Duke Journal of Comparative & International Law: Thomas D. Rowe, Jr., Shift Happens: Pressure On Foreign Attorney-Fee Paradigms From Class Actions, 13 Duke J. of Comp. & Int'l L. 125 (2003)

Cited: 13 Duke J. of Comp. & Int'l L. 125
[*pg 125]

SHIFT HAPPENS: PRESSURE ON FOREIGN ATTORNEY-FEE PARADIGMS FROM CLASS ACTIONS

THOMAS D. ROWE, JR.^*

I. INTRODUCTION

II. THE UNVIABILITY OF CLASS ACTIONS WITHOUT CONTINGENT FEES AND UNDER LOSER-PAYS ATTORNEY-FEE SHIFTING

III. COMPENSATING THE PLAINTIFF CLASS'S COUNSEL
   A. Liability of Unnamed Class Members
   B. Third-Party Funding and Contingent Fees

IV. LOSER-PAYS ATTORNEY FEE SHIFTING
   A. Liability of Class Counsel
   B. Commercial Insurance
   C. Public and Third-Party Private Funding
   D. Modifying or Abandoning Loser-Pays

V. CONCLUSION

FOOTNOTES

"You know what they say about paradigms -- shift happens."¹

I. Introduction

The class-action device, still unique in its degree of prevalence in American civil litigation,² has begun to spread -- some would no doubt say metastasize³ -- beyond the United States. Five Canadian provinces now have some form of the device,⁴ as do the Australian [*pg 126] Federal Court and the state of Victoria.⁵ For over a decade Brazil has authorized collective actions on behalf of private parties by designated government offices and by private associations with relevant institutional purposes.⁶ The People's Republic of China has since 1991 provided for class-action-like representative lawsuits,⁷ and the Indonesian Supreme Court recently adopted a regulation to authorize and govern class actions.⁸ A Swedish law authorizing class actions was to take effect January 1, 2003,⁹ and Finland, Norway, and Scotland have considered or are considering adoption of the device.¹⁰ The South African Law Commission in 1998 recommended recognition there of class and public-interest actions.¹¹

This essay does not discuss these developments in depth; nor does it wade into whether, and if so how, other nations should implement class actions. Rather, it explores the tensions between the class-action device and norms governing attorney-fee liability and class-action financing practices in most of the world outside the United States, the pressures resulting from those tensions, and [*pg 127] possible resolutions. Given differences among legal systems that have, or might adopt, class actions, the essay also largely avoids arguing for or against particular choices that might be made. I hope that this exploration can, by clarifying some of the issues that are likely to arise and the alternatives available, be useful where adoption or modification of the class action is being or may be considered -- perhaps even including in the United States.

II. The Unviability of Class Actions Without Contingent Fees and Under Loser-Pays Attorney-Fee Shifting

As others have long recognized, class actions could find barren soil if they were transplanted to systems that, like much of the world, maintain bans on contingent fees for plaintiffs' lawyers and adhere to the near-universal loser-pays rule on liability for a winning side's attorney fees.¹² The American class action exists in a system under which losing plaintiffs are rarely, and losing defendants only sometimes, liable for the attorney fees of their victorious adversaries;¹³ contingent percentage fees are allowed and are the dominant means for financing plaintiffs' non-class and class damage [*pg 128] litigation;¹⁴ and entrepreneurial plaintiffs' lawyering -- with attorneys often being the main impetus behind, and principal persons financially interested in, a class action for damages -- is at least tolerated.¹⁵ In nearly all the rest of the world, prevailing practices and attitudes hew in varying degrees to an opposite paradigm in which losers in civil litigation are usually liable for a substantial portion of winners' reasonable attorney fees (loser-pays or the English as opposed to the American rule¹⁶); contingent fees -- percentage or hourly¹⁷ -- have been frowned upon, with the client at least in principle obligated to pay the lawyer the same rate no matter whether success be great, small, or nil;¹⁸ and lawyers' financing of and stakeholding in litigation have tended to be regarded as unacceptably commercial and unprofessional.¹⁹ Moreover, financing of litigation by third parties [*pg 129] aside from clients and their own lawyers could run afoul of the traditional common-law barrier to "maintenance."²⁰

American practices are in several ways obviously hospitable to the flourishing of plaintiffs' class actions for damages. While small chance of a significant recovery may (or at least should) deter the pursuit of class claims, individual class representatives' or class members' fear of down-side liability for large defense fees will not. Nor need the class representative or members worry about paying the class's own attorneys in case of small or no recovery; the contingent fee means that class counsel bears the risk of nonpayment. And class counsel's substantial percentage stake in a large recovery, even one made up of many awards too small to be worth pursuing in individual actions, can make it financially attractive for entrepreneurial attorneys to front the considerable costs of class litigation if the chance of eventual reward appears great enough.

By contrast, the contrary paradigm -- which has prevailed in varying degrees in civil-law and non-U.S. common-law systems alike -- would discourage significant use of plaintiffs' class actions in not one but several ways. Whatever the economies of litigation in class form or the strength of a class's claims, the class device increases stakes and likely costs on both sides,²¹ threatening the one or few class representatives with down-side liability for their own side's and the adversary's fees that can be disproportionate to their share in a likely class recovery.²² As Professor John Coffee has nicely put it in another context, the prospective plaintiff may be looking at -- and [*pg 130] understandably reluctant to enter into -- a "one horse, one rabbit" trade.²³

The disincentives can work somewhat differently, albeit with similar effect, whether class members' individual claims are large or small. A small claim, even a strong one, may not be economically viable on its own and therefore unlikely to be pursued in an individual action.²⁴ Yet despite the purpose of class actions to make such claims economically viable when pooled,²⁵ it may be hard to find a willing representative with such a claim to take advantage of the pooling benefits of a class action, given the threat of large costs in the event of defeat or even quite small recovery; hence such claims will tend to be pursued if at all in individual rather than class form.²⁶ And the client who has an individually viable larger claim, while more likely to proceed in the first place, will most probably prefer to pursue the claim on its own rather than in a class action, if the client's recovery will be about the same in the event of success in either form but with the down-side fee liability risks smaller if the claim proceeds by itself.²⁷

[*pg 131]

These kinds of risks to individual clients, to be sure, can lead to some form of risk-pooling or performance of an insurance function. That is precisely the role played by an entrepreneurial American plaintiffs' class-action law firm working for a contingent fee, typically handling a portfolio of multiple cases so that success in some can subsidize those that yield little or no recovery and finance cases still in process. Under the contrary paradigm -- which of course may not exist in pure form anywhere,²⁸ and which has been diluted in some systems in recent years²⁹ (but the existence of pressures for such changes is precisely a main point of this essay) -- the risks are greater than under American approaches and the means of dealing with them more limited. Not only would the loser-pays rule pose the threat of a class plaintiff's being out of pocket for the other side's large fees in the event of defeat; but a ban on no-win, no-pay contingent fees would keep the plaintiffs' attorney from bearing the risk of no or small recovery and spreading that risk across multiple cases -- thus im-[*pg 132] posing the risk of being liable for the probably large fees of class counsel upon individual clients who are less equipped and likely less disposed to take it on in the first place.³⁰ In short, the realities of litigation finance and risk allocation under the contra-American paradigm would make for little if any class damage litigation, except possibly when success and substantial recovery looked virtually certain -- and even then the incentive structure might lead to the choice of litigation in individual rather than class form.³¹

For damage class actions to be viable, then, something -- or more than one thing -- would have to give. The rest of this essay explores possibilities for modifications in the contra-American paradigm that could make class litigation practically feasible. A preliminary and non-exhaustive word may be in order about basic desiderata -- goals to be sought in choosing cost approaches if a system does opt to have some form of class action. The class action is meant to further various ends, including access to civil justice, enforcement of public norms, efficiency in resolution of mass controversies, and repose for defendants who prevail or pay a class-action judgment or settlement. An additional valid concern is what can be called filtering or screening -- trying to calibrate incentives so that access comes, at least as little as possible, at the cost of making it feasible for nuisance or strike suits to be brought.³² Further, any system would want to take into [*pg 133] account norms of particular domestic concern, such as the balance it seeks between private and public enforcement and the types of professional conduct it would like to tolerate or encourage -- although concerns for access and law enforcement without high public expenditure may force rethinking of some such norms.

