of the spill). As in other MDLs, District Judge Barbier set out to establish a common benefit fund by assessing 4%–6% of the gross amount of all individual settlements, to be paid into a fund to compensate attorneys engaged in common benefit work. 200Oil Spill I, 2011 WL 6817982, at *1. He initially ordered that the assessment applied not only to cases in the MDL that settled, but also some state court cases, and even claims paid through the GCCF, 201Id. at *6 (applying the order to state cases where plaintiffs were represented by attorneys who also had cases in the MDL and had access to MDL work product); see also Dodge, supra note 4, at 1255–56 (discussing the GCCF). a unique, nonjudicial, private forum for compensation of claims, established by BP after the Gulf oil spill disaster. 202See Issacharoff & Rave, supra note 8, at 398, 400.

After strenuous objections from claimants who never filed any lawsuits and instead were paid only through the private GCCF, the court eventually modified this order and exempted GCCF claimants from the assessment. 203See Oil Spill II, MDL No. 2179, 2012 WL 161194, at *2 (E.D. La. Jan. 18, 2012). The order offered no reason for the modification exempting GCCF settlements. The assessment remained in effect, however, against state court plaintiffs represented by counsel who participated in the MDL or had access to MDL discovery. 204Id.

Unlike Vioxx, in the Deepwater Horizon litigation, none of the state court parties opted in to any global settlement giving the court supervisory authority over fee allocations. These parties filed suits in state courts, litigated entirely in state courts, and settled in state courts. Or they participated in a private dispute resolution proceeding. They were never part of the MDL. Nevertheless, they found themselves bound by a federal court order instructing the defendant not to pay them the full amount of their settlement agreements they had executed because 4%–6% of their gross settlement was owed to the federal court for the common benefit work by the MDL’s PSC that purportedly was useful to settling plaintiffs outside the MDL, without regard to whether those plaintiffs made individual use of that work.

The series of cases from Guidant through the Deepwater Horizon litigation shows a pronounced evolution in the approach of these MDL courts to global settlements. Each new case builds on and expands upon the power of the MDL court over parties outside the MDL. These courts’ understanding of the controversy over which they have power now encompasses parties outside the MDL, at least in the context of global settlements and common benefit fundsand, as the Deepwater Horizon litigation shows, may now even be creeping toward claimants who never consented to the MDL court’s power over their individual settlements. In this way, the new breed of MDL is highly aggregative (sweeping masses of claims into a global settlement, even claims not pending in the MDL itself), and yet highly disaggregative (effecting individual settlements on a massive scale).

II. The Persistence of the Federalism Problem in Multidistrict Litigation

Given the demise of mass tort class actions, Professor Burch has aptly called MDLs “the new mass-tort frontier.” 205Burch, supra note 101, at 88. The shape of that frontier appears to be changing rapidly.

Part I demonstrated that a new form of transjurisdictional, aggregated, global settlement of mass torts is emerging in MDL, despite the demise of Rule 23 class actions. Each iteration of MDL innovation has yielded a subsequent expansion by a later court: the quasi–class action concept initially relied on coordinated orders from state courts to enforce the MDL court’s decisions; then MDL courts independently asserted power over state litigants, but only when they opted into a global settlement; finally, federal power was exercised over claimants who never had a case before the federal court, and who had not even opted into a global settlement. The modern MDL judicial culture is admirably efficient and yet also self-perpetuating. These cases cite each other, relying on the groundwork laid by each prior application of the quasi–class action analogy to extend it to new frontiers.

This Part argues that the emerging transjurisdictional MDL settlement trend has not solved the federalism problemit has merely submerged it. Part II.A demonstrates that, despite the pressure to aggregate, a small but growing number of MDL judges are declining to exercise mass settlement power over claims not in the MDL because of the persistence of the federalism problem. Part II.B argues that the new conception of transjurisdictional settlement authority in MDL ultimately succeeds in forging global peace only because of state court acquiescence. The exercise of transjurisdictional settlement authority thus is grounded in a view of federal power that has functioned successfully only because of a high level of deference by other cooperating courts, attorneys, and parties. The authority has no intrinsic structural foundation in the MDL device itself.

A. The Narrow Conception of MDL Power: Locating the Federalism Problem in MDL’s Narrow Statutory Definition

Despite the utility of this evolution of MDL power in administering global settlements, there has been a persistent stream of resistance. A number of MDL courts have refused to issue orders binding parties not in the MDL. The resistance stems from the persistence of the federalism problem.

Some MDL courts have refused to follow the expansive quasi–class action, transjurisdictional, aggregate settlement model. They have persisted in applying the older, narrow understanding of the MDL court’s power over parties. These courts adhere to the view of the Fourth Circuit in Showa Denko. 206In re Showa Denko K.K. L-Tryptophan Products Liability Litigation–II, 953 F.2d 162, 165–67 (4th Cir. 1992). Under this alternative model, the MDL courts have declined to seize control of the settlement of cases not in the MDL (i.e., cases pending in other court systems or involving parties making nonjudicial claims). 207See, e.g., In re Fedex Ground Package Sys., Inc., Emp’t Practices Litig., No. MDL-1700, Cause No. 3:05-MD-527-RM, 2011 WL 611883, at *3 (N.D. Ind. Feb. 11, 2011) (finding the court lacked jurisdiction to order a hold-back in cases never part of the MDL); In re Genetically Modified Rice Litig., No. 4:06 MD 1811 CDP, 2010 WL 716190, at *1 (E.D. Mo. Feb. 24, 2010) (finding the court lacked jurisdiction to order hold-backs in state court cases, even though plaintiffs in state court cases had derived substantial benefit from the MDL’s common benefit work); In re Baycol Prods. Litig., No. MDL 1431MJD/JGL, 2004 WL 1058105, at *3–4 (D. Minn. May 3, 2004) (declining to assess a hold-back against settlements by parties not participating in the MDL, and finding the court lacked both original and supplemental jurisdiction over such parties); In re Linerboard Antitrust Litig., 292 F. Supp. 2d 644, 663–64 (E.D. Pa. 2003) (finding the court lacked jurisdiction to assess a hold-back in state court cases never removed to federal court or transferred to the MDL, but enforcing its hold-back order against state court cases that had been remanded after improper removal to federal court in the MDL).

This alternative, disaggregated model of MDL views coordination with state courts as the only possible solution to the federalism problem. Indeed, these cases tend to suggest rather hopefully that state judges handling the state cases may order participation in the MDL’s common benefit fund, where those state plaintiffs received a substantial benefit from that work. 208In re Genetically Modified Rice Litig., 2010 WL 716190, at *1, *4–5. They avoid, however, taking control of an amorphous “litigation” in its entirety, or even of a global settlement fund, based on the understanding that they lack jurisdiction over parties and issues not in the MDL.

A split among MDL judges has thus emerged. Other than the admittedly dated Hartland/Showa Denko line of authority, the split has generally not been reflected in appellate decisions, as it is a product of settlements that evade appellate review. The split is important, however, because it goes to the fundamental scope of the MDL court’s power to aggregate related claims for settlement purposes.

Reconceptualizing the “litigation” to encompass a global settlement involving thousands of claimants who sued in state and federal courts all over the country is a marked departure from the traditional understanding of MDL as a procedural vessel into which the JPML pours individual cases (and class actions), each with its own separate identity as a litigation unit. The traditional understanding still has a strong following among federal judges—for good reason.

