Optimal class size, opt-out rights, and indivisible remedies

George Washington Law Review

Volumn 79, Issue 2, 2011, Pages 542-576

  Betson, David ^a   Tidmarsh, Jay ^b  

^a UNIVERSITY OF NOTRE DAME   (United States)

^b NOTRE DAME LAW SCHOOL   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 79956061583     PISSN: 00168076     EISSN: None     Source Type: Journal    
DOI: None     Document Type: Article

References (233)

1. PRINCIPLES OF THE LAW OF AGGREGATE LITIG. (2010).
2. See THOMAS E. WILLGING ET AL., FED. JUDICIAL CTR., EMPIRICAL STUDY OF CLASS ACTIONS IN FOUR FEDERAL DISTRICT COURTS: FINAL REPORT TO THE ADVISORY COMMITTEE ON CIVIL RULES 21 (1996) (reporting data that sixty-one percent of class actions were opt-out class actions and thirty-nine percent were mandatory class actions).
3. the first exception to this rule is the limited-fund class action brought under Rule 23(b)(1)(B) or equivalent state-law counterparts.
4. See FED. R. CIV. P. 23(b)(1)(B).
5. In limitedfund class actions, a court distributes proceeds from a fund insufficient to satisfy all class members.
6. See Ortiz v. Fibreboard Corp., 527 U.S. 815, 834-35 (1999).
7. In dicta, the Supreme Court has indicated that, under proper conditions, Rule 23(b)(1)(B) allows limited-fund class actions to be maintained on a mandatory basis.
8. Id. at 838-41.
9. Second, some courts permit the award of "incidental" monetary relief in mandatory class actions brought under Rule 23(b)(2).
10. See FED. R. CIV. P. 23(b)(2).
11. Compare Allison v. Citgo Petrol. Corp., 151 F.3D 402, 425-26 (5th Cir. 1998) (restricting incidental relief to awards given on a classwide basis),
12. Robinson v. Metro-N. Commuter R.R., 267 F.3D 147, 164 (2d Cir. 2001) (permitting more individualized monetary relief),
13. Dukes v. Wal-Mart Stores, Inc., 603 F.3D 571, 622 (9th Cir.) (en banc) (holding that individualized backpay awards are available under Rule 23(b)(2), but remanding for a "comprehensive analysis" regarding whether an award of punitive damages would make monetary dam-ages predominate),
14. Cert. granted, 79 U.S.L.W. 3128 (U.S. Dec. 6, 2010) (No. 10-277).
15. Third, courts sometimes allow parties to opt out of a mandatory class action. Compare Cnty. of Suffolk v. Long Island Lighting Co., 907 F.2d 1295, 1304-05 (2d Cir. 1990) (permitting plaintiff to opt out of a limited-fund class action),
16. Thomas v. Albright, 139 F.3D 227, 235 (D.C. Cir. 1998) (holding that the district court abused its discretion in permitting plaintiffs to opt out).
17. See Phillips Petrol. Co. v. Shutts, 472 U.S. 797, 812 (1985) (holding that the Due Process Clause requires that class members without minimum contacts with the forum state be given the right to opt out of a state-court class action seeking damages);
18. id. at 811 n.3
19. (limiting the holding to "class actions which seek to bind known plaintiffs concerning judgments wholly or predominately for money damages"). the reach of Shutts has spawned much uncertainty. Among the questions are its applicability in federal-court class actions, its applicability to limited-fund class actions that involve monetary awards but are usually treated as mandatory under Rule 23(b)(1)(B) (or state-law equivalents), and its applicability to class actions under Rule 23(b)(2) (or state-law equivalents), for which monetary relief is an "incidental" aspect of injunctive relief.
20. See Ortiz, 527 U.S. at 846-48 (raising in dicta the possible applicability of Shutts to mandatory limited-fund class actions);
21. JAY TIDMARSH & ROGER H. TRANGSRUD, MODERN COMPLEX LITIGATION 286-90, 419-24 (2d ed. 2010) (discussing ambiguities in Shutts).
22. Based on earlier antecedents, such as the bill of peace, the modern class action developed in equity as a mandatory device. As a general matter, equity had no authority to award damages;
23. hence, the ideas of "mandatoriness" and injunctive relief traveled together. the optout class action is a more recent device, built in 1966 on the foundation of the "spurious class action," in which litigants seeking vindication of a "several" but "common" right could opt into the class and (according to some courts) were also able to take subsequent advantage of a favorable outcome in the case even if they did not opt in. Because "several" rights usually involved claims for money, the ideas of opting out and monetary relief also traveled together. On the history of the class action,
24. see generally STEPHEN C. YEAZELL, FROM MEDIEVAL GROUP LITIGATION TO THE MODERN CLASS ACTION (1987) (describing the emergence of the modern class action from medieval roots);
25. Zechariah Chafee, Jr., Bills of Peace with Multiple Parties, 45 HARV. L. REV. 1297 (1932) (discussing the origins of bills of peace). On the development of the more innovative opt-out class action from the concept of the spurious class action,
26. see generally 7A CHARLES ALAN WRIGHT ET AL., FEDERAL PRACTICE AND PROCEDURE 1752-1753 (3D ed. 2005);
27. Benjamin Kaplan, Continuing Work of the Civil Committee: 1966 Amendments of the Federal Rules of Civil Procedure (I), 81 HARV. L. REV. 356, 375-400 (1967).
