Civil Justice Report
No. 4 March 2002
Class Actions: The Need for a Hard Second Look
Richard A. Epstein
About the Author
Richard Epstein received a B.A. in philosophy summa cum laude from Columbia in 1964. He received a B.A. in law with first class honors from Oxford University in 1966, and an LL.B., cum laude, from the Yale Law School in 1968. Upon his graduation he joined the faculty at the University of Southern California, where he taught until 1972. In 1972, he visited the University of Chicago, and became a regular member of the faculty the next year. He was named James Parker Hall Professor in 1982 and Distinguished Service Professor in 1988. He served as Interim Dean of the Law School from February to June of 2001.
Mr. Epstein has written extensively in many legal areas. His books include: Principles for a Free Society: Reconciling Individual Liberty with the Common Good (1998); Mortal Peril: Our Inalienable Right to Health Care? (1997), Simple Rules for a Complex World (1995), Bargaining with the State (1993), Forbidden Grounds: The Case Against Employment Discrimination Laws (1992), and Takings: Private Property and the Power of Eminent Domain (1985). Mr. Epstein is also the editor of Cases and Materials in the Law of Torts (7th ed.) and has written a one-volume treatise, Torts (1999). He has also written many scholarly articles on a broad range of common law, constitutional, economic, historical, and philosophical subjects.
He spent the 1977-78 year as a Fellow at the Center for Advanced Studies in the Behavioral Sciences at Stanford. From 1981 to 1991 he was editor of the Journal of Legal Studies. Since 1991, he has been an editor of the Journal of Law & Economics. He was elected a Fellow of the American Academy of Arts and Sciences in 1985.
The topic of class actions is one of the most ubiquitous in modern civil law. It is only a small exaggeration to say that any major innovation in liability, if brought about by litigation, will be either created by or reflected in class actions. The reason for their omnipresence lies in their versatility. Class actions can be used in theory to amalgamate large numbers of claims brought by separate individuals, regardless of their subject matter. Any lawyer who works with antitrust, corporations, securities, discrimination, lending, real property or torts, will necessarily be familiar with class action litigation as a normal part of his or her work. It is hard to describe class actions as a specialty when so many lawyers both pursue and defend these suits on a daily basis.
Nor is it hard to see why class actions have surged to prominence in recent years. As litigation becomes ever more complex, the willingness and ability of individual plaintiffs to bear its costs is correspondingly diminished. The opportunities for gains, however, remain substantial, so the void is quickly filled by entrepreneurial lawyers who hope to profit by organizing a class of potential plaintiffs and bringing their joint claim to a successful conclusion. As might be expected, a development this conspicuous has not escaped examination. Defenses and condemnations of class actions are staples of the modern literature. This essay is one more attempt to put the entire subject into perspective, by seeking to identify both the uses and the limitations of the class action in modern civil litigation.
In order to undertake that task it is, I think, at the outset necessary to set out some conceptual framework to deal with the knotty conceptual and procedural issues that class actions raise. Once articulated, that framework should facilitate a more detailed examination of class action litigation. It may not be possible to resolve all outstanding issues on the role of class actions once the relevant trade-offs are identified, but at least we should be able to make some sensible first approximations. The first section of this paper therefore outlines in brief compass the general approach that I take to this, and indeed all legal matters. In it I try to develop the appropriate balance between two imperatives, each accepted as a good in its own right: the desire for personal control of one’s own claim, and the need for the coordination of separate claims brought in related matters.
The second portion of this paper then uses these principles to explain the ends to which class actions should be devoted and the mechanisms that might help advance their efficient use. In dealing with this issue, I begin the discussion with class actions in the context of the law of groups. The field is vast and covers the full range of voluntary associations, including partnerships, charitable organizations and corporations. The main, by no means exclusive, field of action lies in the area of corporate law, both through derivative suits and direct actions by shareholders. There is no exact cubbyhole for these actions under Rule 23 (b)(1)(B) of the Federal Rules of Civil Procedure. But it appears that suits of this sort “would create a risk of adjudications with respect to individual members of the class which would as a practical matter be dispositive of the interests of the other members not parties to the adjudications or substantially impair or impede their ability to protect their interest.” That mouthful gives the uninitiated little information as to the paradigmatic case that falls within that provision. But one of the notes to the Federal Rules helps fill that gap. “In an action by policy holders against a fraternal benefit association attacking a financial reorganization of the society, it would have hardly been practical, if indeed it would have been possible, to confine the effects of a validation of the reorganization to the individual plaintiffs.” (Supreme Tribe of Ben-Hur v. Cauble, 255 U.S. 356 (1921). Other examples to which I shall revert include the declaration of corporate dividends or the handling of various other corporate distributions.
These cases are not the locus of the current controversy over class actions, but I discuss them first because they illustrate the situations in which class actions, owing to the fungible interests of all group members or shareholders, have the greatest utility. The purpose of the discussion is to show the formidable difficulties that remain in the implementation of the class action in the soil most congenial to its growth. Once that pattern is accepted, we should be in a position to explore the added complexities when class actions are removed from the associational and corporate contexts to cover claims brought under, for example, tort, discrimination, and antitrust law. The final result here is neither praise nor condemnation but the clear appreciation that the only way to overcome the imperfections of the ordinary rules of civil procedure used for the prosecution of individual claims is to invite a different, and sometimes larger, set of imperfections under the class action rubric.
Autonomy and Forced Exchanges
A somewhat outdated name for civil procedure is adjective law. The substantive law determines the rights and duties of ordinary individuals, and thus is the chief concern of any legal system. The “adjectival” rules of procedure are sidekicks to the substantive law, because their major function is to translate abstract claims into concrete cases whose outcomes comport with principles of the substantive law. In dealing with those substantive issues, it is often said that the bedrock principle of the common law lies in its respect for individual autonomy or self-rule. Each person is said to be the owner of his own body, and can decide when to steer clear of certain transactions and when to enter into them. Individual autonomy allows ordinary individuals immunity from external aggression; it explains why they are allowed to acquire the ownership of property and why they are allowed to sell their labor only on terms that they regard as personally satisfactory.
These abstract entitlements, however, remain inchoate until one person asserts that the actions of another have violated one of these rights. To stick only with the simplest case, the right to individual autonomy may allow all individuals to have exclusive control over their own body, but the content of that right becomes most clear when some other individual invades that person by, say an assault and battery. Similarly, individual rights to property are crystallized only when some other person takes away or destroys what they own. The right to contract is made vivid only when some other person breaches an undertaking to the promisee.
At this point we have to ask this question: who holds the cause of action for damage to the person, for loss of property, for breach of contract? Now bedrock principles of substantive law start to blend in with the rules of civil procedure. The usual “right” answer is that any right of action belongs to that individual whose rights were invaded. That proposition seems to apply not only to the traditional common law claims just mentioned, but also to other forms of individual claims that apply to the violation of other forms of right created under statute. The victims of discrimination are normally entitled to sue the perpetrators of that discrimination; the victims of monopoly practices normally hold their own claims.
