Edited by Gerrit De Geest
Chapter 4: Class action
[In: Volume 8, Chris William Sanchirico (ed) Procedural Law and Economics]
The class action is a device that allows one or more parties, called “class representatives,” to sue on behalf of many other similarly situated persons, called “absent class members” or “absentees.” It traces its early roots to medieval forms of group litigation and its modern shape to the courts of equity of the seventeenth and eighteenth centuries (Yeazell 1989; Bone 1990).
The class action played a relatively minor role in American civil litigation through more than half of the twentieth century. In 1966, the federal class action rule, Rule 23 of the Federal Rules of Civil Procedure, was revised and expanded and ever since the class action has sparked intense controversy. Class action supporters celebrate its potential for achieving efficiency gains, redressing imbalances of litigating power, and enforcing the substantive law effectively. Class action critics condemn its potential for enriching class attorneys, pressuring risk-averse defendants to settle, and giving high damage plaintiffs only average recovery. Indeed, the class action has been called a “Frankenstein monster” and “legalized blackmail” (Miller 1979, at 665–6, noting the sharp rhetoric).
Although Rule 23 authorizes class actions for plaintiffs and defendants, virtually all class actions are brought on the plaintiff’s side (Willging et al. 1996). The class need not have existed as a group prior to the litigation. Class members can be, and usually are, total strangers to one another, with nothing in common except the fact that they have all been injured by the same or similar legal wrong. For example, one or more persons harmed by a mass-marketed drug might seek to represent a class of all persons injured by the same drug, or one or more investors might seek to represent a class of all shareholders who held shares during a defined period and suffered loss as a result of the company’s fraud.
The class action in effect functions as a complex preclusion device. It serves to adjudicate the claims of many plaintiffs in one proceeding and bind them all to the result. The specific structure of the class action reflects p. 68this goal. That structure seeks to strike an optimal balance between the benefits of preclusive class treatment, on the one hand, and the costs of aggregation, on the other, including the costs of binding parties who do not in fact participate or exercise meaningful control.
The following discussion analyzes the plaintiff class action from an economic perspective. It briefly surveys the benefits and costs and describes some of the more notable reform proposals. It is worth mentioning, however, that many class action scholars also worry about rights-based and fairness values. For example, the fact that the class action binds absent class members who have no meaningful opportunity to participate is said to intrude on the right to a personal day in court, and the fact that class settlements provide only average recovery to absentees with above-average claims raises concerns of compensatory fairness (Bone 2003b, at 487–9). These rights-based and fairness concerns are beyond the scope of this discussion.
2. The Benefits of Class Action Treatment
According to the available empirical evidence, most class actions seek damages (Willging et al. 1996). Nevertheless, classes are sometimes certified for injunctive and declaratory relief. These non-damage class actions raise fewer controversial issues, as their benefits are relatively clear and their costs much less serious than those associated with damage class actions.2
The remainder of this discussion focuses on the damage class action. There are two broad types of damage class action (Coffee 1987, at 904–6). The first type, called a large claim class action, consists mostly or entirely of plaintiffs with marketable claims, that is, plaintiffs who have enough at stake to justify hiring an attorney and suing individually. The second type, called a small claim class action, consists mostly or entirely of plaintiffs who have unmarketable claims, that is, plaintiffs with too little at stake to justify hiring an attorney and suing individually.
2.1 Large Claim Class Actions
There are two main benefits to the large claim damage class action. The most frequently cited benefit has to do with economies of scale achieved by litigating common issues of fact and law in one proceeding (e.g., Hensler p. 69et al. 2000). Consider a mass tort class action in which class members all suffer serious personal injury from an allegedly defective drug. If the class members were to sue individually, each would have to prove that the particular drug can cause the general type of injury from which she suffers and that the drug is in fact defective given the risks and the information available to the defendant. With strict limits on nonparty preclusion, each plaintiff can litigate the common issues over again even if previous plaintiffs lost. In a large mass tort, repeated litigation of common issues generates substantial duplicative litigation costs, and the class action avoids these costs by aggregating all the related lawsuits in one proceeding.