28. See also generally Debra Lyn Bassett, Just Go Away: Representation, Due Process, and Preclusion in Class Actions, 2009 BYU L. REV. 1079 (providing a brief history of the precursors to Rule 23);
29. David Marcus, Flawed but Noble: Desegregation Litigation and Its Implications for the Modern Class Action, 63 FLA. L. REV. (forthcoming 2011) (discussing the history of the Rule 23(b)(2) class action).
30. See generally, e.g., David Rosenberg, Class Actions for Mass Torts: Doing Individual Justice by Collective Means, 62 IND. L.J. 561 (1987) (arguing for class actions in mass tort cases);
31. David Rosenberg, Mandatory-Litigation Class Action: The Only Option for Mass Tort Cases, 115 HARV. L. REV. 831 (2002) (arguing for mandatory-litigation class actions and critiquing "put option" class actions).
32. Compare In re Telectronics Pacing Sys., Inc., 172 F.R.D. 271, 284-86 (S.D. Ohio 1997) (certifying a medical-monitoring class action under Rules 23(b)(1)(A) and (B)),
33. Day v. NLO, Inc., 144 F.R.D. 330, 335-36 (S.D. Ohio 1992) (certifying a medical-monitoring class action under Rule 23(b)(2)), with In re St. Jude Med., Inc., 425 F.3D 1116, 1121-23 (8th Cir. 2005) (reversing class certification of a medical-monitoring class action under Rule 23(b)(2)).
34. See WILLGING ET AL., supra note 2, at 52-53 (reporting that the median percentage of opt-outs was 0.1% to 0.2% of the class membership and noting a few cases in which significant numbers of plaintiffs opted out);
35. see also JAY TIDMARSH, FED. JUDICIAL CTR., MASS TORT SETTLEMENT CLASS ACTIONS: FIVE CASE STUDIES 10-11 (1998) (noting opt-out percentages of 2% and 8% in two mass tort settlement class actions and further noting significant numbers of optouts in other mass tort settlement class actions, but failing to provide exact percentages due to uncertainty about the size of the class membership).
36. Parties who opt out can sometimes extract larger payments than they could have from remaining in the class action
37. See TIDMARSH, supra note 8, at 39 (reporting that the 259 plaintiffs who opted out of a mass tort class action settlement received significantly more in subsequent settlements than they would have in the class action);
38. Reed R. Kathrein, Opt-Outs, MFNs and Game theory: Can the High Multiples Achieved by Opt-Outs in Recent Mega-Fraud Settlements Continue, A Discussion Draft, SECURITIES LITIGATION & ENFORCEMENT INSTITUTE 2007, at 583, 587-90
39. (PLI Corporate Law & Practice, Course Handbook Ser. No. B-1620, 2007) (discussing numerous securities fraud cases in which opt-out institutional investors received settlement awards that were numerous multiples larger than the settlement awards received by class members).
40. For an important recent effort focusing especially on constructing settlements that reconcile the competing values of individual autonomy and global peace
41. Richard A. Nagareda, Autonomy, Peace, and Put Options in the Mass Tort Class Action, 115 HARV. L. REV. 747, 751 n.8 (2002).
42. See also generally Richard A. Nagareda, the Preexistence Principle and the Structure of the Class Action, 103 COLUM. L. REV. 149 (2003) (explaining the structural distinction between mandatory and opt-out rights by means of a principle that sees the proper scope of class actions to be the resolution of preexisting legal rights).
43. PRINCIPLES OF THE LAW OF AGGREGATE LITIG. 2.04(c) (2010);
44. see also id. 2.07(c).
45. Id. 2.04(c).
46. With somewhat greater specificity, section 2.07(c) then provides that class members should not be afforded an opt-out right when a mandatory class is necessary "to manage indivisible relief fairly and efficiently."
47. Id. 2.07(c).
48. Id. 2.04 cmt. a ("'[D]ivisible remedies' are claims typically handled under Rule 23(b)(3).").
49. Id. 2.04(b).
50. the Principles countenance mandatory limited-fund class actions on somewhat more expansive terms than those permitted by the Court.
51. Compare Ortiz v. Fibreboard Corp., 527 U.S. 815, 838-42 (1999) (presuming that mandatory class actions are only appropriate in limitedfund cases in which all plaintiffs proceed under the same legal theory),
52. PRINCIPLES OF THE LAW OF AGGREGATE LITIG. 2.04 cmt. a (2010) (allowing mandatory class actions in limitedfund cases even if some plaintiffs also claim additional damages that warrant individualized treatment). At another point, however, the Reporters' Notes state that section 2.04 "is designed to explicate with greater precision the approach taken in recent years by courts under the auspices of Rules 23(b)(1)(A) and (b)(2)."
53. Id.
54. .04 reporters' notes. Because limited-fund class actions are certified under Rule 23(b)(1)(B), not 23(b)(1)(A) or 23(b)(2), the Reporters' Notes imply that section 2.04 was not intended to deal with any monetary claims, including limitedfund claims. the Reporters make no attempt to reconcile this statement with Comment a, which had recognized that section 2.04 might be used to create mandatory class actions in limited-fund cases. In another arguable change to existing law, the Principles permit the award of individualized (and variable) incidental monetary relief to class members in a mandatory class action seeking an injunction-an issue on which some courts have taken a more restrictive position.