Again, I am not advocating class actions, at all or in any particular form, for the rest of the world; but given the limited spread of class actions elsewhere in recent decades, and the interest being shown in some systems that have not yet adopted the device,³³ I hope it will be useful to explore the available fee-liability alternatives. The possibilities are numerous, which poses for other systems the intriguing chance of choosing approaches that seem best suited to yield class litigation of a nature and scale suited to a legal system's norms and perceived needs. Adoption of the class action need not, in other words, involve a headlong plunge into what for many may be uninviting American waters. And modifications of the contra-American paradigm are already numerous enough in practice to provide some experience in seeing how they really work, not just speculating about how they might.

III. Compensating the Plaintiff Class's Counsel

To start with the plaintiff's side's own lawyer's fees: If disapproval of contingent fees is not somehow relaxed, the options are limited and damage class actions are likely to be few. Even in successful cases, the individual class representative cannot practically be liable for all of the fees of class counsel. However reckoned, on an hourly or percentage or some hybrid or other basis, any solicitor-and-[*pg 134] client fees owing after an award of party-and-party fees³⁴ against the defendant could still exceed, or at least greatly diminish, the individual representative's share of the class recovery and make claimants unwilling to play the necessary role of representative. And to deal with possible lack of success, a rule of individual-representative liability if seriously enforced (ignoring the rules is always one possibility, although it seems better to set up systems in which participants don't predictably have to do so) would at the least threaten to eliminate all but sure-thing class actions. If there cannot be some form of risk-spreading or risk-shifting, one way of dealing with risks is always not to take them in the first place.³⁵

A. Liability of Unnamed Class Members

In theory, it might be possible to look to class members other than the representatives to share in paying the fees of class counsel, but none of the possible ways of doing so seems particularly promising. Trying to rely on voluntary contributions -- in effect, passing the hat among the class members -- would almost certainly not be viable on a regular basis, given the opportunities for free riding presented by class actions.³⁶ Efforts at collecting from class members by some involuntary means (other than taking the fee out of a recovery before distribution to class members, and in the event of success only -- which is exactly what a contingent fee does) would encounter huge problems of administration, cost, and collection, especially if the class were numerous and scattered and all the more if the individual members' claims were small. Moreover, damage class actions -- involving distinct claims that class members are legally entitled to bring on their own -- would likely have to have the [*pg 135] American opt-out feature;³⁷ and a notice to class members that they might face down-side liability could produce an opt-out stampede on the part of those who pay any attention to such notices, perhaps leaving an unviable class action and in any event subverting the end of economically wrapping up most or all of a widespread controversy in a single proceeding.

These concerns could lead to consideration of an opt-in approach, which might be viable for some actions involving large claims and at least fairly high likelihood of success. The choice facing such claimants could be between proceeding with an individual action and probably having to pay a fairly high fee to one's own attorney win or lose (still assuming that contingent fees were not allowed), versus opting into a class action³⁸ with the economies of scale that could yield a larger net recovery or at least smaller fees for each member to pay to class counsel in case of defeat. This is, indeed, the approach taken in English rules on group litigation adopted in 2000,³⁹ which is too recent for any evaluation of the success of the scheme.⁴⁰ Such an opt-in approach, in any event, would be highly unlikely to work for small-claim class actions, given both notice costs to numerous prospective class members and probable lack of interest on the part of many in signing up to be part of an effort to get at most small individual [*pg 136] recoveries. This setup would result, for better or worse, in the class device not serving the end that some in the United States and elsewhere see as an important function of class actions -- the ability to force disgorgement of gains from, and thus act as a deterrent to, widespread small rip-offs.⁴¹

B. Third-Party Funding and Contingent Fees

These difficulties make it sensible for a system considering class actions to relax, if it has not done so already, the requirement that clients -- class representatives, the represented class members, or both -- be liable for the fees of class counsel in the event of defeat or for payment of a full fee in case of only a small recovery. The risk of limited or no success would then have to be borne by either counsel for the plaintiff class under some form of contingent-fee system, or by a third party such as a legal-aid program, a special fund to finance class actions, or perhaps in some cases an association or union with interest in the subject matter of the litigation. (There appears to be no imperative to choose only a single one of these mechanisms; class counsel might be allowed to decide they were willing to take cases on a no-win, no-pay basis while there also existed, say, a public fund to support class actions -- which could also be useful to help with actions other than damage cases that might be meritorious but less likely to produce a recovery as a source for a substantial fee.) The last-mentioned approach, financing by an organization such as an association or union, seems innocuous -- at least under modern relaxed attitudes about maintenance⁴² -- and even useful if a group has the interest and resources, but unlikely to be forthcoming often enough to be a source principally relied upon.

1. Third-party funding.

Professor Vince Morabito's 1995 article⁴³ nicely presented the issues and experience up to then in the Australian context, with reference to Canadian developments as well, [*pg 137] and this part of my essay will do little more than summarize and update. Legal-aid organizations might be authorized to fund class actions, and even to profit from financing of successful cases, but could probably provide at most a limited response. It would be a diversion from their core mission of meeting the legal needs of those with limited means to authorize their funding of class actions that include those not eligible for legal aid, while the problems of financing class actions are largely independent of the means of the potential class representatives and members.⁴⁴ Yet if legal-aid groups could fund only actions on behalf of those meeting eligibility criteria, the response to the barriers to viable class actions would be only partial and the resulting system perhaps rickety.⁴⁵ In the United States, furthermore, class actions conducted by the federal Legal Services Corporation's grantees proved politically controversial enough to lead to a Congressional restriction on such litigation.⁴⁶

For a system preferring to include some regulatory element instead of leaving matters entirely to the private market by reliance on contingent fees, the creation of public funds dedicated to financing class actions could be preferable to legal-aid funding. The funds can be designed to be self-perpetuating after an initial appropriation, taking reimbursement for money paid to the class representatives plus a modest percentage of awards in successful cases. Such funds exist in Ontario and Quebec, and have been proposed in Australia.⁴⁷ The experience in the two Canadian provinces has contrasted [*pg 138] significantly, with the Quebec fund being quite active⁴⁸ and the Ontario one relatively inactive⁴⁹ -- which suggests the need for adequate funding and careful attention to the features of the funds. The Quebec fund illustrates both the possibility of coexistence with private financing and the possibility of including a potentially desirable assessment of the substantive merit of the claim as part of the decision whether to provide funding.⁵⁰ Since the greater significance of these funds appears to be in connection with protection of class members from liability for fees of winning defendants,⁵¹ further discussion will be saved until later.

2. Contingent fees.

Third-party funding can play a valuable role, but it is subject to the generosity, interests, and assessments of others besides class members and counsel -- and, in the case of public funding, to political swings and governmental budget priorities. [*pg 139] Subject instead to a financial self-interest that may be more nearly aligned to that of class representatives and members is, of course, one or another form of contingent fee under which class counsel is paid only in the event of success (by judgment or, as a practical matter more often, settlement) and out of the class recovery. Movement toward contingent fees as a way of financing plaintiffs' damage litigation, along with cutbacks in costly legal aid, has already taken place in the context of individual actions in England;⁵² and, despite occasional assumptions to the contrary,⁵³ there is nothing incompatible about having contingent fees and retaining the loser-pays rule if a system wants to do so.⁵⁴ Plaintiffs can remain liable for the fees of successful defendants while not being obligated to pay their own lawyer if they lose. Turning the class's lawyer into an insurer by shifting the risk of nonpayment to the lawyer calls, of course, for a premium in the form of some increase over what would be at least theoretically risk-free alternative work at an hourly or otherwise fixed rate, and the question then becomes how to reckon the premium.