The MDL statute itself lends support to the traditional understanding. The statute contemplates the transfer of civil actions (i.e., lawsuits already filed) into the MDL, and the eventual remand of such actions back to their home districts. 20928 U.S.C. § 1407(a) (2012). Under the statute’s language, the MDL court’s power over the dispute encompasses only the right to exercise “coordinated or consolidated pretrial proceedings” over the civil actions transferred into the MDL. 210Id. Nonremovable cases filed in any state court and claims brought privately do not fall within § 1407’s language. The existence of a privately negotiated global settlement does not alter that limited grant of statutory authority.

The exercise of transjurisdictional settlement authority cannot be traced to Congress’s grant of authority to federal judges overseeing consolidated MDL proceedings. In the case of opt-in global settlements, the transjurisdictional settlement practice often seems to envision the parties to the settlement as creating private authority for aggregated resolution, by opting into the terms of the settlement appointing the federal MDL judge as settlement administrator. Part II.B demonstrates that this private authority does not resolve the federalism problem inherent in nonclass aggregated settlements.

B. The Broad Conception of MDL Power: Locating the Federalism Problem Lurking in Mass Settlements

Part II.A showed that, though MDL is sometimes classified as a kind of formal aggregation, 211See e.g., Dodge, supra note 4, at 1256. the MDL statute created MDL as a mere procedural vessel into which the JPML transfers individual lawsuits for consolidated pretrial proceedings. The new breed of MDL challenges this understanding by conceiving the MDL court’s power to extend to the entire litigation, at least at the settlement phase. Part II.B argues that the emergence of transjurisdictional MDL settlement does not solve the federalism problem inherent in federal complex litigation. It merely submerges it through private agreements.

As an initial matter, it is worth noting that mass settlements of some sort have nearly always been a defining feature of MDL. Mass tort disputes that survive dismissal march inexorably toward settlement—nearly always on a collective basis. 212Fallon et al., supra note 58, at 2340 (“As in traditional tort litigation, the endgame for a mass tort dispute is not trial but settlement . . . . [and] the most ambitious settlements seek to make and enforce a grand, all-encompassing peace in the subject area of the litigation as a whole.” (alterations in original) (quoting Richard A. Nagareda, Mass Torts in a World of Settlement, at ix (2007)) (internal quotation marks omitted)). This flows as much from the parties as the courts. 213“The business of mass litigation dictates a collective approach,” according to Professor Erichson, who observes that plaintiffs’ attorneys often seek collective resolution because they represent large inventories of claims and plaintiffs, whereas defendants desire inclusive resolution. Erichson, supra note 54, at 1773–75.

In multijurisdictional mass tort litigation, MDL courts must find a path to aggregate settlement, dispose of the dispute piecemeal, or confront the need for remanding thousands (or tens of thousands) of cases for individual trials. 214Grabill, supra note 188, at 126 (“[C]ourts have struggled in applying established principles concerning the scope of judicial authority to evaluate and oversee the implementation of traditional settlements in the unfamiliar context of private mass tort settlements.”). The exercise of broad settlement authority is arguably a matter of self-preservation for the federal court system: if meritorious claims do not settle, the system lacks the capacity to try all of the federally filed claims individually. Settlement is thus the goal. 215See supra note 212 and accompanying text. The only question is whether the settlement will occur piecemeal or in some collective form. In contrast to the trend of defendant preferences for disaggregated settlements in private sector alternative dispute resolution forums and claim facilities, 216See Dodge, supra note 4, at 1255–56 public litigation often reflects a strong defendant preference for collective settlements. 217See supra note 35 and accompanying text.

In the post–class action era of mass torts, MDL courts are struggling to find a path to global peace without the benefit of Rule 23. 218Grabill, supra note 188, at 126. Unlike Rule 23 mass settlements that bind all class members not opting out, MDL mass settlements aggregate by seeking the parties’ express agreement: persons with claims sign onto the settlement agreement. In the emerging transjurisdictional type of MDL settlement, individuals may participate in the settlement regardless of whether they brought a case in the MDL. Their decision to participate in the mass settlement overseen by the MDL court becomes a substitute for class certification. The settlement agreement thereby uses the claimants’ consent as a method of accomplishing a result akin to what nationwide class action settlements once did, through the parties’ own choice. While there are no absent class members, there may be tens of thousands of claimants who opt in and seek compensation from the MDL claims facility, hundreds of lawyers seeking payment for common benefit work by applying to the MDL court, and a complex settlement administration infrastructure supervised by the MDL court—all reminiscent of class settlements.

Despite the similarities, the formal procedures that enable the collective melding of the litigation unit under Rule 23 are missing in MDL. The MDL statute itself lacks mechanisms to ensure cohesiveness and the procedural ordering that defines the class action. The lack of cohesiveness in the MDL creates an inherent instability to any attempted conceptual unity of the litigation unit. The separate controversies continue to have separate identities, though these identities are temporarily suppressed by the aggregative device during pretrial proceedings. The MDL judges invoking the quasi–class action concept for settlement purposes are nevertheless correctly observing MDLs behaving like class actions at the settlement stage. The much heralded death of aggregation in Rule 23 thus merely relocated the aggregative pressure from the certification stage to the settlement stage in federal litigation.

The utility, efficiency, and effectiveness of the new transjurisdictional MDL settlement mechanisms manifest in the scale of the settlements the MDL courts achieve and oversee—billions of dollars, and tens of thousands of plaintiffs’ claims settled, often in a single “litigation.” Instead of solving the federalism problem, these innovations merely shift attention from it, while leaving the problem lurking just out of sight: these mass settlements avoid the federalism problem only because the participating parties, the MDL court, and the state courts with parallel proceedings affected by these settlements privately agree not to notice the federalism problem. In other words, they jointly agree to disregard the federalism problem in order to settle the dispute.

Setting aside concerns about subject matter jurisdiction over settling parties who have nonremovable claims that could not have been brought in federal court, or that were never filed in any court (which are beyond the scope of this Response), the system requires complicity by all of the actors in it: federal and state court judges, settling plaintiffs and defendants, and counsel for both sides. Without such complicity, global settlements cannot function. In the absence of congressional authority to exert transjurisdictional authority beyond the federal cases transferred to the MDL, the MDL court’s transjurisdictional power only exists by the agreement of all who are affected. As long as all participants march ahead toward a resolution, there is no one to challenge the MDL court’s exercise of power over non-MDL parties—these non-MDL parties themselves want to participate in the settlement.

The difficulty and expense of negotiating a mass settlement likely means that once one exists, it may exert a powerful gravitational force over parties, attorneys, and judges, pulling them all inexorably toward aggregate resolution. However, imagine for a moment that any one of these constituencies were to balk at the MDL court’s administration of any aspect of the settlement and claims facility. Were a state court litigant to contest the settlement agreement’s implementation and convince the state judge with jurisdiction over that case to order a result different than the one ordered by the MDL judge, the charade would be over. Smith v. Bayer Corp. makes clear that MDL courts cannot enjoin parallel proceedings in state court in complex litigation. 219131 S. Ct. 2368, 2375–76 (2011). The MDL court thus likely could not prevent state interference with the settlement administration, at least as to parties who filed in state court.