My initial query about these everyday outcomes is the heretical why? Here it hardly counts as a logical contradiction to assert that A’s rights were violated but that B has the right to sue. Indeed, if B were an ideal claimant and A were hopeless at litigation, then this odd regime would have some real attractiveness not only for the Bs of this world, but for the entire system as a whole. The defendant cares not one whit who gets his money, but only about the likelihood and magnitude of payment. Let B be the perfect professional plaintiff, then these defendants will face higher liabilities, and thus will take greater steps to avoid harm. They will know a similar fate that awaits them from violating the rights of other individuals. Accordingly, they will refrain from the deliberate invasion of these rights and will take care to avoid the accidental violation of these rights as well. The greater security in the person, in the protection of property, in the performance of contracts is enjoyed not just by the Bs of the world, but by the As of the world as well. They might be quite pleased to be stripped of their rights, so long as they believe that their substantive rights will be protected by others that bring suit in the event of loss.
The point of this fairy tale is not to defend the proposition the holders of rights of action should not be the people who hold the initial substantive entitlement. It is rather to show that no necessary or logical contradiction arises from the simple fact of that separation. Putting the matter in this particular fashion requires us to think hard about the question of who gets the right to sue, and why. It thus forces us to fashion a functional explanation for the unity of substantive and remedial rights, which in turn leaves open the possibility that this unity may be desirable in most cases, but not in all. The conceptual difficulty of this exercise has real payoff in understanding some of the peculiar features of class actions. But for the moment I shall defer dealing with that question, and ask the simpler point: how do we make a case for the “self-evident” proposition that A should have the cause of action for the violation of A’s rights?
The first point is that it is one thing to contemplate the separation of substantive entitlements from rights of action, but quite another thing to determine who has that right of action. Once A and B could be different people, it is necessary to posit some rule to decide which B is entitled to the right of action for the substantive violation to the rights of each A. How might this be done? Well one possibility is to assign the rights arbitrarily, but that I think will not commend itself to anyone. A has suffered a substantial loss while B gets an undeserved windfall. It looks as though we should rather have the damage payment (if such it be) neutralize the loss, rather than to give it to someone whose name is drawn out of a hat.
Even if (or, after) we reject random assignment, we might with an eye to efficiency think that the right of action for the violation to A could be auctioned off by the state, so that the winning bidder may bring the suit, and keep the proceeds of settlement or litigation. After all, auctions are used all the time to sell paintings and tulips, so why not causes of action for broken legs and undelivered goods? But the language of an auction invites other questions. The auction presumably involves some form of payment for the right of action. It therefore becomes necessary to ask, who gets the proceeds? One possibility is to pay A in lieu of compensation for the wrong he has suffered. But that solution is less than ideal if the government has to bear the costs of running the auction without recovering its overhead expenses. And if we pay A after all, why not let him keep the right of action to begin with?
The next question is who gets to bid on the cause of action. It should not be assumed that A and B have to be different people. Perhaps A should be allowed to bid in to the state auction in order to recover the right to sue on his own behalf. It is likely that A will be an impressive entry into the bidding wars. Prosecuting any lawsuit for the violation of A’s rights will require the cooperation of A. If liability turns on whether X struck A in self-defense, then A’s testimony will be critical to overcoming this defense. In general, A’s cooperation is needed for all aspects the case, but once A gains nothing from suit, he has no incentive to cooperate with the winning bidder. Other individuals will therefore discount their bids by the attendant costs of securing cooperation. A does not labor under this disability and therefore should have an inside track for the winning bid, at least if he has the resources in question. But, of course, he might not. The injury could have drained him of cash resources, and our sophisticated capital markets do not regard a potential cause of action as adequate collateral for a loan. (No bank could lend even $10 on an asset worth $1,000 if that value consists of a 50 percent chance of $2,000 and a 50 percent chance of $0. The high variance in payoffs leaves the bank with a huge downside and no participation in the upside.)
The auction rule does not seem to be all that attractive in the abstract. This situation does not involve individuals who auction their own property, keeping the proceeds for themselves. Rather, it is a state device for deciding who gets the right to own the cause of action in the first place. Many individuals, most notably Ronald Dworkin, have suggested that all property in the state of nature be auctioned off by the state to the highest bidder.1 But that suggestion loses its operational appeal once we realize how difficult it would be to organize its operation before many people died of starvation. We (by which I mean all early societies without exception) therefore adopt a rule of first possession for the acquisition of land and chattels from the state of nature. Reluctantly, we reach a similar conclusion here. A rule that assigns the cause of action to the victim of the breach is less expensive to operate than any auction that we might set up; it enjoys legitimacy with a populus that quickly tires of strange mind games played by fevered law professors; it usually ends up giving the right of action to an individual who is in a good position to prosecute the suit himself; and, most critically, it allows that person to enter into side contracts with other individuals (call them lawyers) for the prosecution of that suit, if it turns out that they do not have the skills to do it themselves. Indeed, in principle it could allow the individuals to sell the claims to other persons, even by auction, if that seems appropriate, as it often is in cases that involve the collection of receivables. But most often the contingent fee arrangement is the vehicle of choice because it gives access to adjudication to aggrieved persons with both limited wealth and limited ability to monitor the conduct of their lawyers. The bottom line here is that the initial allocation of the right of action to the person who suffers from its breach looks on examination to comport not only with the shadowy dictates of natural justice, but also to have real efficiency justifications that no global auction can duplicate. It gives a quick, clear and determinate owner to the right of action in question, regardless of its substantive content, and that person can make voluntary dispositions of the cause of action in the event that he is not the ideal claimant. Auctions are possible, after a fashion. But they are run not by the state, but by the owner of the cause of action, where they need not be for sale but for the ubiquitous contingent fee arrangement.
Once we have reached this simple empirical conclusion, then we can ask the painful but necessary question. Is it good in each and every case? As often happens in mental exercises of this sort, our mental map of the transactions in question takes a particular form that is nowhere required by the statement of the problem itself. In this case, our usual view that a cause of action flows from the violation of a substantive right takes on the following salient features. First, there is one and only one person who is victim of the violation of the right. It is defendant who has taken plaintiff’s plow; no other plow has been taken. The value of the plow is large not only in absolute terms but also relative to the costs A has to incur to recover it from B. The legal system, moreover, will yield reliable results, such that A can prosecute his suit for a small cost, with some confidence that he will win on a meritorious claim.
Once we make these implicit assumptions explicit, then we can identify why the system of private rights is workable. By adopting an inflexible rule that each owner of property retains the right of action for its theft or destruction, we have eliminated a major stumbling block in running the legal system. A of course must still prove that the plow was his, for the defendant will win if he can show that he lent the plow to the plaintiff, with the understanding that it would be returned on demand. But no system of procedure can eliminate that factual dispute. Our rule only gets rid of the distraction that arises when some third party is vested with this cause of action. Thus, the plaintiff here will act generally as a self-interested person, which means that he will bring suit only if he estimates at the outset that his expected gain from the suit will exceed his expected costs. Both the cost of the legal system and the reliability of its processes enter into the plaintiff’s crude calculations. Thus, if the plow is worth $1,000 and the cost of suit equals $100, then, if recovery is certain, A will sue, for a net $900 leaves him better off given that the plow was taken, even if he is worse off than if the plow had never been taken. But this calculation ignores the risk of loss. If plaintiff thinks that he has only a 75 percent chance of winning the case, his expected gain drops to $650, equal to the $750 he expects to recover (i.e. 0.75 x $1,000 - $100). But let the costs go up to $500, and the chances of success drop to 40 percent, and all of a sudden, the suit does not look attractive no matter how sound the underlying cause: a $400 recovery is less than $500 cost.