Any evaluation of economies of scale, however, must count the litigation costs that the class action itself generates. A class action is a complex procedural device that invites more intensive litigation and requires much more judicial oversight and case management than an individual suit. The additional costs it creates must be subtracted from the costs it saves (Bernstein 1978).
The second benefit of the large damage class action has to do with reducing the asymmetry of litigation costs and stakes and equalizing litigation investment across the party line (Rosenberg 2000; Hay and Rosenberg 2000 and 2002a; Note 2004). The first thing to note is that the class action eliminates a cost asymmetry that can skew the error risk in the defendant’s favor when suits are brought individually. In a class action, all the plaintiffs invest only once in preparing common questions compared to duplicative investments in separate suits. This gives the class plaintiffs the same cost advantage as the common defendant, which helps to even out litigating power and avoid skewed error.
The class action also reduces the pernicious effect of asymmetric stakes on error risk. Suppose that in our drug example, there are 1000 plaintiffs in the class, P-1 to P-1000, and that each plaintiff has one million dollars at stake. If P-1 sues alone, the defendant D has much more at stake than P-1. D knows that it faces preclusion or stare decisis in future suits if it loses on the common issue in P-1’s suit. Thus, the outcome in P-1’s suit will affect D in future suits, and D invests accordingly. By contrast, P-1 has only his own suit at stake and thus invests less than D. When one side invests more than the other, the party investing more is more likely to win, all other things equal. Thus, individual litigation creates a condition of asymmetric stakes, which generates asymmetric litigation investment and produces outcomes skewed in favor of the defendant.
The class action reduces the asymmetry and with it the skewed error risk. In our drug example, an attorney representing the class who takes the case on a contingency expects to receive a percentage of the total class recovery as a fee. Since his fee depends on total recovery, the attorney will p. 70consider the stakes for all plaintiffs in the class and invest up to the point where the marginal benefit in terms of an enhanced fee just equals the marginal cost.3 This result does not necessarily equalize litigation investment across the party line, as the class attorney receives only a fraction of the total class recovery while the defendant pays it all (Coffee 1987). However, it does reduce the asymmetry and thus the skewed error risk.
These two benefits – achieving economies of scale and mitigating the effects of skewed error – assume that the class action is litigated all the way through trial. However, virtually all class actions settle (Willging et al. 1996; Bohn and Choi 1996; Garth 1987). The prospect of settlement reduces the economy-of-scale benefits of the class action, since cost savings are less when issues are not litigated. However, it does not necessarily reduce the error-reduction benefits, since litigation asymmetry affects settlement as well as trial. When parties anticipate a probability of trial success skewed in defendant’s favor, they are likely to settle for an amount skewed in the same direction. Insofar as the class action reduces this asymmetry, it also produces settlements less tainted by skewed incentives and closer to the optimal amount for deterrence.4
2.2 Small Claim Class Actions
Because a small claim class action, by definition, includes claims too small to justify individual suits, authorizing a class action creates litigation where none would exist otherwise. Thus, the small claim class action is p. 71not about litigation cost savings or asymmetry-reducing benefits. Instead, it is about empowering parties as private attorneys general to enforce the substantive law and deter wrongdoing, where the legal violation systematically creates small harms for large numbers of geographically dispersed victims. Kalven and Rosenfeld (1941) and Dam (1975) provide classic treatments of this justification (see also Wright 1969). And numerous commentators since 1980 have examined it in more detail (Macey and Miller 1991; Rubenstein 2006).
The central idea is that small stakes disable private enforcement through individual lawsuits, and geographical dispersion coupled with free riding incentives makes voluntary organization and collective action virtually impossible. As a result, there would be no lawsuits if victims were limited to individual suits or voluntary joinder. The class action overcomes the collective action barriers and enables private lawsuits that internalize the costs of legal violations. The result is enhanced deterrence.