This risk can be illustrated by a counterfactual example from the Vioxx case. If a state judge with power over cases participating in the Vioxx litigation had disagreed with the MDL order capping attorneys’ fees in all the settled cases, and if that state court had instead ordered state plaintiffs to honor their retainer agreements by paying the full amount of fees promised upon settlement, conflicting orders would have resulted. Plaintiffs and their attorneys would be under conflicting obligations with regard to the same settlement. The unified nature of the settlement oversight would then collapse.

The MDL transjurisdictional settlement agreements succeed when state judges allow the MDL court to take the reins of the litigation and the global settlement, at the parties’ request. This is a form of judge-shopping, though the courts do not identify it as such. Settlement oversight requires significant time and court resources likely to be lacking in state courts—as well as familiarity with the case, parties, attorneys, and issues. 220See Fallon et al., supra note 58, at 2340–41 (explaining the role of the MDL court in presiding over global settlement negotiations by virtue of the “centralized forum”). Moreover, experienced MDL judges who developed collaborative relationships with state court judges during earlier stages of litigation prior to the settlement (as was the case in the Vioxx litigation) are unlikely to suffer state judicial rebellions in the settlement phase—the state judges are likely to be well aware how difficult and time-consuming the mass settlement was to accomplish and have no incentive to sink it. 221Kenneth R. Feinberg, Reexamining the Arguments in Owen M. Fiss, Against Settlement, 78 Fordham L. Rev. 1171, 1173 (2009) (“What are we going to do with these cases in the absence of efficiency and in the absence of political will on the part of the other branches of government? What’s the reality of the situation?”). Injured parties, after all, benefit from efficient compensation.

Despite the MDL courts’ broad assertion of power, transjurisdictional settlements in MDL are a very fragile form of aggregation. They ultimately still rely on state court acquiescence, despite the MDL courts’ overt appearance of aggressively asserting federal power. It works only because the state court judges let it work—and the incentives align to minimize the likelihood of nonacquiescence.

The federalism problem thus persists in MDL at the settlement phase, lurking beneath the surface of these mass settlements. The settlements function through an implicit delegation of power—a silent judicial agreement to let the settling parties choose the federal MDL court to oversee the enormously complex details of the claims facility. This sort of radical “super-aggregation” thus turns out to be not so different at its core from the traditional judicial coordination. The transjurisdictional settlement oversight, in this way, derives authority from the implicit assent of the state courts that defer to the federal court in the settlement administration, as well as the parties opting into the settlement. These MDL courts appear to assert global settlement power because no one else wants it, and the parties feel someone needs to have it to effectuate settlement.

In a sense, transjurisdictional global settlements are a uniquely postmodern feature of mass tort litigation. They settle claims in all jurisdictions at once without all of the settling cases being fully under the power of any one jurisdiction to adjudicate. They exist in a liminal space that is neither a class action nor an individual action, administering settlements without actually having statutory power over all the parties, using consent as a proxy for formal procedural devices, and tacit state judicial deference as an alternative to congressional authority.

Conclusion

Over two decades ago, Judith Resnik aptly observed that in complex litigation, cases were “los[ing] their boundaries and becom[ing] part of a ‘litigation.’” 222Resnik, supra note 135, at 5, 50. Such “litigation” is amorphous and potentially spans multiple jurisdictions, with vast numbers of parties in mass tort cases. The systemic pressure to efficiently streamline thousands of related cases resulted in increasing “collectivization” of disputes, transforming them from individual claims into aggregated mass tort “litigation.” 223Id. at 50–51. This melding of cases and blurring of the dispute’s boundaries has long been visible in the trend to centralize mass torts, to the extent possible, before a single judge in federal court in order to craft national solutions to “litigation” rather than relying on local solutions to remedy individual suits. 224See id. at 55 (“Federalization reflects the growing perception that problems raised in many litigations are about harms suffered by people all over the country.”). MDL settlements have become the principle locus of this blurring of case boundaries in the current era.

The vanishing utility of Rule 23 in mass torts is creating pressure upon MDL courts to oversee transjurisdictional, global settlements. Despite disaggregative trends in the private sector, MDL reflects pressure in the public courts to aggregate in the settlement phase, even as the formal authority to aggregate under Rule 23 has become increasingly unavailable.

The goal of this Response has not been to situate transjurisdictional settlements in the boundaries of federal subject matter jurisdiction or resolve problems relating to the constitutional limits on federal power over state parties. Rather, the goal has been to identify the transjurisdictional MDL trend that has gone largely unnoticed by scholars and appellate courts, and connect it to the federalism problem in complex litigation. It responds to Professor Dodge’s mapping of a new branch of disaggregation by juxtaposing that branch to aggregative forces still at work within the federal courts.

The emergence of the transjurisdictional MDL settlement trend illustrates that “aggregation” and “disaggregation” are not monolithic concepts in public litigation. Nor are they even poles on a continuum, where cases might be more or less aggregated. Rather, they are features of case management that can be located simultaneously within the same litigation—and magnified or diminished in ad hoc ways as needed to drive a case inexorably toward settlement.

The vexing federalism problem nevertheless persists in ad hoc aggregate settlement mechanisms in MDL. Transjurisdictional global settlements can submerge the federalism problem beneath the overt agreement of the settlement parties. Such agreements allow the federalism problem to drop out of view, but they do not solve it.

Footnotes

Assistant Professor of Law, Louisiana State University, Paul M. Hebert Law Center. I thank the LSU Law Center for providing research support for this project. I am grateful for insightful critiques of an earlier draft of this Article from Robert Bone, Edward Cantu, John Devlin, Gordon Hylton, Liz McCuskey, Allan Stein, and the participants in the 2013 Junior Faculty Federal Courts Workshop at Brooklyn Law School and the 2013 Marquette Law School Junior Faculty Workshop. I thank Jaime Dodge and the editors of the Emory Law Journal for inviting my participation in this colloquy.I also wish to acknowledge the helpful contribution of my research assistants: Jacob Gower, Catherine Sens, Brian Lindsey, and Charmaine Borne, and the past LSU Law Center students in my Federal Complex Litigation course whose fruitful discussion of Vioxx provided the inspiration for this project. Any remaining errors are entirely my own.

1Howard M. Erichson, Informal Aggregation: Procedural and Ethical Implications of Coordination Among Counsel in Related Lawsuits, 50 Duke L.J. 381, 469 (2000).

2See infra Part I.

3For clarity, this Response refers to the federal court presiding over the MDL as the “MDL court.” Some federal judges use the term “MDL transferee court” to refer to this same court (as cases are transferred to this court from somewhere else). See, e.g., Eldon E. Fallon, Common Benefit Fees in Multidistrict Litigation, 74 La. L. Rev. 371, 372–73 (2014) (using the term “MDL transferee court”). This Response treats these terms as synonymous.

4See Jaime Dodge, Disaggregative Mechanisms: Mass Claims Resolution Without Class Actions, 63 Emory L.J. 1253 (2014).

5Id. at 1276–77.

6See id. at 1257–58.

7See id. at 1262 (“[W]hile parties elect private disaggregative mechanisms because they yield better outcomes than the default litigation system, in many disputes a public disaggregative mechanism would offer a superior option.”).

8See generally Samuel Issacharoff & D. Theodore Rave, The BP Oil Spill Settlement and the Paradox of Public Litigation, 74 La. L. Rev. 397 (2014) (comparing the Gulf Coast Claims Facility to the class action settlement that replaced it, finding that the class action results in greater payments to claimants).