Enter the Class Action
We are now in a position to understand the origin and appeal of the class action in some, but not all, cases. Quite simply, the unthinkable about the ownership of claims becomes thinkable when the basic scenario changes. All that we need to do in order to make this happen is to alter three parameters. First, the number of individuals similarly situated with respect to a common defendant becomes very large. Second, the loss sustained by each party is relatively small. Third, the administrative costs of individual suit turn out to be quite high. In these circumstances we can now see the consequences of a rule that allows each aggrieved individual to bring his own suit. Quite simply, he will not accept this invitation if the costs of litigation exceed the level of recovery, which could easily happen with the high price of lawyers. Within the framework of voluntary transactions, we might expect A to sell his claim, but any individual buyer will face all the problems that beset A and still have to enlist A’s support in order to make his claim good.
The obvious escape hatch to this impasse in a voluntary world is for all the individuals to pool their claims together (under the rules of permissive joinder, as authorized under Rule 20 of the Federal Rules) in order to take advantage of what they hope will prove to be economies of scale. These rules limit the use of permissive joinder to cases in which the parties pursue their rights “in respect of or arising out of the same transaction, occurrence, or series of transactions or occurrences.” But this limitation on the use of permissive joinder hardly binds at all, for unless this condition is satisfied the individual plaintiffs have little reason to pursue cooperative activities anyhow. After all, the hope of parties in a permissive joinder situation is that the cost of suit will rise less rapidly than the value of the amalgamated claims, so that in union they will find strength.
But these negotiations will be fraught with difficulty for someone has to arrange for the pooling of expenses and recovery. Since we are, by hypothesis, still in a world of free bargaining, nothing compels any person to accept a pro rata share of expenses and gains to join the pool. It is possible for individuals to hold out for a larger share of the gain on condition of joining the business. The process could take place quite subtly as when one party insists on a minimum level of recovery out of the common pool, which leaves other people at greater risk and could induce aggressive demands that in total exceed the size of the pie.
What makes this problem so difficult to deal with is that holdouts, wishing to avoid social rebuke, often adopt concealed and not brazen approaches for their own aggrandizement. They assert that the claim really is worth a great deal more than anyone else thinks and calibrate their demands to the perceived value of their interest. Bargaining breakdown is likely in these circumstances, as in any common venture where all have to agree for any to proceed. We can think of the defendant, as seen through the eyes of members of the plaintiff class, as though he were a common pool asset (say oil and gas under the land of multiple landowners), and the defendants each as claimants to the some fraction of the pool. Often it happens that the surface owners cannot agree on any appropriate split of expenses and recovery, so each takes an independent course of action that leads to excessive costs of extraction whose necessary byproduct is the reduction of the total oil and gas taken from the pool. Permissive joinder in these cases can work in some circumstance. But often it does not. It is the failures that explain the rise of the class action.
To press the point on a bit, suppose that the question at hand is disputed behavior in a corporation, partnership, or some voluntary association. A common version of the complaint is that a key corporate officer has purchased in a cozy transaction a collective asset for a sum well below its market price. The remedy in this case is to unravel the transaction so that the asset is returned to the corporation and the cash to the individual buyer. (I ignore all complications with the time value of money, subsequent transactions and the like.) The question is: who is in a position to maintain the suit to undo this transaction? In principle the action belongs to the corporation, but notwithstanding its lofty legal status, a corporation has no powers of generation. Usually the directors act as “its” agents and in this case they have fallen asleep at the switch. So at least one shareholder has to step up to the plate for the corporation, AKA the other shareholders in question. So, how is this transaction organized?
One possible way is to think of a permissive joinder suit among shareholders, but this fails for a number of critical reasons. First, the chronic coordination problems can arise with shareholders every bit as much as it can with surface owners. Second, we have the nature of the relief in question, which is indivisible. As noted earlier the proper procedure is to unravel the transaction so that the thing is restored to the corporation, usually with its purchase price refunded to the insider. That form of relief benefits all shareholders whether they participate or not: it is not just the case as “a practical matter,” as Rule 23 says; rather it derives necessarily from the structure of the corporation itself. In this setting, we now have the worry that some shareholders will simply choose to free ride on the efforts of others. They will bear none of the costs of running the suit (and the consequent risk of failure). Yet, they will stand to gain equally with all other shareholders once the corporation has recovered the asset in question. The danger therefore in this situation is not that of excessive and ungrounded suits by rapacious class action attorneys. Unless something is done to fix up the imbalance, the real risk is that serious wrongdoing at the corporate level will go unchecked for want of a champion to respond to a common problem.
So at this point we can now see how in these core cases the class action operates as a system of forced exchanges that works for the benefit of the individuals who are subject to the state-generated coercion, at least with the derivative suit brought by shareholders in the name of the corporation, or its analogues for associations and partnerships. The basic logic is this: the knight who steps forward to maintain the suit is paid by the corporation out of the winnings of the action. This simple expedient has all the right features. By piercing the corporate veil, we discover that the champion has worked for the benefit of all the other shareholders. Since these individuals all hold fungible interests, we can treat the fractional interest in the corporation as their precise stake in the outcome of the litigation. It is not merely that the claim is, in the words of Rule 23 “typical” of those of other class members. It is that these actions are all “identical,” so that from a structural point of view we cannot conceive of a better class representative.
Once the action is successfully concluded, then the payment issue can be solved by ordering the corporation (i.e. the shareholders) to make an appropriate payment to the outside champion. Each member of the class bears the same fractional interest in the payment as he obtains from the successful recovery, so that this procedure divides up shareholder gains from the transaction in accordance with their respective investments. No one is allowed to opt out of this particular class—see Rule 23(c)(2)—which is just as it should be, because the nature of the relief—restoration to the corporation—works to the benefit of all even if they do not participate. But who is likely to complain about the mandatory inclusion, when the alternative is to get nothing at all?
Structuring the Class Action
Thus far it looks as though the derivative suit is the world’s perfect class action in that it forces all shareholders to deviate from their initial property holdings in ways that leave them better off than before. But while the overall conclusion is correct, the full solution requires us to tie up a large number of important loose ends.
Who leads the charge? The first question is: who should be the champion of the shareholders? On this point, Rule 23(a) federal rules specify that the class representative should be someone who has claims and defenses that are “typical” of those of all members of the class, such that the class representative (which means his lawyer) “fairly and adequately protect the interests of the class.” The question is how to make this definition operational. In the context just mentioned, the incentives are rightly aligned no matter who takes over the burden of the suit. So the real issue here is not whether this type of class action makes sense, but who shall control its operation. If only one party appears, then the problem more or less takes care of itself. But if the illegal transaction is evident from the face of the record, then many different shareholders and shareholder groups could have access to the information. If two or more individuals or groups try to run the single lawsuit as “the” representative of the class, then there is chaos on the plaintiff’s side, and an undue burden on the side of the corporate directors who are targeted by the litigation.
So who gains control over the suit? The Federal Rules do not tell us much about the selection of rival pretenders. Even if they did, as a matter of theory once we enter the world of the class action, we have to face in earnest the question raised earlier in this paper: who “owns” the cause of action? Here we cannot say that the ownership follows necessarily from the ownership of property because common property (i.e. the corporation) does not have a single owner. Put otherwise, all of these individuals have a claim in virtue of their shareholdings, so that the basic rule that gives each individual a cause of action for the violation of his rights does not tell us who (or whose lawyer) gets to represent all.