55. See supra note 3;
56. PRINCIPLES OF THE LAW OF AGGREGATE LITIG. 2.04 cmt. b, illus. 5 (2010).
57. See infra text accompanying note 23. On the idea of "small- stakes" or "negative value" suits,
58. see generally Jonathon R. Macey & Geoffrey P. Miller, the Plaintiffs' Attorney's Role in Class Action and Derivative Litigation: Economic Analysis and Recommendations for Reform, 58 U. CHI. L. REV. 1 (1991).
59. See also Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 617 (1997) ("A class action solves [the] problem [of lack of incentive due to small recoveries] by aggregating the relatively paltry potential recoveries into something worth someone's (usually an attorney's) labor."
60. (quoting Mace v. Van Ru Credit Corp., 109 F.3D 338, 344 (1997)));
61. Castano v. Am. Tobacco Co., 84 F.3D 734, 748 (5th Cir. 1996) ("the most compelling rationale for finding superiority in a class action-the existence of a negative value suit-is missing in this case.");
62. In re Rhone-Poulenc Rorer Inc., 51 F.3D 1293, 1299 (7th Cir. 1995) (Posner, J.) ("In most class actions- and those the ones in which the rationale for the procedure is most compelling-individual suits are infeasible because the claim of each class member is tiny relative to the expense of litigation.").
63. See FED. R. CIV. P. 23(a)(1) (requiring, as a condition of class certification, that the class be "so numerous that joinder of all members is impracticable").
64. We could extend the modeling to account for the amount of care that the defendant could have taken and consequently make the amount of harm a function of A, i.e., Hi(A). For now, however, we are not worrying about the question of the optimal amount of care.
65. there are alternatives, including permissively joining with other plaintiffs,
66. see FED. R. CIV. P. 20(a)(1), filing separately and consolidating with similarly situated plaintiffs,
67. see 28 U.S.C. 1407(a) (2006);
68. FED. R. CIV. P. 42(a), and filing a competing class action;
69. see Rhonda Wasserman, Dueling Class Actions, 80 B.U. L. REV. 461 (2000) (discussing the problems associated with the filing of class actions where membership and claims overlap).
70. We recognize the importance of these alternatives in describing the "game" of how putative plaintiffs might be expected to behave in response to one plaintiff's filing of a class action. For our present purposes of describing the contours of an allocatively efficient opt-out right, however, we do not introduce this complication.
71. In most cases, RIi will be equal to the harm that the individual suffered (Hi), but, for purposes of generality, we leave RIi unspecified.
72. We can also represent this assumption with two equivalent formulations: (1) (PIi × RIi) - CIi > 0, or (2) NBIi > 0.
73. We assume that the plaintiff cannot find a lawyer to handle the case on a contingency basis. If the plaintiff can do so, a contingent-fee arrangement offloads some of the costs of an unsuccessful action onto the lawyer-a result that can affect the decision whether to sue. Our assumption that no lawyer will handle the case on a contingency seems warranted in most cases because the lack of net benefit to the plaintiff limits the fee award to the lawyer. In any event, finding such a lawyer does not affect the assumption that a plaintiff must obtain a net benefit i to bring suit. It affects only the point at which the putative plaintiff is willing to sue.
74. See supra note 16 and accompanying text.
75. For instance, if the first plaintiff prevailed, later plaintiffs might be able to use offensive collateral estoppel and eliminate their costs of proving the defendant's liability.
76. But see Parklane Hosiery Co. v. Shore, 439 U.S. 322, 331-32 (1979) (stating that a court can deny issue-preclusive effect to a finding of liability when the plaintiffs in later cases could have joined the first case). the elimination of that cost might be sufficient to make the later plaintiffs' lawsuits worthwhile.
77. Similarly, the costs of discovery in the subsequent cases might be substantially reduced because of the work done in the first case.
78. See United States v. AT&T, 461 F. Supp. 1314, 1342-43 (D.D.C. 1978) (ordering, over defendant's objection, that documents that plaintiffs in a prior antitrust case had selected during discovery be produced to a plaintiff in a subsequent case).
79. Although it is possible for a court to require parties who benefit from the creation of a common fund to share in the costs of creating that fund,
80. see Sprague v. Ticonic Nat'l Bank, 307 U.S. 161, 170 (1939), the "common fund" concept has never been extended so far as to require later plaintiffs who sue independently to reimburse earlier plaintiffs whose cases eased their own paths to recovery.
81. We can also show that, if one individual has a positive expected net benefit, then that person will file and others may file later, depending upon the outcome of the first trial and the costs that the first-filed case obviates.
82. See supra notes 24-25 and accompanying text. If two or more individuals have positive expected net benefits from suing individually, then no one will file in the equilibrium condition. Because we are restricting ourselves only to the case in which no one has an individual incentive to file, see supra text accompanying note 23, we do not consider these situations further in this Article.
83. Many games are not prisoner's dilemmas, and judges and legal scholars often overuse (and misuse) the phrase when analyzing the behavioral consequences of legal rules in game theory terms.