9Id. at 400 (“Ambitiously, the GCCF set out to expeditiously resolve all of the private oil spill related claims against BP outside of the court system.”).

10Id. at 398 (“This [GCCF] settlement . . . at least in theory, should have been the best of all worlds.”).

11See discussion infra note 24.

12Issacharoff & Rave, supra note 8, at 400.

13Id.

14Id. at 400–01.

15Id. at 401.

16See id.; see also Dodge, supra note 4, at 1312 (discussing class settlement).

17In re Deepwater Horizon, 744 F.3d 370, 373–74 (5th Cir. 2014) (affirming the district court’s interpretation of the settlement agreement, and recounting the byzantine procedural history). Interpretation of the BP settlement was first appealed to the Fifth Circuit in 2013, remanded with instructions for the district court to reconsider certain data in calculating “Business and Economic Loss,” then returned to the Fifth Circuit in a second appeal. Id. While the interpretation of the settlement agreement was pending before one panel of the Fifth Circuit, a different panel of the same court also considered a separate appeal of the certification of the class by the district court. Id. at 374; see also In re Deepwater Horizon, 739 F.3d 790, 795 (5th Cir. 2014) (affirming class certification).

18See Issacharoff & Rave, supra note 8, at 402 (observing the BP class action settlements included a $600 million reserve fund to cover fees to private counsel, costs of class notice, and discovery in the MDL).

19See discussion supra note 17 (recounting the litigation regarding the interpretation of the settlement agreement).

20Issacharoff & Rave, supra note 8, at 402, 406 fig.1 (depicting a comparison of claim valuations in the GCCF and class settlement); id. at 413 (“We have been unable to find any significant category of recovery in which claimants did better under the GCCF than under the scheduled payments of the class settlement.”).

21See id. at 402.

22Id. at 403.

23See id. at 402.

24The ultimate success of the court-administered settlement is not yet final, though the Fifth Circuit upheld the district court’s interpretation of the settlement agreement. At the time of this writing, BP has sought en banc review, and may yet seek review by the U.S. Supreme Court. On March 17, 2014, BP filed a petition to have the entire Fifth Circuit review the terms of the settlement agreement en banc, after losing before a three-judge panel of that court. See Petition for Rehearing En Banc for Appellants BP Exploration & Production Inc., BP America Production Co., and BP P.L.C. at 5–6, In re Deepwater Horizon, 744 F.3d 370 (5th Cir. 2014) (No. 13-30315), available at https://www.thestateofthegulf.com/media/67350/2014-03-17-Petition-for-Rehearing-En-Banc-for-Appellants-BP-Exploration-.pdf.

25See Dodge, supra note 4, at 1300–01, 1305.

26See Francis E. McGovern, Settlement of Mass Torts in a Federal System, 36 Wake Forest L. Rev. 871, 888 (2001) (defining elasticity in mass torts to refer, at least in part, to “the undefined nature and number of plaintiffs”). Elasticity creates special difficulties in resolving mass tort litigation. Professor McGovern has aptly observed that claims resolution facilities that are especially efficient tend to then result in additional claims being filed:In litigation terms, if the supply is the number of cases processed by the system and cost is the transaction cost of that processing, then an elastic mass tort would have an increase in filings, whereas an inelastic mass tort would not. Aircraft crash cases are, for example, inelastic; asbestos cases are highly elastic. Since only 10–20% of all actionable torts result in litigation, there is a universe that remains unfiled. If there is a claims resolution facility that processes claims quickly and at low costs, one can anticipate a much higher filing rate than one would otherwise expect in the tort system. These larger numbers of claims can create a major dilution of benefits, particularly if they are accompanied by large numbers of false positives that cannot be eliminated.Francis E. McGovern, The What and Why of Claims Resolution Facilities, 57 Stan. L. Rev. 1361, 1383 n.66 (2005) [hereinafter McGovern, The What and Why].

27Dodge, supra note 4, at 1316.

28See id. at 1257 & n.14 (discussing scholarship focusing on smaller class actions and disaggregation).

29It may be useful to clarify what is meant by “aggregation” and “disaggregation” for purposes of this article. True, formal aggregation traditionally involved the joinder of parties or certification of class actions. See, e.g., Erichson, supra note 1, at 409. There are, however, myriad other ways suits can be partially aggregated for group resolution, whether formally in court or informally through coordination. See generally Erichson, supra note 1. This contrasts with disaggregation, which emphasizes the separateness of parties, claims, defenses, and remedies. See Dodge, supra note 4, at 1257.

30See In re Oil Spill by the Oil Rig “Deepwater Horizon” in the Gulf of Mex. on Apr. 20, 2010 (Oil Spill I), MDL No. 2179, 2011 WL 6817982, at *3, *5 (E.D. La. Dec. 28, 2011); see also Dodge, supra note 4, at 1256 & n.8 (discussing the GCCF).

31See In re Deepwater Horizon, 739 F.3d 790, 796 (5th Cir. 2014) (noting that BP worked with the PSC “to transfer claims from the GCCF to a program supervised directly by the district court”); In re Deepwater Horizon, 732 F.3d 326, 329 (5th Cir. 2013) (“In March 2012, the district court granted the parties’ request to implement a process to transfer claims from the GCCF to a court-supervised program that the parties agreed to in principle.”); Oil Spill I, 2011 WL 6817982, at *1, *6; Issacharoff & Rave, supra note 8, at 400–02.

32See Oil Spill I, 2011 WL 6817982, at *1, *3, *5. The court later modified this to exempt GCCF claimants but not state court claimants. See In re Oil Spill by the Oil Rig “Deepwater Horizon” in the Gulf of Mex. on Apr. 20, 2010 (Oil Spill II), MDL No. 2179, 2012 WL 161194, at *2 (E.D. La. Jan. 18, 2012).

33See, e.g., Oil Spill II, 2012 WL 161194, at *1 (ordering defendants to withhold 4% of gross settlements, judgments, or other payments to government entities to deposit into a common benefit fund); id. at *2 (ordering defendants to withhold 6% of same for any other plaintiff or claimant); Issacharoff & Rave, supra note 8, at 413.

34See Issacharoff & Rave, supra note 8, at 404–12.

35Id. at 413. They argue there are thresholds of settlement participation that represent enhanced value to defendants, even when the participation obtained is not total. Id. at 415–16.

36The controversial aggregate settlement of personal injury claims belonging to anyone exposed to Agent Orange in Vietnam was one of the earliest cases to demonstrate the power of Rule 23 to impose global peace. See In re “Agent Orange” Prod. Liab. Litig., 818 F.2d 145, 169 (2d Cir. 1987) (affirming the adequacy of notice to “[a]nyone who believed that he or she had suffered injury as a result of exposure to Agent Orange in Vietnam”); see also Anne Bloom, From Justice to Global Peace: A (Brief) Genealogy of the Class Action Crisis, 39 Loy. L.A. L. Rev. 719, 731–32 (2006) (reviewing the terms of the Agent Orange class action settlement); id. at 734 (observing the impact the Agent Orange class action had on reshaping mass tort litigation); id. at 735 (reporting that, after Agent Orange, federal courts began approving aggregate settlements in other mass tort contexts). Rule 23 was also used to craft a global settlement in the Dalkon Shield product liability litigation, another early example of aggregate resolution of personal injury claims. In re A.H. Robins Co., 880 F.2d 709 (4th Cir. 1989); Bloom, supra, at 735. The floodgates then opened and other cases followed. See Bloom, supra, at 735–36. Indeed, during this period, class treatment of settlements was sometimes prompted by defendants seeking global peace. Id. at 746–47.