The main alternatives in use are some form of the American contingent percentage fee, taken from the whole recovery on the long-established "common fund" theory,⁵⁵ and an hourly rate with a bonus. The latter appears in the United States with the "lodestar" approach of multiplying hours reasonably spent on successful aspects of a case times a reasonable market-based hourly rate,⁵⁶ with the addition of a premium for success,⁵⁷ and in England and Australia as the [*pg 140] "conditional" fee based on the lawyer's hourly rate with an "uplift" -- in both those nations, capped at some percentage above what would have been the hourly-based fee.⁵⁸ The amount of the uplift affects how risky a case a lawyer has an incentive to take -- with a 100% uplift providing an incentive for a risk-neutral lawyer to take cases with as low as a 50% odds of success, and lower amounts encouraging only cases with better chances -- thus permitting a system to encourage or discourage riskier litigation. While both calculation approaches rely to some extent on private incentives and arrangements, in the class context neither can operate with the degree of freedom from judicial control that characterizes non-class litigation. Given the potential for conflict between a class lawyer and class members, with the lawyer having a greater self-interest in maximizing the fee than the return to individual members, court control of awards -- as of other aspects of terms when class actions are settled -- is essential. Indeed, difficulties with settlements and fees of class counsel in the United States have led to recent proposals for amendments to Federal Rule of Civil Procedure 23 that would increase court scrutiny of counsel selection, settlement terms, and fee awards.⁵⁹

A system adopting a class-action procedure or contemplating a change in its attorney-fee regime for existing class actions could also consider experimenting with some form of a simple but sophisticated proposal made by Professor Kevin Clermont and John Currivan to deal with many conflict and incentive problems with both contingent percentage and hourly fees: a hybrid contingent hourly-percentage fee. The fee, payable only in the event of success, would be "the sum of two simple components: (1) the lawyer's time charge for the hours devoted to the case, and (2) a percentage of the amount by which the [*pg 141] gross recovery exceeds that time charge."⁶⁰ This is not the place to expound on the merits of one particular proposal, or to venture near the large theoretical literature on incentive effects of various fee arrangements. More important is the point that a range of possible approaches to contingent fees exists, offering opportunities for different systems to choose ones that seem to suit best their norms and needs. In particular, systems with reservations about perceived American excesses have open to them alternatives with the potential to make for workable class actions without the sometimes very large fees and acute counsel-class conflicts that cause much controversy in the United States.

IV. Loser-Pays Attorney Fee Shifting

When a plaintiff class wins a damage recovery, loser-pays liability for attorneys' fees -- if retained for class actions -- should work with not much greater difficulty than in non-class litigation. Cases tried to a judgment for the plaintiff class could be treated through usual cost-taxing procedures; in negotiations and settlements fee liability would be an element to be added to the recovery, subject to the sometimes problematic judicial scrutiny that is a necessary part of a class-action settlement.⁶¹

The rub would come when the class lost -- or, more precisely, in the difficulties and disincentives created by the ex ante prospect of having to pay at least a substantial part of a winning defendant's reasonable attorney fees should a class action go to a litigated judgment for the defendant. The two most obvious possible sources for the payment, the individual representative(s) and the class as a whole, are largely non-starters. For reasons discussed above,⁶² the threat of down-side liability for large defense fees along with those of class counsel would at least mostly scare off potential class [*pg 142] representatives, defeating the class device's purposes of promoting access, economical processing, and repose.⁶³ Similarly, looking to members of a losing class seems at least as unpromising to collect the fees of a winning defendant as it does to pay the fees of the class's own lawyer,⁶⁴ and is either banned or not provided for in Canada⁶⁵ and Australia.⁶⁶

[*pg 143]

If it would not be consistent with viable class actions to have either class representatives or class members liable for counsel fees of winning defendants, the question then becomes what the alternatives might be. Four principal possibilities seem to exist: making class counsel liable; commercial insurance; public funding; and -- what might be, along with contingent fees, the most wrenching paradigm shift -- abolishing or sharply limiting the loser-pays rule for class actions. The rest of this part briefly discusses each alternative in turn.

A. Liability of Class Counsel

One way of viewing the liability of class counsel for fees of winning defendants would be as an extension of the principle behind the contingent fee, with counsel bearing the risk of liability for the other side's fee as they already do with the risk of nonpayment of their own fee in the event of defeat. Professor Deborah Hensler and I explored the possibility of either-way loser-pays for damage class actions, with plaintiffs'-side fee liability on class counsel, in the American context -- where losers, and especially losing plaintiffs, are usually not liable for winning defendants' fees -- in a recent article.⁶⁷ Viewing the losing plaintiff class's lawyer as unsuccessful [*pg 144] entrepreneur rather than malefactor deserving sanctions, we emphasized the possibly favorable access and screening effects of the ex ante prospect of fee recovery and threat of fee liability.⁶⁸

We recognized, though, that many complex details would accompany and perhaps bedevil such a rule,⁶⁹ cautioning that the positive access and screening effects of loser-pays with class-attorney liability "might be quite small, and the practical challenges of grafting such a rule onto our civil justice system might turn out to be so high that it [would] not [be] worth the effort."⁷⁰ If opposition to the idea would be strong from at least some segments of the bar in the United States,⁷¹ in other nations with a less entrepreneurial view of the function of class counsel in particular and lawyers in general it would no doubt be anathema to many, a step much too far over the line that has been crossed only with difficulty to allow a tamer form of contingent fee than exists in America.⁷² The idea may still be "worth further scholarly consideration,"⁷³ but it is highly unlikely to be a realistic alternative for other systems with class actions any time at all soon.

B. Commercial Insurance

If class counsel are unlikely to act as insurers against the down-side risk of fee liability for a losing class, protection against the risk might come from commercial insurers. Legal-expense insurance is fairly common in some Continental nations,⁷⁴ partly because of the established threat of liability for a winning adversary's fees, and has spread in England as Legal Aid cutbacks took away the protection from such liability that had been afforded to covered cases.⁷⁵ But the [*pg 145] same collective-action dilemma that arises with liability of class representatives or members for fees.⁷⁶ seems to pose an obstacle to the workability of such insurance: it would impose a disproportionate cost on individual representatives if they were liable for the full premiums, and it would usually be prohibitively difficult to collect pro rata shares from the class members. Only if class counsel were liable for shifted fees might it be workable to rely on insurance, as they could decide whether to self-insure or seek other coverage. But the near-certain political unviability of counsel's liability in the first place⁷⁷ makes this combination appear at most a highly distant possibility.

C. Public and Third-Party Private Funding

If class representatives, members, and counsel and commercial insurance all seem unpromising as ways to deal with the risk of liability for a shifted defense fee, still other sources might be possible. Although financing by a private group such as an interested association or union might be conceivable in some cases for the fees of a class's own counsel, it seems too much to expect such entities to volunteer to bear loser-pays fee liability at all often when they are not parties facing it involuntarily. So the main possible source would be the kinds of publicly created funds that exist in Ontario and Quebec, which can be liable for both the fees of class counsel and those of a winning defendant.⁷⁸ After an initial appropriation, such funds can be structured to be self-financing out of a contingent share of recoveries in successful cases, although there is always the possibility of fund-depleting awards that could require a perhaps difficult political decision whether to provide another subvention.

[*pg 146]

Further, it might limit access too much to rely solely on such a fund, by requiring its approval for any private damage class action to proceed, to cover the risk of liability for a class's winning adversary's fees; caution, error, backlog, or insufficient resources could keep a fund from aiding what might be viable and meritorious actions. And under loser-pays, such actions might be too unlikely to proceed at all without support from the fund. The high activity of the Quebec fund⁷⁹ makes it appear that such an entity can be a significant source of support, but if fund financing should not be the sole way of bringing a class action a system could find itself back at the hard choice: keep loser-pays with its likely evasions and access-inhibiting features, given the inadequacy of the alternatives canvassed above -- or abandon or modify the loser-pays rule for some or all class actions. The next section considers different ways in which a system could do so, although variations in local preferences and conditions counsel against my recommending any one.