As an example, consider a company that fails to disclose material information to its investors. Most of the defrauded investors own too little stock in the company to warrant individual litigation, and transaction costs and free-rider obstacles stand in the way of their organizing a litigating group on their own. With the class action, however, an attorney can sue on behalf of all the investors, with one or more shareholders acting as class representatives, and obtain a large aggregate recovery and a large fee. Suppose there are 500,000 defrauded investors in the class and the average loss for each investor is $100. Also, suppose an attorney considering whether to take the class action expects a settlement of $20 million. If, as is usually the case, class attorneys receive fees roughly equal to 25% of class recovery (Alexander 1991; Eisenberg and Miller 2004b), the attorney can expect a fee of about $5 million, an amount sufficient to attract skilled counsel.
However, even a perfectly functioning small claim class action generates costs of its own. These include the costs of litigating claims that would not otherwise be brought and the delay and opportunity costs for other lawsuits. These costs must be balanced against the class action’s deterrence benefits (Bernstein 1978). Moreover, given the class action’s enforcement goals, striking the optimal balance must also take account of public enforcement mechanisms. The Securities and Exchange Commission, for example, enforces the federal securities laws; the Federal Trade Commission and the Department of Justice enforce the federal antitrust laws, and various agencies enforce consumer protection laws. From an economic perspective, the goal must be to coordinate public and private enforcement in an overall scheme that minimizes social costs.
Indeed, the problem is particularly acute because of the serious risk that p. 72private enforcement through the class action can produce overdeterrence (Rosenberg and Sullivan 2006). For example, many private antitrust class suits follow on the heels of public enforcement actions. If public enforcement already achieves deterrence, the class action might end up overdeterring, especially with the threat of treble damages and high litigation costs.
The literature includes very little systematic treatment of this coordination problem. A notable exception is Rosenberg and Sullivan (2006) which offers an interesting approach. The authors’ complex proposal, which they apply to antitrust enforcement, has three main features. First, it vests initial and exclusive enforcement authority in a public enforcer. Second, it requires the public enforcer to auction the class action to prospective private enforcers, with the public enforcer retaining an option to buy back the class action at the winning bid price. Third, it requires that the class action be a mandatory-litigation class action, meaning that it is automatically certified and must go to trial if it does not settle and no class members can opt out. Rosenberg and Sullivan argue that this procedure produces the optimal mix of private and public enforcement.
3. The Costs of Class Action Treatment
As already mentioned, class actions create high litigation costs for courts and parties, and by generating additional litigation, the small claim class action creates delay costs for other litigants. This section examines other types of cost, including costs associated with agency problems, adverse selection and opt out, and frivolous litigation.5
3.1 Agency Costs
There is a vast economic literature analyzing the principal-agent problem in many different settings. Miller (1987) gives the classic treatment for litigation in general, and numerous scholars have analyzed agency problems for the class action in particular (e.g., Dam 1975; Coffee 1987; Macey and Miller 1991; Hay and Rosenberg 2000; Klement 2002). Class agency problems take two principal forms: conflict of interest between the class attorney and the class, and conflict of interest between class representatives and other class members. The first is the more serious of the two.
It is well accepted that the class attorney controls class litigation. This has led many to analyze class actions as a type of “entrepreneurial litigation” with the attorney as the litigating entrepreneur (Coffee 1987). The dominance of the class attorney is most obvious for the small claim class p. 73action, in which representative plaintiffs function as little more than “figureheads” (Macey and Miller 1991). Few class members have enough at stake to justify monitoring class counsel and those who do normally lack expertise to perform the task effectively. Indeed, the class attorney is often the one who seeks out the named representatives and he is likely to choose representatives who are willing to comply with his wishes.6
The situation is much the same for the large claim class action. Although in these cases there are class members with enough at stake to justify monitoring class counsel, very little monitoring in fact occurs. There are several reasons for this. One has to do with lack of expertise and limited access to information about attorney performance. These problems also exist in ordinary litigation, but they are particularly serious for the class action because of its broad impact. The second reason for inadequate monitoring has to do with free riding. Monitoring is a public good that benefits the class as a whole, so each class member has an incentive to free ride on the monitoring efforts of others. The result is suboptimal investment in monitoring. Judges sometimes make incentive payments to class representatives to counteract the free-rider effect (Eisenberg and Miller 2006). But there are limits to what judges can do (Silver and Dinkin 2008).