37After the Court’s decisions in Amchem Prods., Inc. v. Windsor, 521 U.S. 591 (1997), and Ortiz v. Fibreboard Corp., 527 U.S. 815 (1999), attempts to use the class action device under Rule 23 to aggregate nationwide mass tort claims to facilitate global settlements have been increasingly doomed because federal courts often decline to certify such classes. See Bloom, supra note 36, at 747; Samuel Issacharoff, Private Claims, Aggregate Rights, 2008 Sup. Ct. Rev. 183, 208 (“As a result, class actions seemed to drop out of the available set of tools for attempting to settle most mass torts . . . .”); see also Jeremy Hays, The Quasi–Class Action Model for Limiting Attorneys’ Fees in Multidistrict Litigation, 67 N.Y.U. Ann. Surv. Am. L. 589, 601 (2012) (discussing the waning utility of class actions under Rule 23 and the imperfect nature of MDL as a replacement). This demise of class actions was long heralded by scholars. See Charles Silver, Comparing Class Actions and Consolidations, 10 Rev. Litig. 495, 500 (1991) (observing in the early 1990s that it was already true that narrow interpretation of class action rules made it difficult for district judges “to craft tort class actions that survive review”).

38Fallon, supra note 3, at 372–73 (observing the increasing significance of MDL).

39Id. at 373 (citing John G. Heyburn II & Francis E. McGovern, Evaluating and Improving the MDL Process, Litigation, Spring 2012, at 26).

40See 28 U.S.C. § 1407(a) (2012).

41See McGovern, The What and Why, supra note 26, at 1377 (discussing consent of the settlement beneficiaries as a vehicle for achieving legitimacy of the claims resolution facility).

42Cf. Manual for Complex Litigation (Fourth) § 20.31 (2004) (discussing federal jurisdictional boundaries in MDL litigation).

43Erichson, supra note 1, at 415 (“[A] state court action cannot be consolidated with a federal court action unless the state court action is first removed to federal court, which in many cases cannot be accomplished. Due to these restrictions, consolidation has limited utility as a method of aggregating dispersed cases.” (footnote omitted)).

44Professor Dodge identifies such private claims facilities as emerging disaggregative mechanisms. See Dodge, supra note 4, at 1272.

45This Article does not address the limits on MDL subject matter jurisdiction at the settlement stage, or how broadly the U.S. Constitution might allow the MDL courts to define the “constitutional case” (in the sense of the litigation unit). Those questions are the subject of a forthcoming follow-up article by this author.

46H.R. Rep. No. 1130, at 2–3 (1968), reprinted in 1968 U.S.C.C.A.N. 1898, 1899–1900; see also In re Phenylpropanolamine (PPA) Prods. Liab. Litig., 460 F.3d 1217, 1230 (9th Cir. 2006) (discussing and quoting the legislative history of the MDL statute).

47See Judicial Panel on Multidistrict Litig., Calendar Year Statistics of the United States Judicial Panel on Multidistrict Litigation: January Through December 2012, at 12 (2013), available at http://www.jpml.uscourts.gov/sites/jpml/files/JPML_Calendar_Year_Statistics-2012.pdf (noting that as of December 31, 2012, pending MDLs included cases related to products liability, sales practices, securities, air and other disasters, antitrust, contracts, employment practices, and intellectual property).

48Although “mass torts” is a flexible term, it is generally understood to encompass situations where many people are injured either from a single accident or event, or use of or exposure to the same product, and each has a claim for individual damages. See L. Elizabeth Chamblee, Unsettling Efficiency: When Non-class Aggregation of Mass Torts Creates Second-Class Settlements, 65 La. L. Rev. 157, 164–65 (2004) (discussing the definition, and four categories: mass accidents, dispersed/personal injury mass torts, property damage mass torts, and economic loss torts).

49See, e.g., William W. Schwarzer et al., Judicial Federalism in Action: Coordination of Litigation in State and Federal Courts, 78 Va. L. Rev. 1689, 1691–95 (1992).

5028 U.S.C. § 1407(a) (2012). The meaning of “coordinated or consolidated” proceedings has received scant attention from MDL courts, even when it impacts their decision-making. Cf. In re Equity Funding Corp. of Am. Sec. Litig., 375 F. Supp. 1378, 1384 (J.P.M.L. 1973) (“We have repeatedly declined to attempt to determine in what way and to what extent the litigation should be coordinated or consolidated. From the very beginning we have left that determination to the discretion of the transferee judge.” (emphasis added)).

51See 28 U.S.C. § 1407(a)–(b).

52Id. § 1407.

53See, e.g., Deborah R. Hensler, The Role of Multi-districting in Mass Tort Litigation: An Empirical Investigation, 31 Seton Hall L. Rev. 883, 906 (2001) (concluding that MDL in mass tort cases facilitates “the growth of mass tort litigation” despite contractions in availability of class certification).

54Howard M. Erichson, A Typology of Aggregate Settlements, 80 Notre Dame L. Rev. 1769, 1770 (2005) (“Because settlements in non-class actions need no court approval, they rarely generate reported decisions. In addition, confidentiality agreements frequently prevent publication of settlement terms. For both of these reasons, aggregate settlements tend to fly under the radar of most observers.” (footnote omitted)).

55See Hensler, supra note 53, at 894 (“[C]ollecting and transferring like claims to a single judge often encourages settlement of these claims, which may be the real goal of the parties requesting multi-districting.”).

56See, e.g., Dodge, supra note 4, at 1256.

5728 U.S.C. § 1407(a).

58Eldon E. Fallon et al., Bellwether Trials in Multidistrict Litigation, 82 Tul. L. Rev. 2323, 2344 (2008).

59E.g., In re Vioxx Prods. Liab. Litig. (Vioxx III), 802 F. Supp. 2d 740, 760 (E.D. La. 2011); Fallon et al., supra note 58, at 2338.

60E.g., Vioxx III, 802 F. Supp. 2d at 760–61; In re Vioxx Prods. Liab. Litig., MDL No. 1657, 2007 WL 3354112, at *1 (E.D. La. Nov. 9, 2007).

61See generally Vioxx III, 802 F. Supp. 2d 740 (allocating common benefit attorneys’ fees).

62Barbara J. Rothstein et al., A Model Mass Tort: The PPA Experience, 54 Drake L. Rev. 621, 625–27 (2006) (discussing strategies MDL courts use to manage discovery).

63See, e.g., In re Phenylpropanolamine (PPA) Prods. Liab. Litig., 460 F.3d 1217, 1223–24 (9th Cir. 2006) (discussing use of case management orders); see also McGovern, The What and Why, supra note 26, at 1388 (“Just as litigation raises different issues than do global resolutions, the type of information that parties seeking a global resolution may need is typically different from the individualistic, case-by-case approach of common law adjudication. Often there is even opposition to the collection of global data to be used in individual trials; a focus on the entire case can be maintained in the context of settlement without adversely impacting the litigation of individual cases.”).

64McGovern, The What and Why, supra note 26, at 1380–81 (observing that class actions under Rule 23(b)(1)(B) and (b)(3), “state class actions, bankruptcy, [MDL], and mass settlements all reach closure with a claims resolution facility”).