At this point, the earlier discussion that compares first possession and auction rules for individual claims comes back to influence the overall analysis. One possible way to deal with the issue is to hold that the control of the lawsuit belongs to that shareholder who is the first to file the suit in question: all others have to remain on the sidelines. But there is one critical difference between the use of capture rules to determine ownership of a seashell or a vacant plot of land and the “ownership” of the class action. The former holds the property on his own account and has no fiduciary duties to others. The “owner” of the derivative suit represents all individuals and thus bears just these fiduciary duties.
It may well be that the first possessor is a convenient owner who can be counted on to protect his own interests. It is a much more difficult inference to assume that the first shareholder (and the first shareholder’s lawyer) who makes it to the courthouse door is going to be in the best position to represent the interests of the class. Indeed one bad side effect of this particular rule is that it might encourage the emergence of legal “sooners,” those who decide to file first and think later, solely to preserve their rights to bring claims. This preemptive use of lawsuits to stake claims could easily have bad consequences on defendants who are forced to mount at least one round of defense, even if this particular suit turns out to be groundless.
On the other hand, the other shareholders (and by implication other class members more generally) could easily find themselves in an unhappy position if the initial claimant is ill-equipped to maintain a strong action against the corporation. All this is not to say that the first possession rule for class actions has to be rejected as a matter of course. The rule is very cheap to operate, and there is nothing that makes the first to file the worst potential representative of the class. Indeed, a priori there may be some reason to think that he will on average be a pretty good class representative. There well may be, for example, some partial voluntary organization among shareholders (who could be related persons who have organized a permissive joinder covering some but not all shareholders); there are costs that have to be borne even if the suit is so hopeless that it is not filed at all. So, on balance, the simple rule of first-to-file might work well in most cases. But the matter is still unresolved. One possible way of looking at it is this: can we find some alternative rule to deal with the question of claim control that is sufficiently better in its overall operation to justify the additional costs of its administration?
At this point, we now have to think seriously about the auction system alluded to earlier. The program, of course, needs to have at least two elements: first an auctioneer and, second, two or more bidders into the contest. One possible way in which the system would work is this. The role of auctioneer could be filled by the trial judge before whom the initial action is filed. A period of, say, 90 days is allowed after the filing of suit to allow other potential class representatives to enter the bidding. The first mover will be protected if no one else moves within that time. But what if a second entrant appears? At this point we can have an auction on fees, where the winner will be the firm that offers to take the lowest stake of the prospective recovery. The judge could then allow sealed bids on this question and award the suit to the lawyers who win the bid.
But here, too, we can never overcome a collective action or coordination problem without creating some alternative, hopefully smaller, problem in its wake. Ask yourself whether a single individual who owned all the stock would necessarily choose his lawyers by taking sealed bids on the percentage of the recovery that they would be willing to accept. The great fear here is that the low bidder might put in less work than the higher bidder who demands a larger share of the take. If the contingent fee lawyer who demands 25 percent of the whole only recovers $1,000,000, then the members of the class are worse off than they would have been if they had taken the bid from a firm that demanded 40 percent of the case but who would have, in light of the larger fraction of the gain, have worked hard enough to bring home $2,000,000. Most shareholders would rather have $1,200,000 over $750,000 in net gains. If so, then this lowest-fee bidding system could easily work to the benefit of the defendant.
Once again the nub of the problem is not all that hard to isolate. It is one thing to ask people to bid on assets that they will own outright once the sale is done. Then all the incentives, ex post the sale, are fully in place in virtue of the 100 percent ownership of the asset. But when the “ownership” of the claim involves both rights and liabilities, we cannot be so confident that the shareholders get the right behavior from the class representatives solely because of the low bids. So one might want to require for class actions what are done in bids for other projects. The cash bid is only part of a lengthier process. It may be preceded by a beauty contest, in which the applicants display their wares before the judge. Of those who make presentations, only some will be allowed to bid. The selection process thus works similarly to that for the selection of an investment banker or an architect, in which many firms submit preliminary proposals, from which some smaller number are chosen to make fuller presentations. Yet here again, each new iteration brings with it new problems. Beauty contests require judges, and it is far from ideal to assume that the judge who has to decide the case on the merits should be actively involved in the selection of the attorney. But, referring the matter to a shareholder committee raises other difficulties. Whobecomes members of the committee, and how do they decide which bid to accept?
Nonetheless, it appears that some combined process is gaining ground today. The most famous of recent auctions involved the civil action brought against Christie’s and Sotheby’s.2 In that case (where defendants’ collusion was stipulated) five established class action firms appeared before District Court Judge Lewis A. Kaplan in the Southern District of New York to announce that they had reached a consensus among themselves to divide the role of lead counsel. Judge Kaplan quickly rebuffed them and ordered an auction. He did not simply ask for the percentage that the lawyer’s were prepared to accept from the case. Owing to the defendants’ admitted liability, any fee on the first dollars was a free gift to the attorneys in question. Rather, with far more sophistication, he fixed the contingent fee percentage at 25 percent, and told the lawyers to submit one sealed bid which indicated the amount of money on which they were prepared to receive no commission. The operative formula was thus one-quarter of the difference between the total recovery and the bid figure. In this system, the highest exemption stated the winning bid. This formula allowed for easy comparison because it required a bid on only one number; and gave in effect a strong incentive to push hard for the highest amount. In that bidding war, and without reviewing the evidence available to the government, David Boies and his firm (Boies, Schiller & Flexner) came up with a winning bid of $405 million, with all the other bids clumped around $130 million. (The other bidders knew the winning bid, but the lawyers for the defendant did not.)
It should not be supposed, however, that Judge Kaplan’s system was foolproof. In fact, the procedure he adopted created a real conflict between the firm and its client. More specifically, this fee structure created the risk that the law firm might refuse a settlement that was in the interest of the clients to accept. Thus suppose that after investigation, Boies had concluded that there is a 50 percent chance that the final court award would be $250 million and a 50 percent chance of $500 million. (Assume that everything turns on the doubtful admissibility of one piece of evidence, or on the choice of rules in calculating damages). Under a fee arrangement that allowed the plaintiff’s lawyers to recover, say, 10 percent of the total award, the lawyers would take a settlement of $400 million, which their clients would want because $400 million is greater than the $375 million they could expect to recover, even before litigation costs are taken into account. Their fees would now be $40 million instead of $37.5 million less trial expenses. But under Judge Kaplan’s rule, Boies would reject this offer because it would have yielded his firm nothing even as it gave the class a favorable settlement. Owing to the chosen fee structure, it would have been prudent for them to roll the dice, at which point they could recover $25 million in fees half the time, for an expected value of $12.5 million.
So the conflict of interest is not eliminated by this fee system, even with the likelihood of a high minimum settlement. In fact, Boies’s team turned down an initial $100 million offer from Christies, from which they would of course receive nothing for their pains. In the end the entire suit settled for a sum of $512 million from both auction houses. Putting all the complications aside (including criminal prosecutions, the use of script, the payments from Sotheby’s chairman), the firm received a commission of $26.7, or about 5 percent of the total, far less than the usual amount. In this case, the bidding system worked wonders for everyone but the defendants. Clearly, the terms on which the auction are conducted matter.