D. Modifying or Abandoning Loser-Pays

A threshold concern with possible changes in loser-pay rules, to accommodate perceived needs of class actions, would be whether to make the changes applicable generally or to limit them to the class context. The latter would be an obvious starting point, if the justifications for change came from the peculiarities of class actions such as the difficulty of making loser-pays work in that context; but so limiting any changes would raise questions of discrimination and possibly skewed incentives for choice between class and non-class litigation. (Unless the modification took the form of a primarily one-way pro-prevailing-plaintiff rule, which may be unlikely to win acceptance in systems with a strong two-way loser-pays tradition,⁸⁰ it is not clear that softening or eliminating loser-pays would regularly cut one way or the other in affecting parties' choices. Tempering or abolishing the rule would mean that plaintiffs would be less likely both to collect fees from the other side when they won and to have to pay them when they lost. Risk aversion would push toward choosing the class form to avoid down-side risk, but whether those with deciding voices were risk-averse would depend on who bore risk.) In any event, the class-action tail should probably not wag the entire dog [*pg 147] when a system is deciding on whether to keep, modify, or abolish loser-pays; and the seeming unlikelihood of its abolition in systems where loser-pays has been so long and deeply entrenched makes it seem sensible to focus on possible abolition or modification in the class-action context alone.⁸¹

Abolition of the loser-pays rule, so that neither losing defendants nor someone on the side of a losing plaintiff class would be liable for the other side's fees, is exactly what the Ontario Law Revision Commission recommended (with limited exceptions) for class actions there.⁸² Interestingly, the OLRC's recommendation was not followed in Ontario but was largely adopted in British Columbia. Quebec has still a third approach retaining loser-pays but reducing the class representative's liability to a nominal sum. As Professor Watson explains:

British Columbia adopted the Ontario Law Reform Commission's recommendation, rendering the representative plaintiff virtually immune from paying costs -- she is only liable if the action is "frivolous or vexatious" (merely losing is not enough). Quebec allows for costs against the representative plaintiff, but after one very large costs award the legislation changed allowing only nominal costs to be paid (on the scale of the small claims court). Ontario is the most extreme. The legislation provides that costs can be awarded against a losing representative plaintiff unless the court is of the view that the action was a "test case, raised a novel point of law or involved a matter of public interest." To date, relatively few significant costs awards have been made against plaintiffs, and representative plaintiffs have largely avoided having to pay costs.⁸³

With still different approaches, Australia and England complete an illustrative array of possible ways of dealing with the loser-pays rule in class actions where that rule remains generally the norm. Australia retains the rule, but with specific provision that class members other than the representative cannot be liable for a winning [*pg 148] defendant's fees.⁸⁴ And England, with its unique opt-in approach, also retains the rule while making all class members liable for common and individual costs.⁸⁵ Putting aside the English approach as unsuited for other nations' regimes with their opt-out features, two of the remaining four alternatives seem particularly questionable. The Australian system creates incentives for evasions and counter-measures,⁸⁶ and Ontario's limited-exception approach has the same problem for cases in which class representatives might be liable⁸⁷ while adding ex ante uncertainty about whether a case will come within one of the exceptions. Still, both Australia and Ontario seem to have significant class-action practice. With the device available it appears that at least a few will find ways to use it despite the obstacles, although with some costs in both hypocrisy and satellite maneuvering and probably at a lower rate of use than would otherwise be the case.

V. Conclusion

As this survey of potential approaches to attorney-fee issues in class actions makes clear, this limited part of the class-action field can involve what may be surprisingly numerous and complex choices. Still one further potential feature should be mentioned, if only to explain why it should not apply in class actions: that of an offer-of-judgment rule like Federal Rule of Civil Procedure 68 and many counterpart state rules in the United States, or the payment-into-court device in England. Such rules, designed to encourage settlement, permit a defendant to make a formal offer of a certain amount, with the plaintiff subject to a consequence -- such as liability for defendant's post-offer non-fee costs and/or attorney fees -- if the plaintiff does not accept and does not do better in the final result. Whatever the virtues or defects of such rules generally, they should not apply in class actions because they presume that plaintiffs have autonomy to accept or reject such offers, which a plaintiff class does not given the requirement of court approval of settlements. "It would be unfair to hold claimants liable . . . if they sought to accept a formal offer but did not win court approval."⁸⁸

[*pg 149]

In addition to that merciful if small bit of simplification, this survey can conclude on a note of optimism. While the choices among approaches to fee issues that face framers of class-action rules in systems with loser-pays rules and restrictions on contingent fees can be difficult and may become politically sensitive, these choices nonetheless offer considerable opportunity for trying to shape the class device so that it is usable in meritorious cases while reducing conflicts of interest and incentives to pursue questionable claims. And although having class actions does raise tensions with some existing norms and practices concerning attorney fees in many systems, the fee provisions applicable in class actions can be framed in ways that accommodate them at least somewhat to traditions that claim strong adherence. Some adjustments to existing fee rules, perhaps uncomfortable ones, seem essential for the creation of a viable class-action regime, but there is a considerable range of choices that are not limited to American-style approaches with what strike many as their excesses. So one can say to those otherwise inclined to adopt the class device, when it comes to framing rules governing attorney fees, "Come on in -- water fine."