84. Richard H. McAdams, Beyond the Prisoners' Dilemma: Coordination, Game theory, and Law, 82 S. CAL. L. REV. 209, 214-18 (2009).
85. But this case is indeed a true prisoner's dilemma; the equilibrium condition of doing nothing leaves all participants in the game (the group of putative plaintiffs in N) worse off than an alternative (aggregation) that relies on the cooperation of others.
86. See id. at 215-16.
87. Because that cooperation cannot be guaranteed, however, and because any individual who sought aggregation might end up even worse off (due to the expense of unsuccessfully trying to aggregate) than if the person had done nothing, no party has an incentive to switch his or her strategy from "do nothing" to "aggregate."
88. More than one person can serve as a class representative.
89. FED. R. CIV. P. 23(a) ("One or more members of a class may sue or be sued as representative parties ⋯"). We simplify and assume that a single member of the class can do so. 29 On the numerosity requirement,
90. see supra note 17 and accompanying text. 30 Other requirements that the class representative and class counsel must satisfy are com monality, typicality, adequacy of the class representative, and adequacy of the class counsel.
91. See FED. R. CIV. P. 23(a)(2)-(4), (g). the class representative must also meet the related due process requirement of adequate representation.
92. See Hansberry v. Lee, 311 U.S. 32, 43-45 (1940). In prior work, one of us has suggested that a class representative meets the adequacy requirement as long as the expected net benefit to each class member (NBCi) equals or exceeds the expected net benefit from individual action (NBIi) and the expected net benefit to each class member is at least zero (NBCi = 0).
93. Jay Tidmarsh, Rethinking Adequacy of Representation, 87 TEX. L. REV. 1137, 1176-77 (2009).
94. these conditions ensure that the class action does not make class members worse off than they would have been in choosing between filing an individual suit or doing nothing. Because the equilibrium condition in negative value cases of the kind we are considering is to do nothing, the class representative must, under this theory, ensure that each class member receives an expected benefit of at least $0-in other words, must not be madeworse off by class treatment than by doing nothing. This very minimal condition might seem evident, but it has not always been honored.
95. See Kamilewicz v. Bank of Bos., 92 F.3D 506, 508, 512 (7th Cir. 1996) (refusing to overturn a state-court settlement of a negative value class action in which one member of the class received $2.19 in benefits and was assessed a fee of $91.33 to cover costs). This floor can be criticized as too minimal a description of adequacy.
96. See Patrick Woolley, Collateral Attack and the Role of Adequate Representation in Class Suits for Money Damages, 58 KAN. L. REV. 917, 944-47 (2010). Although this floor is relevant to our analysis at various points,
97. see infra notes 50, 57, 63, 74, 78 and accompanying text, our analysis is not affected if a court imposes a stronger adequacy of representation requirement.
98. the expected gross benefit could decrease if members of the class are made worse off by the class action's attempt to rectify the harm that the defendant caused. This is unlikely in most cases, especially involving monetary claims, but the situation could arise if some class members were involved in the alleged impropriety and might suffer adverse consequences if the class action's allegation were proven. In injunctive cases, it is easier to imagine situations in which all class members might be harmed by the injunction that the class representative seeks.
99. See, e.g., Hansberry, 311 U.S. at 37-38. Because the class representative could not adequately represent harmed individuals,
100. see supra note 30 and accompanying text, the class definition would need to exclude them.
101. See, e.g., Little Caesar Enters., Inc. v. Smith, 172 F.R.D. 236, 240-41 (E.D. Mich. 1997) (noting that a class definition in an antitrust case excluded, among others, the defendants, the officers and directors of the principal defendant, and its subsidiaries). Once these individuals are excluded, the condition described in the text necessarily holds.
102. In mathematical terms, an increasing gross benefit to group G can be expressed as EB(G), where the first derivative with respect to G is positive and the second is negative. Experimental data suggest that, as group size increases, the likelihood of the group's recovery increases, even if awards to some plaintiffs might decrease.
103. See Irwin A. Horowitz & Kenneth S. Bordens, the Effects of Outlier Presence, Plaintiff Population Size, and Aggregation of Plaintiffs on Simulated Civil Jury Decisions, 12 LAW & HUM. BEHAV. 209, 226 (1988);
104. see also Irwin A. Horowitz & Kenneth S. Bordens, the Consolidation of Plaintiffs: The Effects of Number of Plaintiffs on Jurors' Liability Decisions, Damage Awards, and Cognitive Processing of Evidence, 85 J. APPLIED PSYCHOL. 909, 914, 917 (2000) (reporting data showing that the likelihood of recovery increases with the inclusion of more plaintiffs, but the average damage award decreases if more than four plaintiffs are aggregated). This result seems to confirm the common sense intuition that, as the number of victims rises, factfinders are more likely to think that the defendant did something wrong.
105. 33 Cf. Ratner v. Chem. Bank N.Y. Trust Co., 54 F.R.D. 412, 416 (S.D.N.Y. 1972) (declining to certify a class of 130,000 consumers that sought $13 million in $100-per-claim statutory damages, in part because using a class action would impose excessive liability).
106. the area under the marginal benefit curve represents the total gross benefit of a group of size G, EB(G).