65See id. at 1361–62.

66See, e.g., Fed. R. Civ. P. 23(c)(2)(B) (describing the notice required for members of a class under 23(b)(3)); Fed. R. Civ. P. 23(d)(1)(B) (describing the court’s role to protect class members); Fed. R. Civ. P. 23(e)(2) (requiring court approval of class settlements); Fed. R. Civ. P. 23(e)(5) (permitting class members to object to settlement terms); Fed. R. Civ. P. 23(g) (requiring the court to appoint and supervise class counsel); Fed. R. Civ. P. 23(h) (giving the court power over attorneys’ fees for class counsel); see also Elizabeth Chamblee Burch, Aggregation, Community, and the Line Between, 58 U. Kan. L. Rev. 889, 898 (2010) (observing that MDL operates as a “procedural no man’s land—somewhere in between individual litigation and class action litigation, but without the protections of either”).

67During this early period in MDL history, “business was slow.” Fallon, supra note 3, at 372.

68Fed. R. Civ. P. 42(a)(2); see, e.g., In re Equity Funding Corp. of Am. Sec. Litig., 416 F. Supp. 161, 175 (C.D. Cal. 1976).

69See, e.g., In re Equity Funding, 416 F. Supp. at 176 (“[T]he effect of such pretrial consolidation is not and cannot be to ‘merge the suits into a single cause, or change the rights of the parties, or make those who are parties in one suit parties in another.’” (quoting Johnson v. Manhattan Ry., 289 U.S. 479, 496–97 (1933))).

70Fallon, supra note 3, at 375 (describing the modern trend of combining class actions with individual actions to form MDLs).

71In re Equity Funding, 416 F. Supp. at 170–71.

72Id. at 176.

73See id. at 175.

74544 F.2d 992, 997–98 (9th Cir. 1976).

75Id. at 994.

76Id.

77Id. at 994, 996–99.

78Id. at 994.

79Id. at 996–99.

80Id. at 1001–02.

81See id. at 1001.

82See id. at 997.

83Id. at 994, 996.

84Id. at 997–98.

85Id. at 999.

86Id. at 1001.

87953 F.2d 162, 165–66 (4th Cir. 1992).

88Id. at 164.

89Id.

90Id.

91Id. at 164–65.

92The free rider problem is fundamental to MDL case management. The PSC and various other court-appointed plaintiff committees in MDLs typically handle discovery, argue motions, appear in court, conduct hearings, and sometimes even conduct bellwether trialswork that theoretically benefits all plaintiffs in the MDL, but this work is not compensated while the MDL is pending. Fallon, supra note 3, at 373–74. Common benefit funds have thus long been used in this context to ensure this work is eventually compensated out of the proceeds of settlements or trial judgments in favor of MDL plaintiffs. Id. at 375–76.

93Showa Denko, 953 F.2d at 164.

94Id.

95Id. at 165–66.

96Id.

97Id.

98Id. at 166.

99 The Second Circuit was an earlier outlier in its approach to the federalism problem. It endorsed the use of the All Writs Act to remove state litigation to federal court in Agent Orange cases when a group of plaintiffs claimed to have discovered their own injuries after the settlement of a federal class action had concluded. In re “Agent Orange” Prod. Liab. Litig., 996 F.2d 1425, 1430–31 (2d Cir. 1993). In other words, it purported to force aggregation in the federal forum. This aggressive interference with separate state litigation was subsequently forbidden by the U.S. Supreme Court in Syngenta Crop Protection, Inc. v. Henson, 537 U.S. 28 (2002). See Chamblee, supra note 48, at 179–80 (discussing the Second Circuit’s use of the All Writs Act in the Agent Orange case and the subsequent reversal by the Supreme Court).

100See, e.g., Silver, supra note 37, at 500 (observing in the early 1990s that it was already true that narrow interpretation of class action rules made it difficult for district judges “to craft tort class actions that survive review”); see also Hays, supra note 37, at 601 (discussing the “waning” utility of class actions under Rule 23 and the imperfect nature of MDL as a replacement).

101See, e.g., Elizabeth Chamblee Burch, Litigating Together: Social, Moral, and Legal Obligations, 91 B.U. L. Rev. 87, 99 (2011) (“When defendants decide to settle, they want finality. They thus want to sweep as many plaintiffs as possible under the settlement rug.”).

102See id. at 123 (noting the problem of “[fights] over common resources—such as settlement funds”).

103See Manual for Complex Litigation (Fourth) § 20.31 (2004) (“State and federal judges, faced with the lack of a comprehensive statutory scheme, have undertaken innovative efforts to coordinate parallel or related litigation so as to reduce the costs, delays, and duplication of effort that often stem from such dispersed litigation.” (footnote omitted)); Schwarzer et al., supra note 49, at 1700–01 (discussing early efforts to coordinate cases in the 1970s).

104Schwarzer et al., supra note 49, at 1700–01.

105Manual for Complex Litigation (Fourth) § 20.312.

106Catherine R. Borden & Emery G. Lee III, Beyond Transfer: Coordination of Complex Litigation in State and Federal Courts in the Twenty-First Century, 31 Rev. Litig. 997, 1008 (2012) (quoting Conference of Chief Justices, Resolution 2: Directing the National Center for State Courts to Promote Communication and Best Practices for the Management of Like-Kind Litigation That Spans Multiple State Jurisdictions and Federal Districts (Jan. 26, 2011), available at http://ccj.ncsc.org/~/media/Microsites/Files/CCJ/Resolutions/01262011-Directing-NCSC-Promote-Communication-Litigation-State-Jurisdictions.ashx).

107See, e.g., Schwarzer et al., supra note 49, at 1702–03 (discussing a state court’s decision to “track” the federal court’s case management plan in asbestos litigation); id. at 1704–05 (same, in separate asbestos litigation); id. at 1743–44 (discussing the risk of federal judges exerting too much influence over state colleagues).

108See, e.g., id. at 1744 (“In light of federal courts’ greater resources, this tendency [to control the litigation] is understandable, but judges should take care that dominance be avoided if possible.”).

109Borden & Lee, supra note 106, 1019–20 (discussing criticism by Professor McGovern and Judge Schwarzer).

110Id. at 1014 (noting that a majority of judges aware of ongoing state proceedings communicated with their state counterparties, and of those who communicated, 94% did so directly).

111Pretrial Order No. 402, In re Orthopedic Bone Screw Prods. Liab. Litig., MDL No. 1014, 1996 WL 900349, at *3 (E.D. Pa. June 17, 1996) (emphasis added).

112Manual for Complex Litigation (Fourth) § 20.313 (2004).

113Borden & Lee, supra note 106, at 1021 (discussing a survey of MDL transferee judges finding only 13% of the respondents held joint hearings).

114See, e.g., Pretrial Order No. 467, In re Diet Drugs (Phentermine/Fenfluramine/Dexfenfluramine) Prods. Liab. Litig., MDL No. 1203, 1999 WL 124414, at *1 (E.D. Pa. Feb. 10, 1999) [hereinafter In re Diet Drugs] (“[T]he court has conferred with the [California state court judge who] . . . . is the state-wide coordinating judicial officer duly appointed by the Supreme Court of California to administer the diet drug cases filed in California state courts.”); Manual for Complex Litigation (Fourth) § 20.313.

115In re Diet Drugs, 1999 WL 124414, at *1.