As a result, the class attorney has wide latitude to run the class action for her own private gain. The most problematic result is the so-called “sweetheart settlement.” A sweetheart settlement is a settlement in which the class attorney and the defendant collude to maximize their respective gains at the expense of the class. The defendant agrees to pay the class attorney a large fee, and the class attorney agrees to accept a small settlement for the class. In effect, the class attorney trades a low class recovery for a high fee7 (Coffee 1987; Macey and Miller 1995; Coffee 2000; Hay and Rosenberg 2000; Klement 2002).
p. 74The social costs of sweetheart settlements differ somewhat between small and large claim class actions. Since class members in a small claim class action have too little at stake to make compensation significant, the most salient cost is the adverse effect on deterrence (Bone 1994; Gilles and Friedman 2006). In a large claim class action, such as a mass tort suit, a sweetheart settlement impairs compensation as well as deterrence. Suboptimal compensation for class members can have adverse effects on risk-bearing costs, investment in primary activity, and insurance choices.
One has to be careful, however, when assessing these costs. In the case of a small claim class action, it is possible that public enforcement might make up for any deterrence shortfall. Thus, the important question is not whether the class action itself achieves optimal deterrence, but how much of a deterrence gap remains after taking account of all enforcement methods (public and private). Moreover, for large claim class actions, one should compare the costs of class action settlement with the costs of settlement in the absence of a class action. This is significant because experience shows that nonclass settlements of informal aggregations are tainted by agency problems similar to those that plague the class action (Erichson 2000).
Trial judges are supposed to control opportunistic behavior on the part of class attorneys, but they often lack the necessary information to make an evaluation. Furthermore, faced with congested dockets and intense pressure to resolve cases quickly, judges have incentives to overlook attorney self-dealing when a class settlement promises to resolve a large number of cases at once (Coffee 1987; Macey and Miller 1995; Hensler et al. 2000). This judicial tendency is consistent with empirical findings that show a correlation between docket congestion and the size of the fee award judges are willing to approve (Helland and Klick 2007).
Moreover, sweetheart settlements can be structured in ways that conceal the true relationship between the class settlement and the fee award. One relatively simple way to do this involves making side payments to class counsel.8 A more complicated method involves structuring nonmonetary p. 75relief so it appears as though the settlement is much larger than it actually is. Nonmonetary settlements can make class members better off, but they can also compound agency problems and make class members worse off (Miller and Singer 1997).9 For example, antitrust and consumer class action settlements often include coupon distributions. Typically, the defendant and the class attorney urge the reasonableness of the fee as a proportion of the total face value of the coupons, all the while expecting the redemption rate to be low and thus the real cost to the defendant and the real value to the class to be small.10 The Class Action Fairness Act of 2005 (CAFA) targeted this practice in federal court by requiring the class attorney’s percentage fee to be based on the value to class members of the coupons actually redeemed by consumers. However, it is not clear how much of an effect this reform has had, and it is worth noting that CAFA does not reach coupon settlements in state court.11
3.2 Intra-Class Conflict: Adverse Selection and Opt Out
In large claim class actions, such as those arising from mass torts, the class attorney has an incentive to make the class as large as possible because a larger class supports a larger fee. Since there is a limit to the number of strong claims, the likely result is a class heavily stocked with plaintiffs who have weak or moderate claims. Furthermore, because settlement is designed to avoid the costs of determining individual claim values, compensation must be calculated on an average basis and the settlement distributed pro rata (perhaps subdivided by injury category). As a result, class members with above-average claims receive only average recovery. p. 76In effect, the class action transfers wealth from those with above-average claims to those with below-average claims.