So, too, who conducts them. Thus, a third mechanism might well be preferable to either of these two. It involves holding the question of class representation open until each of the potential class representatives can round up votes from shareholders. The class representative who wins the election gets to guide the litigation. One advantage of this procedure is that it allows all the parties to make whatever pitches they choose to the shareholders directly. In the end, once the information is assembled, it can be distributed in preparation for the final vote. It, therefore, skirts both the problem of excessive judicial involvement in the organization of the plaintiff class and the related difficulty of using low bids to determine class representation.
Once again, however, do not think that we can eliminate one nagging problem without creating a second. Thus, to mention but one issue: will the disappointed applicants be allowed to challenge the outcome of the vote on the ground that the winner had made false representations about his intentions or his rival’s ability, during the course of the election? The point here is that there is no simple way in which the question of representation can be resolved. All systems will require some judicial supervision of the process, which may incline toward one solution in the one case and another solution in another, depending in part on what is known about the soundness of the case at the time the auction is conducted. The root problem goes back to the theory of property rights. We know that each shareholder owns his or her own shares. But we cannot deduce from that which of these shareholders (or their eager attorneys) is entitled to control the class action.
Let us now assume, however, that the question of representation is solved by one means or another. Then we have to face three additional questions: first, what are the effects of the class action on subsequent litigation; second, what determines the compensation for the lawyers who obtain cash or noncash relief to the shareholders; and third, what substantive principles should govern the underlying litigation? As with all collective action problems, it is easier to state the difficulties than it is to solve them.
The Effects on Subsequent Litigation: Res judicata. The first question is one that involves the interaction of the law of res judicata with the law of class actions. The general rule of res judicata is a salutary principle of finality. Each and every cause of action should be litigated once and only once on the merits. To block any litigation at all is a clear denial of due process to the plaintiff. To require the same case to be litigated more than once on the merits adds unnecessary costs to the operation of the legal system without offering any assurance to either side that the outcome in the second trial is any better than that in the first. So, the basic instinct of the legal system is to make sure that one trial is done well, after which the parties and their privies are bound. (This last qualification is needed to make sure that the cause of action is not split between parties. If the trustee has lost the suit against the third party, then his beneficiary should be bound by that initial litigation, for otherwise the beneficiary has the incentive to hang back at the outset and then intervene if and only if the trustee loses the first suit. No legal system should encourage the proliferation of suits by several persons who have partial interests in the same basic claim.)
The situation is, however, a bit more complex with class litigation because, by definition, the interests of the individual shareholders are all separate and distinct from each other. Normally, one shareholder could not decide how to litigate the claims of another. But, in this case we have the odd circumstance that the relief in question is a restoration of the property to the corporation from which all shareholders will benefit. It would, therefore, become intolerable to allow them to obtain the benefit of a successful action, but to stand aside and start over if the initial suit fails, at which time the first round of claimants will benefit as well. Note in this particular instance—the niceties of corporate law aside, which raise special restrictions on derivative actions—that there is no real difference between a derivative suit that asks for a restoration to the corporation and one that asks for a dividend distribution to a class of shareholders. Here, too, the corporate rules require that all shares within the same class receive the same dividend, so that relief for one has to be followed by relief for all. So, once again, even those shareholders who wish to hang back from the first action are bound by it for the simple reason that they will profit from its successful prosecution even if they are not part of the suit.
Determining Attorney Fees. The larger controversy over class actions comes in connection with the fees awarded to the attorneys for the successful class. The issue has, moreover, genuine ramifications for class actions that do not fall within the derivative or corporate mode as well, but are in both cases conveniently discussed here. As an ideal matter we can refer to the same test used to frame the discussion of res judicata. The ideal way to think about the matter is to ask what fee would have been negotiated if the class contained a single member who was able to negotiate a fee with his lawyer.
Just putting the problem in that fashion indicates some of the difficulties that are involved. One possible fee arrangement is for the informed client who is capable of monitoring his lawyer’s work to pay an hourly fee. The advantage of this arrangement is that it allows the client to capture the full upside of the lawsuit should the outcome be successful. But whatever its good sense in ordinary litigation, it is stillborn in class action litigation for the simple reason that diffuse class members have no real way to monitor the behavior of their lawyers. The great danger is that lawyers will pad their bills and leave the shareholders (or other class members) empty-handed.
This inability to monitor the conduct of lawyers is not unique to class action members. Most ordinary individuals who pursue a tort case are unable to assess the effort and quality of their attorney’s work. Hence, the usual arrangement here is to have a contingent fee arrangement in place whereby the lawyer recovers fees only if successful in bringing the suit. That partial alignment of incentives removes some of the strain from the client, but it is by no means a perfect response to the question of whether the attorney is a “perfect” agent for the client, that is, one who acts just as he would if he owned the entire claim himself. It is well known that in some cases the ordinary contingent fee lawyer will settle a case sooner than might be in the best interest of his client, and do so in order to reduce the costs of additional litigation (which fall largely or entirely on him) while the benefits of this action are shared, with the lion’s share going to the client.
We are, therefore, back in the usual situation where there is no ideal solution to the question of how to compensate class action lawyers. I think that we can be confident that it is suicidal to allow the lawyers to collect fees on an hourly wage, so in the corporate context the only question is how the corporation (or the individual defendants) compensate the lawyers once their case has proved successful. Here we can treat the money restored to the corporation as part of the common fund from which recovery could be had. The only question is the formula.
Class action litigation lawyers have been able to devise only two approaches to deal with the question of lawyers’ compensation. The first, or lodestar method, represents an effort to build up the fee on a cost-plus basis, taking into account the hazards of litigation. It is a procedure not unlike the rate-making processes sometimes used to set rate schedules for public utilities in regulated industries where, again, no competitive market is available to determine price. The first step in the process is to determine the number of hours that the lawyers spent on the case. As ever, useless hours are cast out of the calculation, so that only useful hours, reasonably computed, so remain. Naturally a host of subordinate issues are raised as to how the judge reviews the hours and the like. But, once the number of hours is determined then the base, or lodestar, rate is developed by multiplying the number of hours for each lawyer (or paralegal etc.) by their customary billing rates in those transactions where they charge customers by the hour. Once that figure is determined, the Court adds in a risk factor to offset the fact that class members owe nothing to the lawyers when no recovery is obtained. The riskier the litigation, the higher the rate of the adjustment. In principle, a lead-pipe cinch would have a multiplier of 1; some cases have had multipliers of about 1.5. Multipliers of 2 are common in the business, and from time to time it has been suggested that 2 should be the maximum figure used. Multipliers of up to 3 are not uncommon, and occasionally the number could reach 4. In addition, some interest figure has to be added in for delayed payment.