116Id. at *4.

117Manual for Complex Litigation (Fourth) § 20.312.

118Id. § 20.312 & n.705.

119Id. § 20.31; Francis E. McGovern, Rethinking Cooperation Among Judges in Mass Tort Litigation, 44 UCLA L. Rev. 1851, 1858 (1997).

120But see Borden & Lee, supra note 106, at 1025 (finding that few MDL transferee judges perceived any reluctance among state judges to coordinate with the MDL proceeding); id. at 1027 (discussing a survey of state judges finding that they did not perceive relationships with federal judges to be problematic).

121See Schwarzer et al., supra note 49, at 1736–37.

122Manual for Complex Litigation (Fourth) § 20.311.

123See supra note 120 and accompanying text.

124131 S. Ct. 2368 (2011).

125In re Baycol Prods. Litig., 218 F.R.D. 197, 201 (D. Minn. 2003).

126In re Baycol Prods. Litig., 593 F.3d 716, 720 (8th Cir. 2010), rev’d sub nom. Smith, 131 S. Ct. 2368.

127Smith, 131 S. Ct. at 2374.

128Id.

129Id. at 2375, 2382.

130See Borden & Lee, supra note 106, at 998 (discussing the implications of Smith for interjurisdictional coordination).

131Professor Erichson coined the term “informal aggregation” to refer to coordinated efforts that treat “the litigation as single, integrated whole,” while still maintaining separate lawsuits with independent claims. See Erichson, supra note 1, at 383.

132See Schwarzer et al., supra note 49, at 1714 (discussing cases in which defendants would not have settled had they been forced to go to trial in state cases).

133Centralized decision-making has been identified as a line federal and state judges should avoid crossing. Borden & Lee, supra note 106, at 1019–20.

134See, e.g., Schwarzer et al., supra note 49, at 1719–20 (observing that as early as 1980, in an MDL concerning fire in an MGM Hotel, the federal judge met with state plaintiffs to assess the value of their claims and facilitate a global settlement, because that was the only kind of settlement defendants would accept).

135Judith Resnik, From “Cases” to “Litigation,Law & Contemp. Probs., Summer 1991, at 50–52.

136Such claimants have the choice whether to sign onto the settlement. They may instead decline to settle and continue to litigate individually. See, e.g., Fallon, supra note 3, at 378–79 (“The [MDL settlement] agreement is usually an opt-in agreement; when a claimant opts into the agreement, the claimant and the claimant’s primary counsel agree to be bound by the terms of the agreement, including the payment of common benefit fees.”).

137In re Zyprexa Prods. Liab. Litig., 424 F. Supp. 2d 488, 490 (E.D.N.Y. 2006).

138See id. at 491 (“While the settlement in the instant action is in the nature of a private agreement between individual plaintiffs and the defendant, it has many of the characteristics of a class action and may be properly characterized as a quasi-class action subject to general equitable powers of the court.”).

139Id. at 490–91.

140Id.

141Id. at 496 (emphasis added).

142Id. at 491.

143The analogy received the stamp of approval in a concurrence in a Second Circuit opinion. See In re Zyprexa Prods. Liab. Litig., 594 F.3d 113, 129–30 (2d. Cir. 2010) (Kaplan, J. concurring). Judge Kaplan agreed with Judge Weinstein about the “substantial similarities” between MDLs and class actions, and the need for common benefit funds of the type utilized by Judge Weinstein in the Zyprexa settlement. See id.

144MDL No. 05-1708 (DWF/AJB), 2008 WL 682174 (D. Minn. March 7, 2008).

145Id. at *6.

146Id. at *1, *10.

147Id. at *4.

148See id. at *3.

149Id. at *4.

150Id. at *11.

151Id. at *16.

152Id. at *13.

153Id.

154Id. at *18–19 (capping attorneys’ fees at 20%).

155See id. at *19.

156See In re Vioxx Prods. Liab. Litig. (Vioxx I), 574 F. Supp. 2d 606 (E.D. La. 2008); Vioxx III, 802 F. Supp. 2d 740, 760 (E.D. La. 2011) (discussing a meeting between MDL judge, and judges from Texas, New Jersey, and California in New Orleans with representatives from Merck and the PSC); Transcript of Status Conference Before the Honorable Eldon E. Fallon, United States District Judge at 22–23, Vioxx I, 574 F. Supp. 2d 606 (MDL No. 1657-L), available at http://vioxx.laed.uscourts.gov/Transcripts/01182008trans.pdf (discussing the federal judge’s conversations with state judges in New Jersey and California concerning settlement administration).

157The NPC included “all counsel appointed to the Executive Committee of the Plaintiffs’ Steering Committee” in the MDL, and also representatives of coordinated state proceedings in New Jersey, California, and Texas. See Settlement Agreement Between Merck & Co., Inc. and the Counsel Listed on the Signature Pages Hereto 1 (Nov. 9, 2007) [hereinafter Vioxx Settlement Agreement], available at http://www.officialvioxxsettlement.com/documents/Master%20Settlement%20Agreement%20-%20new.pdf; see also Vioxx I, 574 F. Supp. at 609 (announcing the settlement agreement between NPC and Merck); Adam Liptak, In Vioxx Settlement, Testing a Legal Ideal: A Lawyer’s Loyalty, N.Y. Times, Jan. 22, 2008, A12.

158See Vioxx Settlement Agreement, supra note 157, at 2; see also Issacharoff, supra note 37, at 218 (discussing the Vioxx settlement terms).

159See, e.g., Issacharoff, supra note 37, at 218 (discussing the innovative Vioxx settlement mechanism).

160 Vioxx Settlement Agreement, supra note 157, at 1.

161Id.

162Vioxx I, 574 F. Supp. 2d at 609 (“The Settlement Agreement expressly contemplates that this Court shall oversee various aspects of the administration of settlement proceedings . . . .”); see also Vioxx Settlement Agreement, supra note 157, § 6.1.1, at 29; Amendment to Settlement Agreement § 1.2.10 (Jan. 17, 2008), available at http://www.officialvioxxsettlement.com/documents/Amendments%20to%20Master%20Settlement%20Agreement.pdf.

163Vioxx Settlement Agreement, supra note 157, §§ 8.1.1–.2, at 33.

164Id. § 8.1.2., at 33.

165See, e.g., id. § 10.1.2, at 37 (“The submission of fraudulent Claims will violate the criminal laws of the United States, and subject those responsible to criminal prosecution in the federal courts.”).

166Id. § 9.2.1, at 35.

167Id. § 9.2.3, at 36.

168See Vioxx I, 574 F. Supp. 2d 606, 612 (E.D. La. 2008) (discussing “this Court’s equitable authority over the global settlement” as well as its “supervisory authority” over attorneys’ fees).

169Id. at 609.

170Id. at 612 (“[T]he Court finds that the Vioxx global settlement may properly be analyzed as occurring in a quasi-class action, giving the Court equitable authority to review contingent fee contracts for reasonableness.”).

171See Fallon, supra note 3, at 378 (discussing section 9.2 of the settlement agreement).

172Vioxx I, 574 F. Supp. 2d at 613.

173Id. at 614 (“The Settlement Agreement expressly grants this Court the authority to oversee various aspects of the global settlement administration.”).

174Id. at 611–12 (invoking the quasi–class action analogy).