107. In mathematical terms, the first derivative of TC(G), the marginal cost of an additional group member, is positive everywhere. This assumption seems realistic in all cases involving monetary relief delivered to individual class members; the costs of identifying the class member and delivering a remedy to that member make costs rise when each new member is added to the class. the same is also true for many cases seeking injunctive relief. For instance, an order to establish a medical monitoring program requires the expenditure of money to identify class members and determine their eligibility for the program. Put differently, if the remedy requires individualization of relief to class members, total costs rise as the size of the class increases. Some injunctions-notably public law injunctions that strike down legislation as unconstitutional- arguably do not meet this description; beyond a certain point, they impose no additional costs as more members are brought into the class. Even in this situation, however, some class members might subsequently attack the judgment in later cases.
108. See, e.g., Hansberry, 311 U.S. at 32, 39-40 (permitting successors in interest to a property owner who was bound by a classwide injunction to challenge the injunction's applicability to them);
109. see also infra note 89 (discussing collateral attacks). Spread across the class as an expected cost, the inclusion of more class members will increase costs slightly; as the size of the class increases, the chances for such collateral attacks also increase. In any event, if total costs do not increase as the size of the class increases, then the marginal cost of adding new class members is zero. Our subsequent analysis is therefore unaffected; in such cases, the optimal class size is N.
110. See infra text accompanying Figures 4,
111. Cf. Hardy v. Johns-Manville Sales Corp., 681 F.2d 334, 346-47 (5th Cir. 1982) (refusing to give issue-preclusive effect to a prior judgment involving only one asbestos plaintiff because the defendants had a limited incentive to expend significant resources in the prior case).
112. See FED. R. CIV. P. 23(c)(2) (discussing when notices should or must be given to class members in mandatory and opt-out class actions).
113. This assumption necessarily holds when every class member obtains the same benefit from the class action. Because the marginal benefit for each class member is identical, we obtain declining marginal costs simply by sorting individuals by their costs of litigation (from the largest amount to the smallest). For reasons that we explain infra notes 47-48 and accompanying text, it is better to sort class members in a different fashion in which marginal costs might start to rise as the class size approaches N. But see infra note 59 (describing situations in which this sorting is unavailable).
114. For instance, the plaintiff with the most highly valued claim might reside in Vermont. the person with the next-greatest claim might live in Georgia, and the third might live in New York. A principal cost of maintaining any class action is providing notice to class members.
115. Eisen v. Carlisle & Jacquelin, 417 U.S. 156, 175-76 (1974) (requiring individual notice to every reasonably identifiable class member in an opt-out class action). When the class members are known, the cost of giving notice should remain fairly fixed for each new class member. When the class includes members who are not readily identifiable, however, substitute notice in print and broadcast media might become necessary. the costs of a substitute notice campaign in Vermont could be significant, but the inclusion of the second plaintiff and the cost of substitute notice in the Georgia market increases the marginal cost of including the second class member. the same is true of adding the third member from New York. At some point, a nationwide notice campaign to reach unidentified class members might become necessary. these media campaigns can be very expensive.
116. See TIDMARSH, supra note 8, at 55, 67-68 (describing costs of $7 million and $22 million associated with nationwide notice in two mass tort settlement class actions). As more members from a state are added as class members, however, these costs should fall for the reasons described in the text.
117. For now, we assume that relief in the nature of cy pres is unavailable, and that each class member must receive relief corresponding to the harm he or she suffered.
118. See PRINCIPLES OF THE LAW OF AGGREGATE LITIG. 3.07 (2010) (discussing the conditions under which cy pres relief is appropriate);
119. supra note 20 and accompanying text (declining to specify the relationship between the harm suffered and the remedy awarded);
120. see also infra note 53 and accompanying text (discussing cy pres relief). We relax this assumption infra notes 53-56, 62 and accompanying text.
121. See infra text accompanying Figures 5-6, 8-9, 11.
122. We specify possible shapes of the marginal cost line infra Figures 7-11.
123. If marginal costs are always rising, then negative value class actions could never form. Because no individual has an incentive to sue, the marginal costs of bringing a class action are higher for the first class members than the marginal benefits of a class action. If marginal costs rise from there, while marginal benefits remain steady or fall, costs would always exceed benefits, and the class action would never be viable. the fact that negative value class actions form is an experiential confirmation of our assumption,
124. see supra notes 35-37 and accompanying text, that marginal costs generally fall as class size increases.
125. See supra Figure 1 and accompanying text.
126. See supra notes 11-14 and accompanying text
127. See supra Figure 3 and accompanying text.
128. Sometimes costs might not vary. For instance, in the identical-remedy scenario, the cost of determining the remedy, which by definition is the same for each class member, does not change. Likewise, identical-remedy cases usually involve consumer fraud and other statutory violations in which individual defenses are limited or nonexistent. Finally, in some consumer fraud cases, such as cases involving holders of credit cards issued by the defendant, the cost of identifying the members of the class also might be constant across the class.
129. In addition, some literature suggests that, as a group's size increases, the incentive of group members to cooperate decreases.
130. See T. K. Ahn et al., Endogenous Group Formation, 10 J. PUB. ECON. THEORY 171, 190-91 (2008);
131. Dale O. Stahl & Ernan Haruvy, Other-Regarding Preferences: Egalitarian Warm Glow, Empathy, and Group Size, 61 J. ECON. BEHAV. & ORG. 20, 33 (2006).