To illustrate with a simple example, suppose a mass tort class includes 100 plaintiffs with strong cases, each of whose claims has an expected trial value of $500,000, and 100 plaintiffs with weak cases, each of whose claims has an expected trial value of only $10,000. The defendant understands the class attorney’s incentives and expects the class to be stocked with weak claims, which should depress the amount of any settlement. Suppose the class action settles for $20 million and the settlement is distributed pro rata to all 200 class members, so each receives $100,000. Those class members with weak claims end up with ten times the expected value of their claims, and those class members with strong claims end up with only one-fifth.
This averaging creates an adverse selection problem. Rule 23 gives each class member in (b)(3) damage class actions a right to opt out of the class.12 Armed with this right, class members with high value claims will have an incentive to opt out to avoid damage averaging, while class members with low value claims will stay in.13 The result is a class action that adversely selects for low value claims.
Che (1996) shows that there are limits to adverse selection when plaintiffs are privately informed about damages. In the equilibrium of his signaling model, some low value plaintiffs leave the class in order to pool with exiting high value plaintiffs and to signal (falsely) that they too have high value claims. Moreover, some high value plaintiffs remain in the class. Still, a significant adverse selection effect remains.14
Adverse selection is costly for two reasons. First, it undermines the efficiency of the class device by encouraging class members to litigate separately and by diverting social resources to compensate low value p. 77claims. Second, adverse selection exacerbates agency problems by driving high value plaintiffs out of the class and making it even less likely that the remaining class members will have enough at stake to monitor the class attorney.
There are other ways in which the interests of class members can conflict (Morawetz 1993), and when they do, there is always a risk that powerful parties will prevail over weaker parties. One solution to this problem involves dividing the class into subclasses with separate representatives and separate attorneys. Subclassing, however, reduces the benefits of class-wide adjudication and raises the transaction costs of settlement bargaining. Miller (2003b) analyzes intra-class conflict and proposes an approach based on hypothetical ex ante consent.15 (See also Morawetz 1993, examining theories of distributional fairness to guide measures for handling intra-class conflict, and Dana 2006, using hypothetical consent as a standard for determining the permissibility of collateral attack.) Kornhauser (1983) offers a framework based on cooperative game theory and the solution concept of the core to analyze the fairness of a class settlement in the presence of interest conflict.
3.3 Frivolous and Weak Class Action Suits
Some observers of the class action worry that the settlement leverage created by class certification might pressure defendants to settle frivolous and weak class action suits. Commentators disagree about the precise magnitude of this risk. Some commentators believe it is at least serious enough to justify some kind of merits review at the certification stage (Bone and Evans 2002; Priest 1997). Others are more skeptical (Silver 2003).
Those who see a potentially serious problem point to several contributing factors, such as the way the class action magnifies stakes and complexity thereby compounding defendant’s litigation, risk-bearing, and reputation costs (Bone and Evans 2002; see also Note 2005, discussing the effect of risk preferences on class action litigation). The judicial tendency to push for settlement only makes matters worse, as do cost and informational asymmetries (Coffee 1986) and the practice in federal court of reviewing the merits of class claims in only a very limited way (Bone and Evans 2002).16
4. p. 78The Special Case of the Settlement Class Action
The settlement class action is particularly susceptible to agency problems. In a settlement class action, the parties enter into a settlement before filing suit and file a class action not to litigate, but to obtain judicial approval of the settlement so they can bind the entire class. The most notorious examples of settlement class actions involve mass torts. In several highly publicized asbestos cases, defendants tried to use the settlement class action to obtain global peace by binding everyone, including those with future claims.