FOOTNOTES

*	Elvin R. Latty Professor of Law, Duke University. My thanks to the University of Sydney Faculty of Law for its Parsons Fellow support during a visit in January-February 2002; to the law faculties at Sydney and at Flinders University in Adelaide for hospitality and comments on earlier versions of this paper; to Justice Murray Wilcox and his colleagues at the Federal Court in Sydney for their help in understanding Australian class-action practice; and to Profs. Antonio Gidi, Vince Morabito, William Rubenstein, and Garry Watson for comments on an earlier draft of this essay. I am also greatly indebted to Herbert Bernstein for the learned help he was always willing to give to amateur comparativists like myself. We all miss him. Errors are, of course, mine.
1.	A fairly widely encountered saying, with no apparent single originator. See, e.g., Bayou Philosophy (attributing quotation to "a cybersage") available at http://david_hewins.tripod.com/floridacajun/id10.html (visited Sep. 16, 2002).
2.	The statement in text is impressionistic, but class actions are available in American federal courts, see FED. R. CIV. P. 23, and in all states but two, see generally, e.g., LINDA S. MULLENIX, STATE CLASS ACTIONS: PRACTICE AND PROCEDURE (2000) (state-by-state discussion of class-action rules, with only Mississippi and Virginia not generally providing for them).
3.	For a good, recent discussion of the controversies surrounding class actions in the United States, see DEBORAH R. HENSLER ET AL., CLASS ACTION DILEMMAS: PURSUING PUBLIC GOALS FOR PRIVATE GAIN (2000). A shorter presentation drawing on the data and observations of the Hensler et al. book is Deborah R. Hensler, Revisiting the Monster: New Myths and Realities of Class Action and Other Large Scale Litigation, 11 DUKE J. COMP. & INT'L L. 179 (2001).
4.	British Columbia, Newfoundland and Labrador, Ontario, Quebec, and Saskatchewan. See, e.g.,, Garry D. Watson, Class Actions: The Canadian Experience, 11 DUKE J. COMP. & INT'L L. 269 (2001) (describing class-action provisions and practice existing at the time in British Columbia, Ontario, and Quebec). The Newfoundland/Labrador and Saskatchewan provisions took effect in 2002 and are modeled after the British Columbia legislation. See Class Actions Act, S. Nfld. 2001 Ch. C-18.1 (in force Apr. 1, 2002); The Class Actions Act, S.S. Ch. C-12.01 (in force Jan. 1, 2002).
5.	See, e.g., S. Stuart Clark & Christina Harris, Multi-Plaintiff Litigation in Australia: A Comparative Perspective, 11 DUKE J. COMP. & INT'L L. 289 (2001).
6.	See Antonio Gidi, Class Actions in Brazil: A Model for Civil Law Countries, 51 AM. J. COMP. L. (forthcoming 2003) (on file with author).
7.	See Note, Class Action Litigation in China, 111 HARV. L. REV. 1523 (1998).
8.	I attended an International Conference on Class Action Procedures and Their Implementation in the Indonesian Judicial System in Jakarta in February 2002, and the Chief Justice of the Supreme Court of Indonesia has since issued a Regulation of the Supreme Court Concerning Class Action Procedures (2002) (on file with author).
9.	See Per Henrik Lindblom & Roberth Nordh, Scandinavian Developments: The New Swedish Act on Group Actions (Sep. 13-14, 2002) (on file with author). For a published description of the Swedish proposal at a late stage in its development, see Roberth Nordh, Group Actions in Sweden: Reflections on the Purpose of Civil Litigation, the Need for Reforms, and a Forthcoming Proposal, 11 DUKE J. COMP. & INT'L L. 381 (2001). See also Per Henrik Lindblom, Individual Litigation and Mass Justice: A Swedish Perspective and Proposal on Group Actions in Civil Procedure, 45 AM. J. COMP. L. 805 (1997) (discussing earlier Swedish proposal).
10.	See Lindblom, supra note 9, at 824 (mentioning proposals put forward by state-appointed commissions in Finland, Norway, and Scotland); Norwegian Civil Procedure Comm'n, Proposed Rules Concerning Group Action (provisional draft Feb. 2000) at http://www.law.duke.edu/grouplit/papers/ProposedRules.pdf (visited Sep. 21, 2002) .
11.	South African Law Comm'n, The Recognition of Class Actions and Public Interest Actions in South African Law (1998) at http://wwwserver.law.wits.ac.za/salc/report/classact.pdf (visited Sep. 16, 2002) . See also 3 COMM'N OF INQUIRY INTO THE AFFAIRS OF THE MASTERBOND GROUP AND INVESTOR PROTECTION IN SOUTH AFRICA, CORPORATE LAW AND SECURITIES REGULATION IN SOUTH AFRICA 651-953 (2001) (surveying studies and practices in several nations and recommending adoption of class actions in South Africa) at http://www.doj.gov.za/commissions/nelcommission/chapter15_18.pdf (visited Oct. 19, 2002).
12.	See, e.g., Donald N. Dewees et al., An Economic Analysis of Cost and Fee Rules for Class Actions, 10 J. LEGAL STUD. 155, 157 (1981) ("in England and Canada, the representative plaintiff's liability for the defendant's costs in a losing action, his inability to demand contribution by class members to the costs of the action regardless of the success or failure of the action, and the widespread prohibition against contingent fees have together created a situation where a representative plaintiff is never better off in economic terms bringing his action in class rather than individual form") (footnote omitted) (emphasis in original); Per Henrik Lindblom, Group Actions in Civil Procedure in Sweden, in SWEDISH NATIONAL REPORTS TO THE XIIITH INTERNATIONAL CONGRESS OF COMPARATIVE LAW 59, 92 (1990) ("If the class action were adopted in Sweden today without changing our rules on costs, very few such cases would be likely to be instituted.") (italics omitted); Vince Morabito, Federal Class Actions, Contingency Fees, and the Rules Governing Litigation Costs, 21 MONASH U.L. REV. 231, 232 (1995) ("the existing costs rules governing litigation, if applied unaltered to class actions, could constitute 'a disincentive to bringing grouped proceedings and might in fact create yet another barrier to access to legal remedies of the kind which the [class action] itself aims to overcome'" (quoting AUSTRALIAN LAW REFORM COMM'N, GROUPED PROCEEDINGS IN THE FEDERAL COURT 106 (1988)); 3 ONTARIO LAW REVISION COMM'N, REPORT ON CLASS ACTIONS 703 (1982) [hereinafter OLRC REPORT] ("the application of the existing Ontario costs rules to class actions has resulted, and will continue to result, in the commencement of very few class actions").
13.	See, e.g., 10 CHARLES ALAN WRIGHT ET AL., FEDERAL PRACTICE & PROCEDURE �� 2675, 2675.2 (3d ed. 1998) (describing general American practice and exceptions under federal law).
14.	See, e.g., Herbert M. Kritzer, The Wages of Risk: The Returns of Contingency Fee Legal Practice, 47 DEPAUL L. REV. 267, 267 & n.1 (1998) ("The contingency fee is one of the defining characteristics of civil litigation in the United States. . . . While a number of countries have some form of contingency fee, the American contingency fee is relatively unique, both in its form and its dominance as a means of funding litigation. What sets the American contingency fee apart from contingency arrangements in most other countries is that it is based on a percentage of recovery. . . .").
15.	See, e.g., Richard O. Faulk, Armageddon Through Aggregation? The Use and Abuse of Class Actions in International Dispute Resolution,* 10 MICH. ST. U.-DCL J. INT'L L. 205, 210 & n.9 (2001) (referring to the "'entrepreneurship' of contingent fees" as "a hallmark of American class action litigation that promotes risk-taking by the plaintiffs' bar," and citing further sources).
16.	See, e.g., Thomas D. Rowe, Jr., The Legal Theory of Attorney Fee Shifting: A Critical Overview, 1982 DUKE L.J. 651, 651 (1982) (referring to loser-pays practice in England and Continental nations).
17.	Because of the prominence of the American contingent fee based on a percentage of the recovery, it may often be assumed that the term "contingent fee" refers to such a contingent percentage fee. But the element of contingency comes from the no-win, no-pay feature and is independent of the means by which the amount of the fee is determined. The conditional fee arrangement used in, for example, England, see infra note 29, is also a contingent fee but is based on an hourly rate instead of a percentage of the recovery. This essay uses "contingent fee" to refer to all forms of contingent fees, not just the American percentage variety, and speaks more precisely when referring to a particular type of contingent fee.
18.	See, e.g., Geoffrey Woodroffe, Loser Pays and Conditional Fees-An English Solution?, 37 WASHBURN L.J. 345, 349-51 (1998) (describing historical roots of and rationales for prohibition of contingent fees in England).
19.	See, e.g., id. at 350 (giving, as rationale for unlawfulness of "champertous" contracts under which lawyer would get share of proceeds of suit, that "if lawyers have a personal financial interest in the outcome of cases so that their reward depends on the result of the litigation, this may affect the objectivity and impartiality of their advice and undermine the integrity of the administration of justice in that the lawyer as an officer of the court will have a conflict of interest and be torn between self-interest and his or her duty to the court and to the client").
20.	See BLACK'S LAW DICTIONARY 965 (7th ed. 1999) (defining "maintenance" as "[a]ssistance in prosecuting or defending a lawsuit given to a litigant by someone who has no bona fide interest in the case; meddling in someone else's litigation.").
21.	See, e.g., Morabito, supra note 12, at 233 ("Class proceedings tend to last longer and be more complex than individual suits," and safeguards to protect the interests of unnamed class members "increase substantially the costs incurred by the representative plaintiff and render a class suit a considerably more expensive form of litigation than individual proceedings") (footnotes omitted).