132. It is not clear that this possibility has great salience in the class action context, given that little active cooperation is expected of class members.
133. See infra note 82 and accompanying text. To the extent that it has salience, however, this possibility also adds another cost that increases as the size of the class approaches N. We thank Beth Burch for calling our attention to this literature.
134. Formally, a class of size G1 is optimal when: (1) EB(X) ≥ TC(X), X < N; (2) x < g1 < n; (3) MBi > MCi for x < i = g1; and (4) MBi < MCi for g1 < i = n. On the first term, see supra Figure 3 and accompanying text.
135. See supra note 30.
136. On this assumption, see supra note 20 and accompanying text.
137. See supra note 33 and accompanying text.
138. the Principles' approach to cy pres relief is consistent with our analysis. It allows cy pres relief to be awarded when providing relief to individuals is "not viable." PRINCIPLES OF THE LAW OF AGGREGATE LITIG. 3.07(c) (2010);
139. See also id. 3.07(a) (requiring that courts provide individual relief when "individual class members can be identified through reasonable effort, and the distributions are sufficiently large to make individual distributions economically viable").
140. For instance, in Democratic Central Committee of the District of Columbia v. Washington Metropolitan Area Transit Commission, 84 F.3D 451 (D.C. Cir. 1996), the plaintiffs alleged overcharging of transit users during the 1960s. By the time that the case settled twenty-five years later, it would have been impossible in some cases and cost-prohibitive in others to identify and compensate the victims. the D.C. Circuit upheld the award of cy pres relief-in particular, the purchase of new buses that benefited present transit users.
141. See id. at 458.
142. Transit users from the 1960s who were still riding buses in the 1990s therefore obtained some indirect benefit from the cy pres award; other transit users from the 1960s (for instance, those who died or moved away) received none. On the other hand, because costs associated with the creation and distribution of the fund were charged against the cy pres fund and not charged to victims,
143. see id. at 458-59,
144. the marginal costs to these victims was comparably low.
145. Including class members g1 through x has the arguable advantage of binding them to the class judgment or settlement. But, in a negative-value situation, their individual claims were worth nothing to begin with; because they were never going to sue, this preclusive effect is, as a practical matter, irrelevant. Moreover, if these class members are assessed the costs of bringing their claims, the Constitution might prohibit a court from extending the preclusive effect of a judgment or settlement to these members.
146. See supra note 30 and accompanying text.
147. See supra notes 51-55 and accompanying text.
148. Formally, a class of size N is optimal when: (1) EB(X) ≥ TC(X), X = N; (2) MBi > MCi for x < i = n. On the first term, see supra text following note 41.
149. Formally, a class of size G2 is optimal when: (1) EB(X) ≥ TC(X), X = N; (2) x < g2 < n; (3) MBi > MCi for x < i ≤ g2; and (4) MBi < MCi for g2 < i ≤ n.
150. On the first term, see supra text following note 41. This formula is equivalent to the formula for determining the optimal group size in an identical-benefit situation.
151. See supra note 49.
152. the analysis suggested by Figures 7-9 has other limits and should be interpreted with care. Unlike the situation of a fixed remedy for each class member
153. see supra text following note 48, we cannot order class members from those with the lowest to the highest individual costs. Although we have assumed that as a class member's expected benefit falls, the class member's cost also falls,
154. see supra notes 35-40 and accompanying text, the association between falling marginal benefits and falling marginal costs is neither inevitable nor continuous. It is not inevitable because, although common costs might remain constant or decline, individual costs associated with identifying and awarding relief to some class members can result in the marginal cost for some class members exceeding the benefit that the class member obtains.
155. See supra text following Figure 4. the decline is not continuous because there are likely to be discontinuities not reflected in Figures 7-9. For instance, the marginal cost might exceed the marginal benefit on a $500 claim of a difficult-to-identify class member, but the marginal cost might not exceed the marginal benefit on a $100 claim of an easy-to-identify class member. One response to this latter difficulty is to remove such people from the class on the theory that they are inadequately represented.
156. See supra note 30. For further discussion, see infra notes 74-75 and accompanying text.
157. the analysis would be the same if G3 and G5 were the equilibrium points,
158. In our analysis, we have focused on situations in which there are two equilibrium points. It is also possible that marginal costs will fluctuate above and below marginal benefits for additional class members, creating a serpentine marginal cost line beyond g3. In such a case, there are multiple equilibrium points (one at each point at which marginal benefits and costs are equal). In terms of determining the optimal group size, the net benefits of all of the groups would need to be considered.
159. See supra notes 51-55 and accompanying text.
160. This analysis suggests how it is possible, even without bad faith or strategic behavior on the part of the class counsel or the class representative, for some class members in a negativevalue case to end up worse off than they would have been had they done nothing. the poster child for such a situation is Kamilewicz v. Bank of Boston, 92 F.3D 506 (7th Cir. 1996);
161. see also supra note 30 (discussing Kamilewicz). From one viewpoint, Kamilewicz represented a case in which the rapacity of class counsel overbore the interests of the class members.
162. Susan P. Koniak, How Like a Winter? the Plight of Absent Class Members Denied Adequate Representation, 79 NOTRE DAME L. REV. 1787, 1808-17 (2004). We do not intend to wade into that discussion. We note only that an alternate explanation is possible.