There are, however, genuine difficulties with the lodestar method for no one has a firm handle on how to count hours, how to price hours, and how to assess the riskiness of litigation. My own preference therefore is to stay away from this cost-plus formula with its intrinsic risk of padding the accounts, and to veer to the same kind of arrangements that ordinary lawyers use in tort actions: the contingent fee. Here the costs incurred for the solution are the lawyers own business; what matters are the results that are obtained for the client as a result of that work. But, once again, this general endorsement only conceals a thicket of problems that have to be worked through in fee deliberations. Contingent fee arrangements are complex agreements that to some extent, at least, should vary in accordance with the parameters of the underlying case. To see the problem here, consider two cases: one in which the question of liability is virtually conceded, and the second in which liability is hotly contested. The ordinary contingent fee lawyer may well demand higher compensation in the second case than in the first case, to take into account the nature of the risk. Certainly if the defendant has offered the individual plaintiff $10 million to settle the case before the intervention of the lawyer, no client would be well advised to give the lawyer 33 percent or more of the first dollar in settlement. (This is in reality a variation of the situation with Christie’s and Sotheby’s). Rather the optimal schedule is one that would try to give a larger percentage of compensation to the recoveries in excess of the $10 million.
Unfortunately, the contingent fee arrangement in class actions cannot be tailored by negotiations prior to the litigation. Since all shareholders in the context of derivative actions are party to the same suit, we do not have any parallel agreements between lawyers and their individual plaintiffs to give us some clue as to the proper design of these fee arrangements. Worse still, we do not have the possibility of making the judgment on the fee structure before the fact, as is necessarily done in ordinary contingent fee litigation. The inability to negotiate the contingent fee in advance therefore creates some level of skepticism in the judges who pass on fee arrangements generally. Occasionally, they do receive some guidance when individual plaintiffs opt out of the class, as they can do for the so called (b)(3) as in Rule 23(b)(3) class actions. See Rule 23(c)(2). But sometimes that guidance may not be precise, and in other cases it may prove unavailable. So judges hedge their bets: if they prefer in principle the contingent fee ideal to the lodestar, they will not place all their eggs in one basket. Rather, they will take at least a sideways glance at the actual work expended so as to avoid making the attorneys’ fees too rich for their own good. Ordinary contingent fee arrangements typically run to one-third, forty percent, and in some difficult cases even as much as one-half. The usual fees in class actions top out between 20 and 30 percent, and are frequently lower.
It has been suggested that the fee in question be “tapered,” so that the percentage take is reduced with an increase in the size of the class settlement or verdict. In general, however, this does not seem to be the right approach. It is precisely the opposite of the form adopted, for example, by Judge Kaplan in setting up his auction of the law suit against Christie’s and Sotheby’s. Similarly, there are few, if any, instances in which contingent fees are tapered in ordinary litigation. Quite the contrary, the home-run settlement might be a reason to increase the fee of the lawyers whose enterprise generated the desirable outcome. What does make sense, perhaps, is a rule that follows ordinary litigation by reducing the contingent fee for cases that settle prior to trial, relative to those which require trial, or in some cases an appeal. It is very difficult to replicate after the fact what the ideal agreement should look like before the fact. In principle we would hope that the terms of the fee arrangement would duplicate those that are obtained from a voluntary action. But the blunt truth is that there is simply no way to recreate the ex ante environment in which ordinary contingent fee arrangements are negotiated. So, judges do the best that they can, because the alternative is to quash the most realistic possibility of relief in the corporate context.
Substantive Principles Shaping the Litigation. The next point of vital concern involves the interplay between substantive and procedural law in dealing with class actions. As noted earlier, treating the class action as a procedural rule carries with it profound implications. Quite simply, the class action functions solely as an aggregation device to allow the pursuit of claims that could otherwise not be brought because of the high rate of administrative cost relative to the anticipated recovery. In principle, therefore, we should hold the substantive law constant regardless of whether the plaintiffs proceed by individual action, permissive joinder, or class action. Thus, the class format does not alter the terms of the basic cause of action; nor does it introduce some new defenses, or eliminate others, in the prosecution of the case. The whole point here is to avoid any extraneous influence that would give parties a reason to either bring or refuse to bring a class action. The substantive outcomes should not be distorted by the choice of procedural vehicle.
This invariance constraint has powerful roots in other substantive areas of law. One common theatre in which the point is raised is bankruptcy. Ordinary firms have relationships with multiple creditors antecedent to and outside of bankruptcy. The ideal bankruptcy system allows for the coordination of multiple claims, the marshaling of the defendant’s assets, and a key decision over whether to liquidate or reorganize the basic business. These issues are hard enough to resolve in their own right, and the ideal set of procedural rules is not one that induces parties to go into or to avoid bankruptcy solely on the grounds of the relative procedural advantages of the various fora. Thus, it would be quite dangerous if the legal position allowed a plaintiff-creditor to defeat a statute of limitations defense available in state court by filing for bankruptcy. At this point they may well choose an inefficient place to litigate in order to gain a partisan advantage. Defendants will have equal and opposite incentives, and the whole system could easily grind to a halt, for in both cases the private advantage creates a social disadvantage.
Likewise, in dealing with the regulation of private land use disputes between neighbors, it is important to keep parallelism between the ordinary tort actions that some neighbors can bring against another and the actions (often class actions) that the state can bring on behalf of some neighbors against others. Let the state be given substantive or procedural advantages not available to the individual plaintiffs, and enforcement will migrate into public hands even if the private law offers systematically superior substantive solutions on the issues. Likewise, if the private law is systematically more advantageous to plaintiffs, they now have an incentive to resist more efficient class actions solely to obtain partisan advantages. In all these cases, then, we should be careful to see that the amalgamation of claims does not alter the balance of power between the two sides except insofar as it overcomes the transactional obstacles that justify the use of the class action in the first place.
This program is, moreover, easy to implement in connection with the ordinary derivative suit against private associations and corporate defendants. Here all the plaintiffs are in precisely the same position, so that the court need only ask itself how it would resolve the suit if all the shares in question were held by a single person who had an action against members of the board of directors. But the concerns here become much more serious in the modern class action to which we now turn.
The Modern Class Action
We are now in a position to turn to more controversial aspects of the class action, which arise outside the corporate and association situations where what is sought is a restoration in cash or kind to the association or corporation. Right off the bat it should be clear that the efficiency of the class action is necessarily reduced as it is carried over into these new situations. No longer do we have any fungible corporate shares that certify the indivisible nature of the class relief and the parallel nature of the individual claims. In addition, it is no longer necessary to limit the class suit to a single (corporate) defendant, which increases the complexity of its administration. The first of these elements falls to the wayside because there is no legal entity to which the damages in question can be paid. When someone is run over by a truck, gouged by a monopolist, victimized by discrimination, consumer or securities fraud, he sustained his loss in his individual capacity. The demand is for cash relief, to be paid to each person separately. These claims, moreover, may be analogous in some ways but different in others. The individual plaintiffs may have been sold goods at different times, for different prices, by different salesmen. Shareholders may have bought and sold stocks at different times in the period before a takeover bid is announced. Often the claims involve individuals in different states with different substantive law. The claims may have some elements in common, but also some important differences. Often the salience of these differences and similarities may not be fully apparent at the outset of the suit, but only become apparent once discovery has been undertaken, or perhaps even at trial.
These two additional sources of complexity do not eliminate any of the difficulties of class administration in the context of derivative actions. But they do add a number of additional elements that require some closer examination. Owing to the want of parallelism, the question arises of whether all persons similarly situated—itself a term of art—must become members of the class whether they want to or not. At least one strand of thought, championed most conspicuously by David Rosenberg, claims that this approach is not a bad idea, and holds, in effect, that the conscription of individual plaintiffs into the class action really works to their benefit, such that they have no reason to opt out of the class to control their own suit. The law can make that judgment for them at lower cost and higher reliability. Indeed in one sense his position goes a step further. Since the real question is deterrence of defendants, he takes the view that there is no particular reason to want to distribute the money to any members of the victim class. Their protection comes in a different form: stronger deterrence reduces the occasions on which compensation is required.