Coffee (1995, 2000) provides an extensive discussion of the agency problems and sweetheart settlement risks of mass tort settlement class actions. These include the incentives of class counsel to negotiate particularly large fees and the incentives of defendants to play lawyers off against one another in reverse auctions to obtain the most attractive settlement arrangement.
5. Reform Proposals
There are many reform proposals designed to improve the class action’s cost-benefit balance. Some have been implemented; others are only suggestions. The following, while hardly comprehensive, discusses some of the more salient proposals.
5.1 Improving the Monitoring of Class Counsel
One set of proposals aims to reduce agency costs by improving the monitoring of class counsel. The Private Securities Litigation Reform Act (PSLRA) implements this approach for securities fraud class actions in federal court. After filing a securities class action, the would-be lead plaintiff must publish notice inviting other class members to apply for the lead plaintiff role. The district judge considers all the applicants and chooses as lead plaintiff the one “most capable of adequately representing the interests of class members.” The PSLRA creates a presumption in favor of appointing the investor or group of investors with the “largest financial stake in the relief sought.” Once appointed, the lead plaintiff selects and retains class counsel, sets the terms of the fee award, and (ideally) monitors counsel’s performance throughout the litigation.
Securities fraud classes often include a few large institutional investors who have enough at stake to justify monitoring class counsel. The goal of the PSLRA is to attract these institutional investors to the role of lead p. 79plaintiff, thereby improving the monitoring of class counsel. Weiss and Beckerman (1995) proposed this idea and Congress implemented it in the PSLRA.
Cox and Thomas (2006) review anecdotal and empirical evidence bearing on the success of the PSLRA in attracting institutional investors to the role of lead counsel. They find a very low (though increasing) rate of participation. They also find that institutional investors produce larger class settlements, which offers some support for an improved lead-plaintiff-as monitor approach. (See Choi (2011) for a review of the empirical literature and a study of the PSLRA’s impact on agency costs.)
Silver and Dinkin (2008) explain why private institutional investors (as opposed to public and union funds) are reluctant to volunteer for the lead plaintiff role, and they propose compensation mechanisms, including a bidding scheme, to attract them.17 Klement (2002) also proposes a bidding scheme to select the monitor, but in his proposal bidders do not have to be class members or involved in the class suit in any way other than as monitors. The winning bidder selects counsel and receives a percentage of the class recovery to compensate for its monitoring efforts.
5.2 Bolstering Judicial Review of Class Settlements
One way to guard against sweetheart settlements is to bolster trial judge review of proposed settlements. Reviewing settlements is problematic in part because no one in the litigation has an incentive to identify problems. Both class counsel and the defendant want the judge to approve the settlement. Absent class members can raise objections, but data show that they seldom take advantage of the opportunity (Eisenberg and Miller 2004a). Judges sometimes appoint a guardian ad litem to advocate on behalf of the class or a special master or court monitor to assist in settlement review. But compensation issues and other problems complicate these appointments, and there is at least some risk of collusion among the appointed agent, the defendant, and class counsel.
The Class Action Fairness Act implements a different approach. CAFA requires that the court inform public authorities of the pendency of a class p. 80action settlement. This is supposed to facilitate the involvement of state and federal government officials, who can advocate for the public interest and the interests of the class.
5.3 Adjusting Fee Award Rules
There is a very large literature investigating fee award rules as a way to align the incentives of class counsel with the interests of the class and with the public interest in optimal deterrence.18 Proposals abound. For example, Macey and Miller (1995) recommend that the fee award be calculated on a percentage-of-recovery basis, so counsel has an incentive to maximize class recovery. Hay and Rosenberg (2000) and Hay (1997b) propose a fee formula that equalizes the fee that class counsel expects from settling with the fee that he expects from going to trial. The idea is to eliminate the incentive to enter into sweetheart settlements by making it impossible to buy off the lawyer with a larger fee. Klement and Neeman (2004) use mechanism design theory to derive a fee schedule that awards a fixed percentage for recoveries above a specified threshold.