163. See supra note 30.
164. See supra text accompanying note 52.
165. Formally, the conditions are: (1) EB(X) ≥ TC(X), X ≤ N; (2) MBn > MCn; and (3) EB(N) - TC(N) > EB(G) - TC(G) for any Gxn. An equivalent, simpler formulation of the latter two conditions is MBi > MCi for x = i = n. We adopt the lengthier formulation to correspond with the formula set forth infra note 67.
166. Formally, the conditions are: (1) EB(X) ≥ TC(X), X ≤ N;(2) MBg* > MCg* (3) MBg*+1 < MCg*+1; and (4) EB(G*) - TC(G*) > EB(G) - TC(G) for any Gxn other than G*.
167. See supra note 32;
168. see also In re Agent Orange Prod. Liab. Litig., 818 F.2d 145, 167 (2d Cir. 1987) (noting "the difficulty in fashioning a distribution scheme that does not overcompensate weak claimants and undercompensate strong ones" in class actions).
169. See supra note 35 and accompanying text.
170. See FED. R. CIV. P. 23(b)(3) (stating that a class action must be "superior to other available methods for fairly and efficiently adjudicating the controversy").
171. See id. 23(b)(3)(A)-(D) (listing four factors to be used in determining superiority, including manageability); In re Bridgestone/Firestone, Inc., 288 F.3D 1012, 1018-20 (7th Cir. 2002) (relying principally on a lack of manageability to reverse a decision granting class certification).
172. FED R. CIV. P. 23(b)(1)-(2). Some courts have held that Rule 23(b)(2) requires a finding that the class is sufficiently "cohesive."
173. See In re St. Jude Med., Inc., 425 F.3D 1116, 1121 (8th Cir. 2005);
174. Barnes v. Am. Tobacco Co., 161 F.3D 127, 142-43 (3D Cir. 1998);
175. cf. In re Welding Fume Prods. Liab. Litig., 245 F.R.D. 279, 315, 315 n.189 (N.D. Ohio 2007) (noting that "whether there is an implicit cohesiveness requirement within Rule 23(b)(2) is not settled within this Circuit" and citing cases on both sides of the issue);
176. New Motor Vehicles Canadian Exp., No. MDL 1532, 2006 WL 623591, at *9-10 (D. Me. 2006) (requiring and finding class cohesiveness for antitrust claims seeking a classwide injunction). Some courts have also held that Rule 23(b)(2) contains a finding that the class be manageable.
177. Shook v. El Paso Cnty., 386 F.3D 963, 973 (10th Cir. 2004) ("Elements of manageability and efficiency are not categorically precluded in determining whether to certify a [Rule] 23(b)(2) class.");
178. Lowery v. Circuit City Stores, Inc., 158 F.3D 742, 758 n.5 (4th Cir. 1998), vacated on other grounds, 527 U.S. 1031, on remand, 206 F.3D 431 (4th Cir. 2001) (holding "that in appropriate circumstances a district court may ⋯ deny certification if the resulting class action would be unmanageable or cumbersome"). Other courts disagree.
179. See Forbush v. J.C. Penney Co., 994 F.2d 1101, 1105 (5th Cir. 1993) ("[Q]uestions of manageability and judicial economy are ⋯ irrelevant to [Rule] 23(b)(2) class actions.");
180. see also Dukes v. Wal-Mart Stores, Inc., 603 F.3D 571, 614-15 (9th Cir.) (en banc), cert. granted, 79 U.S.L.W. 3128 (U.S. Dec. 6, 2010) (No. 10-277) (importing a manageability analysis into Rule 23(a) in a massive Rule 23(b)(2) class action);
181. id. at 617
182. (discussing manageability issues in connection with Rule 23(b)(2)).
183. Cf. Debra Lee Bassett, Implied "Consent" to Personal Jurisdiction in Transnational Class Litigation, 2004 MICH. ST. L. REV. 619, 639-40 (arguing that foreign class action members should have opt-in rights because there are many factors, such as fear and misunderstandings, which might dissuade those potential members from taking part in the class action).
184. For this demonstration, see supra text accompanying Figures 6, 9, 11;supra notes 63-65 and accompanying text. One view of constitutionally adequate representation requires that class members be made no worse off than they would have been had they either chosen to file an individual suit or instead chosen not to sue, whichever decision yielded the greater expected net benefit. But not all cases have adopted this principle, so some negative value class actions might contain members who are affirmatively made worse off by class treatment.
185. See supra note 30.
186. Sometimes, however, it is necessary for these class members to remain in the case. For this discussion,
187. see supra notes 63-65 and accompanying text.
188. See supra Part II.
189. We have assumed in the text that the competitive pressure would come from another class action. Other methods of aggregation are also possible.
190. See supra note 19. Although these methods are unlikely to be useful in many negative value suits, we can generalize the claim in the text so that class members should arguably be given the right to opt out whenever any alternative form of aggregation (including an alternative class action) promises to yield a greater net benefit.
191. See Tidmarsh, supra note 30, at 1190-94. For an argument that an inequitable concentration of benefits might violate the adequacy of representation requirement,
192. see Woolley, supra note 30, at 946-47.