Even if we reject (as current law manifestly does) the view that ex post deterrence is irrelevant, powerful implications still flow for the governance of class action litigation. This position presupposes that the judgment should be collective and not individual, such that a person who objected to the strategies pursued by the class would be required to remain a class member on the ground that the economies of scale in running the class action would leave him better off than before. There is obviously a powerful paternalistic streak in this argument. Surely a consumer class that has 100,000 potential members could operate if some fraction of them decided to opt out, perhaps to form a second class under different leadership. And, it becomes hard to insist on their participation in the class if these dissenters have fundamental strategic disagreements with the lawyers and/or class committee that takes direct control over the litigation. It may well be too expensive to try to recruit individuals into the class, but the transaction costs do not preclude a default position that preserves the individual right to opt out on receipt of notice. To their credit, the current class rules appear to recognize the difference. The rules contain no provisions for opt-outs for (b)(1) and (b)(2) class actions, but recognizes them for class actions brought under 23(b)(3): “the court shall direct to the members of the class the best notice practicable under the circumstances, including individual notice to all members who can be identified through reasonable effort.” Newspaper notification will not work when names and addresses are available.
Stated otherwise, the situation with individual claims differs fundamentally from the derivative suit in that it is no longer the case that the provision of a remedy to one person necessarily provides a like remedy to another. When one person opts out of the class to bring his own suit for money damages, all other individuals may proceed under the class rubric if they please. The key point here is to make sure that those who hang back to do not get the benefit of the offensive use of res judicata should the class action be successful, while reserving the right to bring their individual suits anew should that action fail.
Once individuals are allowed to opt out, they must be able to receive some notice, by publication or in person, about the terms and conditions under which the class action will proceed, as Rule 23 provides. In many cases where the individual sums for the class are small (as with the miscalculation of interest rates on small personal loans), most people will choose to stay put, assuming that they pay any attention to the matter at all. But, nothing about the current structure of the rules of civil procedure limits ordinary class actions to small overcharge cases. Huge tort actions and substantial antitrust claims may also be brought in this form, and here the choice whether to opt out is far weightier. In these cases, moreover, individual plaintiffs may well decide to commence their suit before any class action could begin, so that it is highly doubtful that a plaintiff should be bound to the class unless at the very least he receives actual notice of the suit, and probably not even then unless he agrees to a stay of his own litigation pending the outcome of the class action. After all, if two individuals brought suit, neither would be stopped in his tracks simply because he had notice of the other suit. Some evidence of collusive or opportunistic behavior would seemingly be required.
The issue of class membership has, moreover, important consequences for the defendant, although it is difficult in the abstract to say which way they cut. On the one hand, corralling all the plaintiffs in an individual class action reduces the litigation costs for the defendant, and avoids the possibility of follow-on suits (from which the plaintiffs can learn from earlier strategic mistakes even if they do not have the benefit of res judicata). On the other hand, the creation of a huge nationwide class makes it impossible for a defendant to diversify its litigation portfolio. The litigation may easily assume “you bet the firm” proportions. The high stakes may well induce some juries, and indeed some judges, to adopt a compromise position that is ruinous to the defendant’s interest, and the possibility of error in so complex a lawsuit is something that an innocent defendant should greet with dread. After all, a ten percent exposure to a $10 billion verdict counts as real money, even today.
Defining the Class
The amalgamation of individual claims raises yet another question of no simple proportions. Which claimants, holding which claims, should be eligible to participate in a class action in the first place? That problem was solved almost by definition in the derivative suit because each plaintiff occupied a position that was largely indistinguishable from other members of the class. The issue of class membership is solved in the ordinary action by making each shareholder a member of the class to the extent of his own interest. But with separate claims the matter becomes far murkier. The key trade-off is easy to state. The gains from amalgamation increase as the claims are more similar to each other; but these economies of scale are much reduced to the extent that individual claims differ from one another on some material point.
In 1966, the Federal Rules of Civil Procedure made a conscious effort to liberalize the scope of the class action by adopting a posture that at bottom asks whether “the questions of law or fact common to the members of the class predominate over any questions affecting only individual members.” (Rule 23 (b)(3)). That requirement is generally satisfied in cases of antitrust violation where all sales within a given period were made as part of the same business scheme to the same set of plaintiffs. The level of perceived overcharge (assuming that the members of the plaintiff class have standing to sue, which may be problematic in some cases of “indirect purchasers”) is roughly constant so that once the difference between the monopoly and the competitive price is determined for one party in one transaction, then it is largely determined for all. This argument presupposes that a single scheme controlled multiple separate transactions, such that the outer limits of the class could well be sensitive to changes in the defendant’s pricing policies or its relationships with other firms in the industry. And there remains the constant gnawing problem that distinct state law claims may well be governed by different laws that make their amalgamation harder to justify. But in general these cases will be amenable to some level of class formation. Even if all potential plaintiffs do not fit snugly within the confines of a single class, it is easy to imagine a couple of subclasses that will cover the vast bulk of cases.
The problem of class actions becomes much more difficult in dealing with tort claims, including mass tort claims. The original notes to the 1966 Federal Rules of Civil Procedure made it seem doubtful that any ordinary tort claim could be subject to class actions. “A ‘mass accident’ resulting in injuries to numerous persons is ordinarily not appropriate for a class action because of the likelihood that significant questions, not only of damages but of liability and defenses to liability would be present, affecting the individuals in different ways.” Thus even if two individuals were hit by the same car at the same time, the issues in the two cases could overlap and not be precisely the same. Here much could turn on the theory of liability. If liability were strict, so that the only question was whether this defendant hit both plaintiffs, then the issue could easily be common between the parties. But if liability is based on negligence, then the defendant might have been negligent with respect to the plaintiff in plain view but not with respect to one that was not within his line of vision. Or some jurisdictions could adopt a principle of “transferred negligence,” such that the defendant who was aware that one plaintiff was in the field of danger, is liable in negligence to a second defendant who could have been spared injuries if the defendant had taken only those precautions needed to deal with the plaintiff in plain view.
The situation with mass torts, of course, only becomes more difficult when the defendant has engaged in a similar line of business over a long period of time, such as the selling of asbestos or a pharmaceutical product. In these cases we lose the Aristotelian unities of time and space, so that one might think that only rarely would the class action be appropriate in suits of this sort under the Federal Rules. But in this case a set of ambitious certifications in a wide variety of cases, moving from blood transfusions to cigarettes, indicates how the law has migrated from initial cautious attitudes in these cases to a far more aggressive stance. A similar migration can be found in cases involving misrepresentations, where the 1966 attitude toward misrepresentations, which was prepared to allow many actions where the separate cases had a “common core,” but not in those instances where “although having some common core, a fraud case may be unsuited for treatment as a class action if there were material variation in the representations made or in the kinds of degrees of reliance by the persons to whom they were addressed.” The current attitude seems to be that even if the common issues do not dominate the law suits, then the appropriate response is to use the class action for those issues that are common and, thereafter, to allow the cases to be tried or settled separately: strategic advantage to the plaintiff.