22.	See, e.g., 3 OLRC REPORT, supra note 12, at 656 (footnote omitted): [I]f the class action fails, the representative plaintiff will be liable for two sets of costs. Even if the suit succeeds and there is the usual award of party and party costs [i.e., a partial award of the winning plaintiff's attorney fees], the representative plaintiff will still have to pay to his own lawyer the amount that is not indemnified by the defendant. The extent of the potential liability of a representative plaintiff, whatever the result of the suit, will be affected by the particular nature of class actions, which ordinarily are more complex, lengthy proceedings than individual actions. There will therefore be a commensurate increase in the ancillary expenses and lawyers' fees. This, of course, will augment the financial risk assumed by a representative plaintiff.
23.	John C. Coffee, Jr., The Mandatory/Enabling Balance in Corporate Law: An Essay on the Judicial Role, 89 COLUM. L. REV. 1618, 1643 (1989).
24.	Loser-pays fee shifting may encourage strong small claims more than does the American rule. See, e.g., Thomas D. Rowe, Jr., Predicting the Effects of Attorney Fee Shifting, 47 LAW & CONTEMP. PROBS. 139, 148-49 (Winter 1984). As the text goes on to discuss, however, such fee liability is also highly likely to discourage the choice of class over individual litigation if the claim is being pursued at all.
25.	See, e.g., Phillips Petroleum Co. v. Shutts, 472 U.S. 797, 809 (1985) (Rehnquist, J.) ("Class actions. . . may permit the plaintiffs to pool claims which would be uneconomical to litigate individually. For example, this lawsuit involves claims averaging about $100 per plaintiff; most of the plaintiffs would have no realistic day in court if a class action were not available."); RICHARD A. POSNER, ECONOMIC ANALYSIS OF LAW 585-86 (6th ed. 2003) ("The modern class action generalizes [the] technique . . . [of] aggregating a number of small claims into one large enough to justify the costs of suit-or, stated otherwise, . . . realizing economies of scale in litigation.").
26.	See, e.g., Morabito, supra note 12, at 233-34 ("Potential representative plaintiffs whose claims are individually non-recoverable would be unlikely to commence class actions as the extent of their potential liability for costs would exceed the value of their own claim.") (footnotes omitted); OLRC REPORT, supra note 12, at 659 ("A modest claim will not defray the solicitor and client costs payable to the class lawyer after a successful action, let alone justify the risk of bearing party and party costs."). "Solicitor and client" costs are those "payable by a client to his own lawyer," id. at 647, including attorney fees, while "party and party" costs are "payable to a successful party by his unsuccessful adversary," id. Usually, party and party "indemnity is only partial." Id. at 649.
27.	See, e.g., Morabito, supra note 12, at 234 ("In relation to those potential class representatives . . . whose claims are individually recoverable, individual proceedings constitute a more appealing option than grouped proceedings as they involve lower costs.") (footnotes omitted); OLRC REPORT, supra note 12, at 659 ("A person with an individually recoverable claim is better advised to bring his own action against the defendant, rather than incur the more onerous financial costs of a class action, which can bring him no greater material advantage.").
28.	Actual practice in some systems has long involved moderation of the paradigm described in the text. See, e.g., Herbert M. Kritzer, Fee Arrangements and Fee Shifting: Lessons from the Experience in Ontario, 47 LAW & CONTEMP. PROBS. 125, 130-32 (Winter 1984) (describing de facto existence of contingent fees in Ontario, from plaintiffs' lawyers' tendencies not to collect fees owed by losing plaintiffs and from charging of prevailing plaintiffs based partly on amounts recovered, despite formal disapproval of contingent billing).
29.	Recent developments in some nations, under pressures to reduce the cost of legal-aid programs and to improve access to civil justice, have moved away from the text's paradigm in the American direction, even for non-class actions. See, e.g., Richard L. Abel, An American Hamburger Stand in St. Paul's Cathedral: Replacing Legal Aid with Conditional Fees in English Personal Injury Litigation, 51 DEPAUL L. REV. 253 (2001) (recounting history of adoption of contingent hourly fees in England); Woodroffe, supra note 18 (same); Editorial, Contingency Fees: A Case of Too Little, Too Late, DE REBUS, July 1999 (criticizing new South African measures allowing contingent hourly fees as lagging behind recent spread in practice of contingent percentage fees) available at http://www.derebus.org.za/archives/1999Jul/editorial/editeng.htm(visited Sep. 22, 2002). What I have described as a "contingent hourly" arrangement to distinguish it from the American contingent percentage fee is known, perhaps because of historic British aversion to the idea of the contingent fee, as a "conditional fee agreement" or CFA. Under a CFA, the lawyer would usually receive nothing if the case were lost, but would receive a basic fee plus a success fee if the case were won or settled with a payment to the client. . . . [T]he . . .CFA involves a basic fee which is based on the lawyer's hours worked at an agreed hourly fee and the success fee is a percentage increase of the basic fee. . . . The success fee may never be more than 100% of the basic fee. Christopher Hodges, Multi-Party Actions: A European Approach, 11 DUKE J. COMP. & INT'L L. 321, 351 (2001) (footnote omitted). In this essay I am not assuming that the paradigm exists in pure form in other nations nor implying that class actions would be the only source of pressure to move away from it. What I do maintain is that creating of viable class-action devices elsewhere would be virtually certain to require some modifications to the paradigm, and that the existence of class actions will create particularly acute pressures for changes.
30.	This introductory sketch oversimplifies somewhat, because alternatives to contingent fees such as public funding might substitute for the financial risk-bearing role of the plaintiff class's attorney. Later discussion takes account of such possibilities, see infra notes 43-51, 78-79 and accompanying text; but the essential point remains that some alternative to the bearing of large risk for paying class counsel's fee by the individual class representatives is necessary for a viable class-action regime. And contingent fees in some form are the most prominent of the alternatives.
31.	See, e.g., OLRC REPORT, supra note 12, at 663 ("the existing costs rules have the effect of discouraging all class actions, for reasons having nothing to do with their propriety or merits"). The main focus of this essay is on cost rules in class actions for damages, not such actions for prospective relief like injunctions or declaratory judgments. Relief in actions challenging policies affecting many may have the same effect whether the action is in individual or class form; an injunction against enforcing an invalid law will or at least should keep the government from enforcing it at all, even if the litigation is not formally a class action. Hence the problems with cost rules in class actions are likely to be less unique and acute in suits seeking prospective relief as opposed to damages. Also, contingent fees out of recoveries are viable only in damage actions and thus deserve discussion in that context alone. Incentive problems with loser-pays rules can of course occur in injunctive and declaratory actions, but for the most part they probably call for similar responses-whatever those may be-in class and nonclass litigation.
32.	See Deborah R. Hensler & Thomas D. Rowe, Jr., Beyond "It Just Ain't Worth It": Alternative Strategies for Damage Class Action Reform, 64 LAW & CONTEMP. PROBS. 137, 140 (Spring/Summer 2001) (seeking "to identify mechanisms for enhancing the system's capacity to screen out non-meritorious suits, while preserving access for meritorious actions"). There is also the possibility of trying to rely on other approaches than cost rules, such as various possible tests for certification of class actions, to perform the screening function: [A]n action should be allowed to proceed as a class action only after it has received judicial approval, upon satisfaction of a number of certification tests. These tests are intended to act as a filter, allowing only proper, meritorious class actions to be brought in Ontario. This purpose is particularly evident in the incorporation of a preliminary test on the merits and a "cost-benefit" test, the latter of which permits the court to refuse to certify a class action if it is of the view that the adverse effects of the proceedings would outweigh its benefits. In light of the purposes of the certification tests-to weed out inappropriate class actions-we believe that it is neither necessary nor appropriate for the costs rules to be assigned a similar function. OLRC REPORT, supra note 12, at 663 (footnote omitted). However a system chooses to pursue the screening or filtering function, it is important to have some checks on abuse and overuse of the class device. And if cost rules can be set so as to calibrate incentives at desirable levels, they have the advantage of requiring less judicial management than other approaches.
33.	See supra notes 10-11 (proposals for class actions in Finland, Norway, Scotland, and South Africa).
34.	For definitions of these terms, and explanation that "party-and-party" fees awarded to winners are usually partial (thus leaving a remainder owed by client to counsel), see supra note 26.
35.	It would be possible to allow some sort of bonus payment to class representatives out of recoveries in winning cases to compensate for disproportionate costs they would have to bear if liable for all of class counsel's fees, but such an approach would leave unsolved the serious incentive problems resulting from the chance of defeat-with fee liability still existing if no form of contingent fee arrangement were allowed or subsidy provided.
36.	Professor Morabito attributes the "reluctance of most, if not all, absent class members to contribute to the expenses of the suit" to two major factors. In the first place, class members will be able to enjoy the benefits flowing from a successful class suit whether or not they provide any financial assistance to the representative plaintiff. Another reason for this reluctance is due to the remoteness of any potential benefit at the time the request for contributions is normally made, namely, in the early stages of the proceedings. Morabito, supra note 12, at 236 (footnotes omitted).