175Id. at 612; see also In re Vioxx Prods. Liab. Litig. (Vioxx II), 650 F. Supp. 2d 549, 553–54 (E.D. La. 2009) (reconsidering and reaffirming the sources of authority discussed in Vioxx I).

176See generally Vioxx I, 574 F. Supp. 2d at 611–14. The scope of the MDL court’s order crystallizes when it is read in conjunction with the settlement agreement, which unambiguously included state and federal claims, as well as private claims tendered directly to the defendant, and its appointment of Judge Fallon to preside over the entire settlement. See discussion supra notes 160–61 (observing the settlement agreement’s inclusion of 14,500 private claimants who never filed suit and its resolution of claims filed in state and federal court); discussion supra note 162 (discussing the settlement agreement’s appointment of the MDL judge to preside over the global settlement); discussion supra note 163 (discussing the settlement agreement’s requirement that parties agree to the authority of the MDL judge to preside over the settlement).

177Vioxx I, 574 F. Supp. 2d at 617.

178See id. (acknowledging that 33%–40% was a standard contingent fee).

179Id. at 617–18.

180Vioxx III, 802 F. Supp. 2d 740, 758 (E.D. La. 2011).

181Id. at 762.

182Id.

183Vioxx II, 650 F. Supp. 2d 549, 551 (E.D. La. 2009).

184Id. at 556.

185Id.

186Id. at 557–58.

187See Lynn A. Baker & Charles Silver, Fiduciaries and Fees: Preliminary Thoughts, 79 Fordham L. Rev. 1833, 1847–48 (2011). Professors Erichson and Zipursky have also argued that the enforcement of the settlement itself “pushed the envelope in legal ethics” by obligating plaintiffs’ attorneys to recommend it to clients. Howard M. Erichson & Benjamin C. Zipursky, Consent Versus Closure, 96 Cornell L. Rev. 265, 267 (2011).

188Charles Silver & Geoffrey P. Miller, The Quasi–Class Action Method of Managing Multi-District Litigations: Problems and a Proposal, 63 Vand. L. Rev. 107, 109–11 (2010). Attorney Jeremy Grabill, a former law clerk to Judge Fallon, has added to the chorus critiquing the application of the “quasi–class action” idea in the Vioxx MDL, arguing it “added to the confusion that now exists concerning the proper role for courts to play.” Jeremy T. Grabill, Judicial Review of Private Mass Tort Settlements, 42 Seton Hall L. Rev. 123, 123, 178 (2012).

189Issacharoff, supra note 37, at 219.

190 Id.

191See Dodge, supra note 4, at 1275–79 (describing postdispute disaggregative mechanisms); id. at 1306–07 (analyzing the dysfunctional features of defendant-designed claims resolution facilities).

192See id. at 1310 (discussing problems of fractional enforcement).

193Fallon et al., supra note 58, at 2335.

194Id.

195Id. at 2336.

196Id. at 2335.

197See id. at 2338 (“[B]ellwether trials can precipitate and inform settlement negotiations by indicating future trends, that is, by providing guidance on how similar claims may fare before subsequent juries.”).

198See id. at 2337.

199See Oil Spill I, MDL No. 2179, 2011 WL 6817982, at *1, *2 n.1 (E.D. La. Dec. 28, 2011); Vernon Valentine Palmer, The Great Spill in the Gulf . . . and a Sea of Pure Economic Loss: Reflections on the Boundaries of Civil Liability, 116 Penn St. L. Rev. 105, 105 (2011) (discussing the magnitude of the spill).

200Oil Spill I, 2011 WL 6817982, at *1.

201Id. at *6 (applying the order to state cases where plaintiffs were represented by attorneys who also had cases in the MDL and had access to MDL work product); see also Dodge, supra note 4, at 1255–56 (discussing the GCCF).

202See Issacharoff & Rave, supra note 8, at 398, 400.

203See Oil Spill II, MDL No. 2179, 2012 WL 161194, at *2 (E.D. La. Jan. 18, 2012).

204Id.

205Burch, supra note 101, at 88.

206In re Showa Denko K.K. L-Tryptophan Products Liability Litigation–II, 953 F.2d 162, 165–67 (4th Cir. 1992).

207See, e.g., In re Fedex Ground Package Sys., Inc., Emp’t Practices Litig., No. MDL-1700, Cause No. 3:05-MD-527-RM, 2011 WL 611883, at *3 (N.D. Ind. Feb. 11, 2011) (finding the court lacked jurisdiction to order a hold-back in cases never part of the MDL); In re Genetically Modified Rice Litig., No. 4:06 MD 1811 CDP, 2010 WL 716190, at *1 (E.D. Mo. Feb. 24, 2010) (finding the court lacked jurisdiction to order hold-backs in state court cases, even though plaintiffs in state court cases had derived substantial benefit from the MDL’s common benefit work); In re Baycol Prods. Litig., No. MDL 1431MJD/JGL, 2004 WL 1058105, at *3–4 (D. Minn. May 3, 2004) (declining to assess a hold-back against settlements by parties not participating in the MDL, and finding the court lacked both original and supplemental jurisdiction over such parties); In re Linerboard Antitrust Litig., 292 F. Supp. 2d 644, 663–64 (E.D. Pa. 2003) (finding the court lacked jurisdiction to assess a hold-back in state court cases never removed to federal court or transferred to the MDL, but enforcing its hold-back order against state court cases that had been remanded after improper removal to federal court in the MDL).

208In re Genetically Modified Rice Litig., 2010 WL 716190, at *1, *4–5.

20928 U.S.C. § 1407(a) (2012).

210Id.

211See e.g., Dodge, supra note 4, at 1256.

212Fallon et al., supra note 58, at 2340 (“As in traditional tort litigation, the endgame for a mass tort dispute is not trial but settlement . . . . [and] the most ambitious settlements seek to make and enforce a grand, all-encompassing peace in the subject area of the litigation as a whole.” (alterations in original) (quoting Richard A. Nagareda, Mass Torts in a World of Settlement, at ix (2007)) (internal quotation marks omitted)).

213“The business of mass litigation dictates a collective approach,” according to Professor Erichson, who observes that plaintiffs’ attorneys often seek collective resolution because they represent large inventories of claims and plaintiffs, whereas defendants desire inclusive resolution. Erichson, supra note 54, at 1773–75.

214Grabill, supra note 188, at 126 (“[C]ourts have struggled in applying established principles concerning the scope of judicial authority to evaluate and oversee the implementation of traditional settlements in the unfamiliar context of private mass tort settlements.”).

215See supra note 212 and accompanying text.

216See Dodge, supra note 4, at 1255–56

217See supra note 35 and accompanying text.

218Grabill, supra note 188, at 126.

219131 S. Ct. 2368, 2375–76 (2011).

220See Fallon et al., supra note 58, at 2340–41 (explaining the role of the MDL court in presiding over global settlement negotiations by virtue of the “centralized forum”).

221Kenneth R. Feinberg, Reexamining the Arguments in Owen M. Fiss, Against Settlement, 78 Fordham L. Rev. 1171, 1173 (2009) (“What are we going to do with these cases in the absence of efficiency and in the absence of political will on the part of the other branches of government? What’s the reality of the situation?”).

222Resnik, supra note 135, at 5, 50.

223Id. at 50–51.

224See id. at 55 (“Federalization reflects the growing perception that problems raised in many litigations are about harms suffered by people all over the country.”).