This brief discussion shows how difficult it is to decide at the outset of the lawsuit whether the common issues are sufficient to dominate the separate ones. What is the cart and what is the horse? Normally we would like to know whether plaintiffs are similarly situated before we are called upon to decide what legal theory is relevant in their case. But now it looks as though we cannot decide whether two or more claims are dominated by common issues until we decide which theory of liability is invoked. The problem, moreover, only gets worse when the plaintiffs seek to pursue class actions and unities of time and place are not strictly observed.
For example, in product liability actions the defendant may have sold a given drug in different tablet sizes with different warnings in different locations over different times. It could be that the jurisdiction in question uses a strict liability theory for any defects, at which point the variations in the level of care may not matter. But, if the action relates to the duty to warn, a warning that was effective in 1980, when the knowledge base was more limited, may turn out to be insufficient in 1990 when the level of public knowledge became greater. The situation gets no easier if it turns out the role of intermediate parties, or of the plaintiffs themselves differ in some material way that is found relevant to the law suit.
I think that the level of common elements in many mass torts are far less dominant than others believe, and I would follow, as most federal courts today do not, the lead of the 1966 Federal Rules in presumptively denying class actions in most mass tort cases. I realize that this result may place some plaintiffs at a serious disadvantage, in that the defendant is geared up with a standard form defense that it tweaks in each individual case. But even in these cases, the issues of defendant’s liability, causal intervention, plaintiff’s knowledge or misuse, and plaintiff’s damages could all differ from case to case, so that the standard form defense need not be all that standard. It remains possible, of course, for plaintiffs to form a voluntary loose alliance in which they share information about common issues while controlling their individual causes of action. But it is much riskier to follow the pattern of amalgamation, for once the cases are together then the individual differences in the plaintiff’s cases will be bled out of the equation, so that all suits will appear to be cut from the same cloth. As the choice of forum will normally lie within the control of plaintiffs, it is likely that the substantive law will drift in their favor.
This integration between substance and procedure is evident in yet another sense. It leads courts in close cases to adopt that version of the substantive law that facilitates class action suits. To take one notable securities case, Basic Inc. v. Levinson (485 U.S. 224 (1988)), that involved the question of whether the officers and directors of the corporation had violated the provision of Rule 10b-5 relating to the publication of misleading information, in this case a false denial that the firm was engaged in potential merger negotiations, when in fact it was. One effect of this denial was arguably to lower the price of the shares so as to induce class members to sell before the merger was formally announced. In an ordinary action for common law fraud, the plaintiffs must prove that they have relied in specific transactions to their detriment on the false statements made by the defendant. But, in this case the defendants were not sellers of the shares; nor did they make any specific statements to identifiable purchasers. If each plaintiff had been forced to show his own reliance on these false statements, then it would be impossible to keep the class intact. But once the Supreme Court accepted a “fraud-on-the-market” theory, which presumes that efficient capital markets quickly embed all false information into the price, then the element of reliance flips over from a separate to a common issue, allowing the class to hold together. I have no doubt that one reason why the Supreme Court embraced this substantive theory was to allow the use of class actions in securities case. Yet even here its conclusion can be criticized on the ground that the presumption of reliance is at most rebuttable, so that the defendant could try to show, on a case by case basis, that individual plaintiffs had disbelieved the information when published. But rebuttable evidence is admissible only in a small fraction of cases, so that even the shift in the burden of proof allows the class action to go forward.
There are other situations in which the use of class actions strikes me as far more controversial, and I shall close by mentioning how the disguised use of class action litigation dramatically influenced the outcome of the tobacco litigation. Before the advent of the class action, individual smokers had a difficult time persuading juries to allow their recovery for tobacco-related illnesses. In these cases the winning line of defense usually relied on the integration of two separate techniques. First, the defendant would show that tobacco caused harm only in the sense that it created an increased risk or hazard of illness—one that was often overestimated by smokers. Thereafter, the defendant concentrated on the assumption of risk defense in both its objective and subjective guises. In its objective sense, the defense depended on showing that the risks of smoking were common knowledge to the public at large, including all smokers. The subjective version relied on the constant reminders that each individual plaintiff had received from countless sources warnings of the risks and hazards associated with smoking. The one-two punch was in general decisive.
The great break in the tobacco litigation came when the plaintiffs bar figured out how to get the assumption of risk issues out of the case. What they did was to bring independent actions on behalf of Medicaid for the medical expenses incurred in treating tobacco-related illnesses. By calling these actions independent, they were able to amalgamate large numbers of individual actions into a class-like action brought on behalf of Medicaid programs: hence the reference to disguised class actions. The assumption of risk defense disappeared because the suits were not brought by individual smokers. This procedure should be contrasted with the correct way of doing business, which is to recognize that the medical providers in these cases as everywhere else, can only proceed by subrogation, that is, by taking the right of action for medical expenses by assignment from individual smokers. At this point, the general rule that all assignments, including those partial assignments by subrogation, are subject to affirmative defenses, including, of course, the assumption of risk defenses that are available against the assignee. Under this view of the world, the Medicaid suits are nothing more than concealed class actions in which (it appears) that the separate issues predominate over the common ones. The entire stratagem wholly transformed the tobacco litigation because the plaintiffs were able to do an end run around the invariance requirement, that is, the requirement that class actions rely on the principles used in individual suits.
This capsule account of the tobacco litigation gives us, I think, the critical hint as to why the use of class action should be greeted with such mixed emotions. There is no question that in some contexts it allows plaintiffs with sound substantive claims to gain access to the courthouse that would be denied to them without some method of amalgamation. The class action can be the key for taking the disorganized business of life and structuring it in simplified ways that permit mass adjudication. How could anyone such as myself, who authored a book called “Simple Rules for a Complex World”, be opposed to that development? Yet notwithstanding, I have genuine ambivalence about the use of the class action. In some cases it seems that the mere whisper of class action induces courts to expand the scope of liability in order to allow plaintiffs to amalgamate claims by one device or another. In many cases it is hard to pin down exactly where the process has gone wrong. But sometimes it looks as though marrying the aggressive use of class actions with the aggressive extension of substantive theories of liability creates real negative synergies.
Quite simply, the class action serves as an amplifier for the ordinary principles of civil litigation. Where those are correctly announced, then the class action increases their effectiveness. Yet when these are incorrectly stated, then the class action increases the mischief that these new actions can bring. So, in the end the class action serves as a giant megaphone that amplifies both the strengths and weaknesses of the underlying system of substance and procedure. It is precisely because this effect will be benevolent in some cases and harmful in others that we cannot make a uniform assessment of the overall effects of class action practices. All that can be confidently said normatively is that the more the class action conforms to the older models used in derivative suits for corporations and voluntary associations, the sounder they will be. Yet, we should be confident that most of the recent innovations in class actions have tended to deviate from those ideals, which is why the question of the relative proportions of use and abuse in class actions remains, as of yet, unresolved.
- Ronald Dworkin, Sovereign Virtue: The Theory and Practice of Equality 65-71 (2000).
- On which see, James B. Stewart, Bidding War: How an antitrust investigation into Christie’s and Sotheby’s became a race to see who could betray whom, The New Yorker (of course) October 15, 2001, at 158.