37.	FED. R. CIV. P. 23(c)(2) ("The notice [in any common-question class action maintained under Rule 23(b)(3)] shall advise each member that (A) the court will exclude the member from the class if the member so requests by a specified date . . ."). For a defense of mass-tort class actions with no opt-out feature, see David Rosenberg, Mandatory-Litigation Class Actions: The Only Option for Mass Tort Cases, 115 HARV. L. REV. 831 (2002).
38.	If any action were brought in class form at all, which for reasons discussed supra in note 27 and accompanying text it quite likely would not be-even with claims large enough and promising enough to be worth pursuing in individual lawsuits.
39.	See Neil Andrews, Multi-Party Proceedings in England: Representative and Group Actions, 11 DUKE J. COMP. & INT'L L. 249, 260 (2001)) (footnotes omitted): Those wishing to join and take advantage of group litigation under the 2000 rules must either affirmatively register as parties to the relevant claim, or at least have their particular claims adjoined by judicial consolidation to the group action. Therefore, group actions involve positive opting-in, or at least a positive decision to litigate.
40.	See id. at 266 ("the newly refined 2000 model of group litigation orders must be allowed time to prove itself"). There are, though, reasons for pessimism. An opt-in regime existed for some time in the Australian state of Victoria. See David Kell, The Liability of Represented Persons for Party-Party Costs in Representative Actions, 13 CIV. JUST. Q. 233, 236 (1994) ("persons to be represented must 'opt in' by expressing in writing their consent to being represented"). This requirement was criticized, see Clark & Harris, supra note 5, at 294 n.27, and "[n]o proceedings ever progressed to final hearing under these provisions," id. at 295. The Victoria scheme was replaced in 2000 by "a class action procedure virtually identical to that of the Federal Court," id. at 292, which provides for an opt-out rather than an opt-in mechanism, see id. at 299-300. See also Hensler & Rowe, supra note 32, at 144-47 (critiquing opt-in proposals).
41.	See, e.g., Hensler & Rowe, supra note 32, at 146 n.30 (discussing possible effect of change to opt-in approach on enforceability of consumer-protection legislation); FED. R. CIV. P. 23(b)(3) Committee Note to 1966 amendment ("a fraud perpetrated on numerous persons by the use of similar misrepresentations may be an appealing situation for a class action . . . .").
42.	See supra note 20; Michael Zander, Will the Revolution in the Funding of Civil Litigation in England Eventually Lead to Contingency Fees?, 52 DEPAUL L. REV. 259, 259-60 (2002) (mentioning historic disapproval of maintenance and 1966 Law Commission report detailing forms of maintenance already allowed then, such as litigation funding by trade unions and associations).
43.	Morabito, supra note 12.
44.	See OLRC REPORT, supra note 12, at 661: [T]he principle upon which legal aid is granted in individual actions-the impecuniosity of the plaintiff-does not address the economic problem faced by a prospective plaintiff in the class action context. His dilemma is not necessarily a lack of financial resources, but rather the fact that the potentially enormous costs of litigation so exceed the amount of his personal stake that, regardless of his resources, it would not be economically rational to initiate a class action.
45.	If legal-aid financial-eligibility criteria were "applied not just in relation to the financial means of the representative plaintiff but also to the means of all class members," that would create "obvious administrative problems, especially in opt out schemes which do not require the identification of, and the express consent to the bringing of the class suit by, the class members." Morabito, supra note 12, at 264 (footnote omitted).
46.	See 42 U.S.C. � 2996e(d)(5) (2000) ("No class action suit, class action appeal, or amicus curiae class action may be undertaken, directly or through others, by a staff attorney, except with the express approval of a project director of a recipient in accordance with policies established by the governing body of such recipient.").
47.	See Morabito, supra note 12, at 265-70 (discussing funds in Quebec and Ontario, and proposals by Australian Law Reform Commission, South Australian Law Reform Committee, and Australian Access to Justice Advisory Committee).
48.	See FONDS D'AIDE AUX RECOURS COLLECTIFS, RAPPORT ANNUEL 1999-2000, at 9 (2000) (during period 1985-2000, annual average of 51 requests for aid presented to Quebec fund); id. at 10 (total of 1026 requests made to fund from 1978 to 2000); id. at 11 ("from 1985 to 2000, the Fund has approved 84.7% and has denied 14.3% of the requests on which a decision has been made") (translation by author); id. at 12 (from 1985 through 1999, Quebec fund financed on average 65.5% of class actions docketed each year in Quebec superior court).
49.	See Watson, supra note 4, at 275-76 (Ontario Class Proceedings Fund "has been a failure in that, due to inadequate financing, it has given funding to very few class actions (approximately six to date).") (footnote omitted). An additional reason for limited use of the fund could be caution induced by heavy exposure to defense fees in case of defeat: "Once the [Class Proceedings] Committee decides to provide the applicant with some financial assistance in relation to the class suit's disbursements, the Committee automatically becomes liable for any costs that are awarded against the financially assisted class plaintiff, no matter how modest the grant from the Fund happened to be." Morabito, supra note 12, at 266 (footnote omitted). An additional reason, though, could be lack of need: "[A]pplications to the fund have been few (approximately thirty cases), and most importantly, the 'big, successful' cases have not applied." Watson, supra, at 276 (footnote omitted). Apparently, "class actions are alive and well in Ontario without a successfully functioning Fund"). Id. at 277. For a description of the Ontario fund's procedures and standards for awards, see WARD K. BRANCH, CLASS ACTIONS IN CANADA at 8-1 to 8-4 (2002).
50.	See Morabito, supra note 12, at 265 (Quebec fund's criteria include whether action can proceed without its assistance and, in case of actions not yet judicially certified, whether cause of action appears valid); id. at 266 ("Since contingency fee arrangements are permitted in Quebec, they can be combined with financial assistance" from the fund).
51.	See Garry D. Watson, Ontario's New Class Proceedings Legislation-An Analysis, in GUIDE TO CASE MANAGEMENT AND CLASS PROCEEDINGS 1, 7 (Garry D. Watson & M. McGowan eds., 1995) ("the most important and significant step in class proceedings may not be in court, but will be the class representative's application to the Class Proceedings Committee for what will be (in form) a request for disbursement funding, but is in reality a desire to obtain an immunity from the 'downside risk' of liability for the defendant's costs"). Significantly, British Columbia, which has virtually abolished fee shifting in class actions, see Watson, supra note 4, at 277, alone among the three Canadian provinces with long-established class actions apparently has no such fund.
52.	See, e.g., Abel, supra note 29.
53.	See, e.g., Faulk, supra note 15, at 210 (footnote omitted): The "entrepreneurship" of contingent fees, a hallmark of American class action litigation that promotes risk-taking by the plaintiffs' bar, seems flatly incompatible with the "fee shifting" tradition of most other nations, which discourages litigants and their counsel from taking similar risks.
54.	See, e.g., Michael Zander, If conditional fees, why not contingency fees?, 152 NEW L.J. 797, 798 (2002): It has been argued that if [contingency fees] were allowed, the English fee shifting rule would have to go and that each side would bear its own costs as in the USA. This is not a necessary consequence of permitting contingency fees. In Canada, every province except Ontario permits contingency fees in the full sense of a percentage of the damages but every province adheres to the English rule that costs follow the event. . . .
55.	See generally, e.g., DOUGLAS LAYCOCK, MODERN AMERICAN REMEDIES: CASES AND MATERIALS 930-32 (3d ed. 2002) (discussing theory of and issues arising under common-fund idea for attorney-fee recovery).
56.	See, e.g., id. at 921-24 (discussing lodestar calculations).
57.	See, e.g., id. at 932-33 (discussing contingency enhancements and their justifications; their rejection as an element of a shifted fee by U.S. Supreme Court in City of Burlington v. Dague, 505 U.S. 557 (1992); and their persistence in federal-court common-fund cases-in which awards come from the plaintiffs' side's recovery, rather than from the losing defendant-and in some state systems).
58.	The Australian state of Victoria limits the uplift to 25%, See Vince Morabito, Contingency fees in federal class actions, LAW INST. J., Dec. 1999, at 86, 87. In England-which does not have something closely resembling the class actions that exist in the United States, Canada, and Australia, see Andrews, supra note 39-the uplift is capped at 100%. See Woodroffe, supra note 18, at 353-54.
59.	See Report of the Civil Rules Advisory Committee, May 20, 2002, at 101-22 (proposing and describing amendments concerning class-action settlement approval, appointment of class counsel, and attorney fee awards) at http://www.uscourts.gov/rules/jc09-2002/CVRulesJC.pdf (visited Sep. 28, 2002); Ammendments to Federal Procedure Rules, 71 U.S.L.W. 4253-55 (2003) (Supreme Court promulgating class-action rule amendments, to take effect December 1, 2003).
60.	Kevin M. Clermont & John D. Currivan, Improving on the Contingent Fee, 63 CORNELL L. REV. 529, 547 (1978). Something resembling the hybrid fee may be used in Canada: "The court determines the fees for class counsel, typically using a combination of a multiplier test (i.e., hours worked multiplied by the hourly rate) and a percentage contingency fee." Watson, supra note 4, at 277 (footnote omitted). For critique of the Clermont & Currivan proposal with suggestions for refinements, see John C. Coffee, Jr., The Unfaithful Champion: The Plaintiff as Monitor in Shareholder Litigation, 48 LAW & CONTEMP. PROBS. 5, 44-46, 48 (Summer 1985).