Some commentators have proposed auctioning the role of class counsel, and the auction approach has actually been used in a few securities class action cases. Prospective attorneys for the class submit bids with proposed fee schedules, and the judge chooses the lowest bidder subject to quality constraints. There are, however, serious problems with this approach. For a review of the cases and the critique, along with a novel auction proposal, see Harel and Stein (2004).
One particularly ambitious variant on the auction idea involves auctioning the entire class claim. Macey and Miller (1991) first proposed this idea for securities class actions and derivative litigation, and Macey and Miller (1995) generalize it to include mass tort class actions as well. The idea is to eliminate divided ownership and internalize the costs and benefits of litigation in the party who controls the lawsuit. Thomas and Hansen (1993) apply insights from auction theory to criticize this proposal.
5.4 Screening Frivolous or Weak Class Action Suits
There are several reform proposals specifically addressed to screening frivolous and weak class actions suits. One approach adopted by the PSLRA involves strengthening sanctions for frivolous filings. Another approach, tentatively suggested by Hensler and Rowe (2001), would apply a modified p. 81version of the British Rule to shift some portion of the winner’s attorney’s fee to the losing side.
A third approach would have the trial judge conduct a preliminary merits review at the class certification stage and grant certification only if the plaintiffs can demonstrate some significant probability of success on the merits (Bone and Evans 2002). Although this proposal is likely to increase process costs and erroneous denials of class certification, it will reduce erroneous grants and, according to its supporters, is likely to produce more benefits than costs.
The fourth proposal, by Hay and Rosenberg (2000), aims to reduce the effects of defendant risk aversion in pressuring settlement and attracting frivolous and weak class action suits. Hay and Rosenberg (2000) propose that the judge “hold multiple class trials and base its judgment on some suitable weighted combination of the class verdicts” (at 1382). This approach reduces the defendant’s risk-bearing costs by avoiding the all-or-nothing quality of a single class trial.
This entry is based on the treatment of the class action in Bone (2003a), at 259–98.
These non-damage class actions achieve three main benefits: a remedial benefit in facilitating a group injunction, a heuristic benefit in focusing the judge on the group nature of the wrong, and a practical benefit in guaranteeing class-wide preclusion.
This assumes substantial agency costs; in other words, it assumes that the attorney runs the class action with the exclusive goal of maximizing his own fee. Agency costs in the class action setting are discussed in Section 3.1 below.
This is a highly simplified analysis. A more complete treatment should consider effects on settlement rate and settlement amount. As for settlement rate, if the class action increases total litigation investment, it also increases the potential surplus from settlement. Depending on the circumstances, a larger surplus can increase the likelihood of settlement by enlarging the range of feasible settlement allocations, or reduce the likelihood by encouraging strategic bargaining over the larger settlement stakes. The effect on settlement amount can be seen by considering a simple settlement model where p is the probability of success at trial, w is the expected trial award conditional on success, and CP and CD are the expected trial costs for each side. Assuming both sides agree on values for these variables and there is equal bargaining power, the settlement range is [pw − CP, pw + CD], and the likely settlement is at the midpoint of this range, i.e., pw + (CD − CP)/2. If class treatment elicits greater relative investment from the plaintiffs, it should increase p, which increases the settlement amount all other things equal. But insofar as class treatment aligns litigation investments more closely across the party line, it reduces CD − CP, which reduces the settlement amount. Without more information, it is not possible to determine which of these opposing effects dominates.
There are other costs as well. For example, Che (2002) shows how class action plaintiffs can exploit informational asymmetries in settlement bargaining.
As discussed below, the Private Securities Litigation Reform Act (PSLRA) tries to address this problem for securities fraud class actions in federal court by creating a procedure to select as class representatives large institutional investors who have the incentive and wherewithal to monitor class counsel (Cox and Thomas 2006; Silver and Dinkin 2008).