193. Phillips Petrol. Co. v. Shutts, 472 U.S. 797, 811 n.3 (1985);
194. see also id. at 811-12;
195. supra note 4.
196. On the nature of litigant autonomy and the effect of class actions on that autonomy,
197. see Tidmarsh, supra note 30, at 1140-51.
198. It is in theory possible that a person will misperceive both the value of remaining in the class action (believing that the net benefit will be negative) and the value of an individual suit (believing that the net benefit will be positive). This person will both opt out and bring suit. Because such a suit is worthless in fact, there is little social value in allowing it to be brought in a court system that many already regard as too clogged and, therefore, little reason to honor the person's right to opt out just to vindicate the autonomy to act based on a doubly erroneous view of the world.
199. See Shutts, 472 U.S. at 810 ("Unlike a defendant in a normal civil suit, an absent class action plaintiff is not required to do anything. He may sit back and allow the litigation to run its course, content in knowing that there are safeguards provided for his protection.").
200. Cf. GUIDO CALABRESI, THE COSTS OF ACCIDENTS: A LEGAL AND ECONOMIC ANALYSIS 135, 139 (1970) (arguing that one factor to be considered in imposing liability is the cheapest cost avoider).
201. See generally Michael A. Perino, Class Action Chaos? the theory of the Core and an Analysis of Opt-Out Rights in Mass Tort Class Actions, 46 EMORY L.J. 85, 108-19 (1997) (discussing the ways in which coalitions of class members can disrupt a beneficial class action).
202. See supra notes 63-65 and accompanying text. these are the class members between g3 and g4 in Figures 10 and 11.
203. See supra text following note 67.
204. the sampling of class members to determine the characteristics of their claims is not an uncommon practice. See Smith v. Lowe's Home Ctrs., Inc., 236 F.R.D. 354, 357-58 (S.D. Ohio 2006);
205. In re Airline Ticket Comm'n Antitrust Litig., 918 F. Supp. 283, 288 (D. Minn. 1996). the "losers" to whom we refer are the class members from g1 to n in Figure 6, g2 to n in Figure 9, g3 to g4 in Figures 10 and 11, and g5 to n in Figure 11.
206. As we have previously described, loss spreading is an efficient strategy only with respect to class members between g3 to g4 in Figures 10 and 11.
207. See supra notes 63-65 and accompanying text. For all of the other "losers" mentioned supra note 87, loss spreading is an inefficient solution.
208. See supra notes 52-55 and accompanying text.
209. the capacity of disappointed class members to attack a class judgment or settlement collaterally is a much debated subject. Compare Stephenson v. Dow Chem. Co., 273 F.3D 249, 261 (2d Cir. 2001)
210. (permitting collateral attack, albeit not in a negative value case), aff'd in part by an equally divided Court and vacated in part, 539 U.S. 111 (2003),
211. Samuel Issacharoff & Richard A. Nagareda, Class Settlements Under Attack, 156 U. PA. L. REV. 1649, 1658-59 (2008) (generally arguing against collateral attack when the court in the first case determines that the class's representation was adequate). the Principles also take a narrow view of collateral attacks.
212. See PRINCIPLES OF THE LAW OF AGGREGATE LITIG. 3-14 (2010).
213. In our judgment, should collateral attacks be permitted, the inquiry should be limited to considering whether the expected (or ex ante) benefits were less than expected (or ex ante) costs at the time of certification and (if applicable) settlement. We would not consider the ex post results of the class action, at least in the absence of strategic behavior by class counsel or the class representative to throw the costs onto some class members.
214. See supra note 70 and accompanying text.
215. See supra note 72.
216. See supra note 8.
217. See, e.g., Ortiz v. Fibreboard Corp., 527 U.S. 815, 855-56 (1999) (rejecting a class action settlement in part because of the unfairness of the distribution).
218. FED. R. CIV. P. 23(e)(2).
219. See supra note 4 and accompanying text.
220. See Mathews v. Eldridge, 424 U.S. 319, 335 (1976).
221. See Shutts v. Phillips Petrol. Co., 679 P.2D 1159 (Kan. 1984).
222. See Phillips Petrol. Co. v. Shutts, 472 U.S. 797 (1985).
223. See id. at 807.
224. See World-Wide Volkswagen Corp. v. Woodson, 444 U.S. 286, 291-92, 297 (1980) (creating a two-prong personal jurisdiction test, one dependent on whether a defendant "purposefully avails itself of the privilege of conducting activities within the forum State" and one based in notions of economy and convenience).
225. PRINCIPLES OF THE LAW OF AGGREGATE LITIG. 2.04 (2010).
226. Id. 2.04(b).
227. See supra Part II.A.
228. See supra Part II.B.
229. See supra notes 11-14 and accompanying text.
230. See PRINCIPLES OF THE LAW OF AGGREGATE LITIG. 1.03(a) (2010). the other goals were "enforcing substantive rights and responsibilities,"
231. id. 1.03(b), "facilitating binding resolutions of civil disputes,"
232. id. 1.03(c), and "facilitating accurate and just resolutions of civil disputes by trial and settlement,"
233. id. 1.03(d). In addition to better fulfilling the goal of efficient adjudication, our approach, which makes all optimally sized class actions mandatory, better fulfills the goals in sections 1.03(b) and (c), and fares no worse in fulfilling the goal in section 1.03(d).