61.	See supra note 59 and accompanying text (mentioning proposal to amend Federal Rule of Civil Procedure 23 to increase judicial scrutiny of proposed class-action settlements). For discussion of difficulties with judicial review of proposed settlements, and of possible improvements, see HENSLER ET AL., supra note 3, at 86-93, 486-90.
62.	See supra note 12 and text accompanying notes 21-27.
63.	Professor Watson reports that some class actions do take place in Ontario, which has loser-pays fee liability for class actions albeit with some exceptions plus a public fund that supports a few such cases, and speculates on how the problems of potential class-representative liability are being handled: Although it is not clear what is happening "out in the field," possibly plaintiff class counsels are not properly advising representative plaintiffs of the risks involved, are choosing judgment-proof plaintiffs, or are agreeing to indemnify the representative plaintiff for the costs of the action. If the former-if plaintiff class counsels are not informing their clients of the risks-a malpractice action might be necessary to clear the air. Watson, supra note 4, at 275. Clark and Harris report somewhat parallel evasion in Australia: Needless to say, plaintiffs' lawyers are often careful to nominate as the representative party a person of straw-that is to say, someone who has no assets and who is therefore incapable of satisfying any significant order for costs made in favor of the defendant. Defendants have responded to this tactic by seeking what is known as an order for security for costs against the plaintiff. This is an order that can be made by the court requiring a plaintiff to pay into court, or otherwise give security for, an amount equal to the estimated costs of the proceedings-including attorney's fees. Clark & Harris, supra note 5, at 302. See generally Vince Morabito, Security for Costs and Class Actions in the Federal Court of Australia, 20 CIV. JUST. Q. 225 (2001) (surveying and criticizing practice concerning security for costs). The tactics described by Clark and Harris reflect but do not seem to address adequately the underlying difficulty of the disincentives created by traditional cost and fee rules. Moreover, they raise questions about the adequacy of the class representative and the ethics of class counsel's conduct. If possible, it is preferable to face the problems directly and adopt approaches that do not elicit such maneuvers.
64.	See supra text accompanying notes 36-37 (discussing obstacles to having unnamed class members contribute to fees of class counsel).
65.	In both Ontario and Quebec, the unnamed class members can be liable only with respect to determinations on their individual claims. See, e.g., Terrence J. O'Sullivan, The future of financing class actions in Ontario: considerations arising from a review of bills 28 and 29, in CLASS ACTIONS IN ONTARIO AND QUEBEC: PROCEEDINGS OF THE FIRST YVES PRATTE CONFERENCE 81, 99 (A. Prujiner & J. Roy eds., 1992) (Ontario); Larry M. Fox, Liability for costs: a comparison of bill 28 and bill 29 and the Quebec legislation, in id. at 123, 136 (Quebec). British Columbia, the remaining Canadian province that has for some time provided for class litigation, follows the American rule for such actions, so that neither the class representative nor class members are generally liable for defense fees in a losing case. See, e.g., MICHAEL A. EIZENGA ET AL., CLASS ACTIONS LAW AND PRACTICE � 12.3, at 12.1-12.2 (1999) ("Costs may only be awarded in the event that the court concludes: that the class proceeding itself, or an appeal from a class proceeding, has been conducted in a frivolous, vexatious, or abusive manner" or in other limited circumstances); John A. Campion & Victoria A. Stewart, Class Actions: Procedure and Strategy, 19 ADVOCATES' Q. 20, 42 (1997).
66.	See Clark & Harris, supra note 5, at 301-02 n.72 ("There is currently no provision for members of an unsuccessful class to be ordered to contribute to costs, even where the plaintiff has insufficient finances to satisfy any costs order."); Morabito, supra note 12, at 239 ("party-party costs are to be borne only by the representative plaintiff and not by the class members"). Under England's new group-litigation procedure with its opt-in feature, by contrast, participation will involve liability for costs: Consistent with the basic [loser-pays] costs rule, members of a group claim will incur liability for the opponent's costs if the group loses the action. Group members' costs liability to the victorious opponent might be comp[o]sed of two elements. First, a group litigant's liability with respect to other members of the group will be "an equal proportion, together with all the other group litigants, of the common costs." . . . The second and additional element of costs payable by an individual group litigant is the amount of individual costs incurred by the defendant in meeting that particular litigant's claim. Andrews, supra note 39, at 261-62 (footnotes omitted) (quoting CIVIL PROCEDURE RULES 48.6A(4)(b) (U.K.)). See generally CHRISTOPHER HODGES, MULTI-PARTY ACTIONS 135-204 (2001) (discussing financing of multi-party actions in England and Wales via legal-expense insurance, conditional fee arrangements, and public funding).
67.	Hensler & Rowe, supra note 32, at 152-59. For a critique of our idea, see Marc I. Gross, Loser-Pays-or Whose "Fault" Is It Anyway: A Response to Hensler-Rowe's "Beyond 'It Just Ain't Worth It,'" 64 LAW & CONTEMP. PROBS. 163 (Spring/Summer 2001). See also Morabito, supra note 12, at 253 (liability of class counsel for winning defendant's fees "would exacerbate the conflict of interest problem and . . . would increase substantially the compensation that would need to be paid to the lawyer if the case succeeded.") (footnote omitted). See generally Dewees et al., supra note 12 (analyzing economic viability of alternative class-action fee-liability arrangements including some with class lawyer bearing risk, and finding that some such rules including loser-pays fee shifting with attorney liability could make for economically viable class actions without encouraging claims with low chance of success).
68.	See Hensler & Rowe, supra note 32, at 154-56.
69.	See id. at 156-59.
70.	Id. at 153.
71.	See, e.g., Gross, supra note 67.
72.	For a limited exception in Australia, see Clark & Harris, supra note 5, at 302 (describing case in which Federal Court awarded costs against counsel for losing plaintiff class when "the firm gave no consideration, or no proper consideration, to the question whether the federal claim had any prospect of success.") (footnote omitted).
73.	Hensler & Rowe, supra note 32, at 153.
74.	See, e.g., Abel, supra note 29, at 270 (citing much larger total amounts paid for legal-expense insurance premiums in France and Germany than in Britain).
75.	See id. at 272 (with Legal Aid cutbacks and launching of conditional fees, "[s]everal [British] companies offered plaintiffs insurance against liability for defendants' costs, selling over a thousand policies a month within a year.") (footnote omitted); id. at 302 (reporting premium increases and agreement to make premiums recoverable from losing defendants); Woodroffe, supra note 18, at 355-56 (describing "Accident Line Protect" policies available in Britain after Legal Aid cutbacks). See generally Hodges, supra note 29, at 336 (describing rationale and working of legal aid's exempting covered claims from usual loser-pays rule, and discussing prospect of insurance coverage for risk of fee liability in multi-party actions).
76.	See supra text accompanying notes 34-37.
77.	See supra text accompanying note 71.
78.	See supra notes 49, 51. It would also be conceivable to have the fund's choice to finance a class action mean that winning defendants would not be eligible to an award of fees from anyone on the plaintiffs' side, including the fund, as has been the practice in individual cases that receive legal aid under the British scheme. See, e.g., Woodroffe, supra note 18, at 347 (when "one party is legally aided . . . and the other is privately funded, the latter, when successful, will not normally receive his or her costs against the legal aid fund."). But it would be unfair to defendants to have their ability to recover fees when they win depend on whether the fund had chosen to finance their adversaries. If the loser-pays rule is to be relaxed or abolished for class actions, it would be better for that to come about on a less arbitrary basis.
79.	See supra note 48.
80.	See, e.g., Morabito, supra note 12, at 262 (while favoring public funding and abolition of loser-pays for Australian class actions, arguing against general one-way rule as unfairly imbalanced and smacking of punishment).
81.	But see South African Law Comm'n, supra note 11, at 70-72 (recommending court discretion to apply general loser-pays rule in context of proposed class-action regime, although with little discussion of impact of rule on viability of class actions).
82.	See OLRC REPORT, supra note 12, at 704-09 (recommending "no-way" costs rule except when "it would be unjust to deprive the successful party of costs" when reasonable plaintiff should not have sought class certification, id. at 709 (footnote omitted), and at any stage "in the event of vexatious, frivolous, or abusive conduct on the part of any party," id. (footnote omitted)). See also Morabito, supra note 12, at 255-63 (arguing for rejection of loser-pays rule in class actions).
83.	Watson, supra note 4, at 274-75 (footnotes omitted). For more detail on the background and content of the Quebec nominal-costs approach, see Fox, supra note 65, at 132-33.
84.	See Clark & Harris, supra note 5, at 301.
85.	See supra note 66.
86.	See supra note 63.
87.	See id.
88.	13 JAMES WM. MOORE ET AL., MOORE'S FEDERAL PRACTICE � 68.03[3], at 68-14 (3d ed. 2003) (mentioning also problems with trying to collect from class members, onerousness of imposing liability on class representative, and possible threat to zealous pursuit of class's claims).