Attorneys often compete for the class fee by filing separate and overlapping class actions aimed at the same wrongdoing. The idea is to race to be the first to secure a settlement and with it the class fee. (A settlement in one class action usually precludes class members from pursuing other class actions.) Defendants can exploit these incentives by pitting attorneys against one another in a “reverse auction.” The defendant in effect shops for an attorney willing to accept a settlement on defendant’s terms. For a discussion of the reverse auction, see Coffee (1995). For an analysis of overlapping class actions in general, see Miller (1996).
In mass tort cases, for example, class counsel will often have a large inventory of individual clients in addition to those in the class. By structuring the class action to exclude the inventory clients, the class attorney and the defendant can enter into a deal in which the defendant settles the class claims for a small amount and the inventory claims for large amounts. In this way, class counsel receives a substantial fee from the inventory settlements without disclosing the amount to the judge reviewing the class settlement. In the past, this technique was used as part of a strategy for achieving global class settlements of future claims (Coffee 1995, at 1373–5). Congress targeted this practice in the Class Action Fairness Act of 2005, which requires parties to disclose all side agreements when the judge reviews a class settlement. The same requirement is also imposed by the 2003 revisions to Rule 23 of the Federal Rules of Civil Procedure. But it is not clear how effective these reforms have been.
See also Durand (1981) for an economic analysis of fluid recovery remedies, which distribute class damages in ways only roughly matching individual harm.
There are other problems with coupon settlements. The effect of coupons on price can leave consumers as a group no better off (Borenstein 1996), and the effect of coupons on demand can create deadweight loss (Polinsky and Rubinfeld 2008).
Also, CAFA does nothing to address the reversionary fund settlement, another popular way to disguise sweetheart settlements. In a reversionary fund settlement, the defendant creates a large settlement fund for the class and pays the class attorney a large fee. The large fee appears reasonable as a percentage of the total settlement fund. However, the attorney and the defendant collude to minimize the number of class members claiming on the fund, and the amount of the fund left after a specified period reverts to the defendant.
Rule 23(c)(2). In addition, class counsel must give the best practicable notice to the class, including individual notice to all class members who can be reasonably identified, so that class members can decide whether to exercise their opt-out right.
A high value plaintiff has an incentive to opt out when individual litigation or joinder in a nonclass aggregation promises a larger net recovery. In theory, a class action could unravel completely at the high end, as the expected settlement and pro rata share decline with exiting high value plaintiffs. The incentive to opt out, however, depends on a number of factors, including how much the plaintiff knows about the strength of her claim, the additional cost of litigating separately, and the degree of risk aversion.
Friedman (1996) proposes a mechanism for generating damage payments to class members that reflect the strength of their individual cases. His mechanism extrapolates from trial verdicts in a sample of cases drawn from the class after giving the defendant an opportunity to choose the sample.
For a critical analysis of hypothetical consent arguments generally in procedure, see Bone (2003b).
In recent years, some courts have been willing to investigate the merits more closely at the certification stage, although the inquiry is confined only to those merits issues relevant to certification standards. See, e.g., In Re Hydrogen Peroxide Antitrust Litigation, 552 F.3d 305 (3d Cir. 2008); Initial Public Offering Securities Litigation, 471 F.3d 24 (2d Cir. 2006).
Kobayashi and Ribstein (2004) discuss a special problem that arises when the attorney who prepares the complaint and initiates suit is not the attorney chosen as lead counsel. In such a case, the selected attorney is able to free ride on the efforts of the originating attorney, and this undermines ex ante incentives to investigate, research, and file class action suits. Kobayashi and Ribstein propose that the originating attorney receive special consideration in the appointment of lead or co-lead counsel and be compensated for his efforts by receiving a portion of the class fee.
For an empirical study of class action fee awards, see Eisenberg and Miller (2004b).