Class Action Reform
The Why and the Who

October 30, 2003

Transcript prepared from a tape recording

Proceedings:
MR. GREVE: [In progress] -- with the AEI Liability Project. Thank you all for coming. We're here to talk about class actions. That's a very timely subject and a very big and sprawling subject. In fact, it's so sprawling and big that I should say right now we'll extend this event at the tail end by about five to ten minutes, so you'll be out of here by 12:10.

I'll say a few words at the outset to try to narrow this down a little and explain why we decided to put together this particular two-panel event in the way we have.

Among the precursors of the modern class action, ironically enough, or strangely enough, is British naval policy during the 18th century. Britain encouraged the more entrepreneurial members of its society to seize foreign ships on the open seas, and if you could bring these ships back to Britain, you could find yourself a court to divvy up the proceeds. And that pattern is rather familiar to modern defense lawyers, certainly seize some out-of-state corporation, drag it into a court, which, with distressing frequency, is situated in Morris County, Texas, or Jefferson County, Mississippi, and then you certainly can find yourself a court to divvy up the proceeds. We now call these population trial lawyers instead of privateers. But the basic idea is the same, with one critical distinction. So far as we can tell the privateer action worked, and that is, because they were bounded on both the plaintiff side and the defense side, you actually had to be a privateer to participate. You couldn't say that you wanted to be a privateer or you were once exposed to a privateer. You had to be one, and that bounded the plaintiff class. And at the same time, at the other end of the transaction, the actions were limited to the actual ship and its cargo. You couldn't say, hey, great, now that we have the ship, let's go after the sail maker.

And the second reason why they worked, that within the class you had a certain homogeneity of interests and a reasonable distribution formula that governed the disposition of the proceeds. In other words, these privateer actions, to the extent that they worked, resembled what we now call limited fund class actions, and those by almost uniform consensus still work, and the only debate is under what circumstances are those kinds of actions appropriate and permissible under the rules that we have.

What doesn't work by almost uniform consensus is what used to be called spurious class actions or what are now called sprawling class actions, (b)(3) actions, in other words, where you have unbounded and frequently very heterogeneous classes, always on the plaintiff side, sometimes on the defense side, and no limitations on the size of the fund, at least short of bankruptcy. Those actions, everybody thinks, don't work very well. You have exorbitant transactions costs. You have ruinous costs frequently for the defense side. You have absurd compensation formulas on the plaintiff side, so everybody is clamoring for reform.

With respect to reform, you have to ask yourself at least two questions, I think. The first question is: Reform for what? Do you want to drive down the transaction costs? Do you want to improve the terms one way or the other? Do you want to ensure compensation for the injured? Do you want to improve client control over lawyers? Which is the sort of foremost goal of scholars and the John Coffee school of thought.

You think about this for five minutes, and you realize that all of these goals may be in tension and sometimes in conflict, that what helps in one direction may hurt you in another. And that affects even the most common-sensical and consensual reforms. You have to ask yourself, it seems to me, before you wade into this: What is it actually that I want to do here? What is it actually that I want to accomplish? And what am I willing to give up for it? And that's in large measure a normative question about the purposes of class actions and the liability system at large. And that is the first or that is the topic of the first part or the first panel, which Frank Buckley will introduce in a minute.

No doubt the first panel will succeed in sorting all of that out, and then we can move on to the second question, which is: Suppose you have figured out what it is that you want, where are you likely to get it? And that's a question of institutional commitment and competence. It's a question of commitment because there seems to be a certain diffidence at the class action reform front all around, on the part of the Judicial Conference, on the part of the Supreme Court, on the part of the Congress. Into that breach has stepped the unlikeliest of candidates in many ways, the Federal Trade Commission.

You don't hear very many confident pronouncements, "This is our job to fix this problem, and here's why it's our job." You hear lots of confident pronouncements that it's somebody else's job. And the competence question has to do partially with expertise where I think every agency that is involved in this mini-movement one way or the other has some claim to expertise with the possible exception of the Congress. In other part, it's a matter of capacity and possible bias in these institutions.

The Supreme Court, arguably, underestimates the difficulties that trial courts confront in these mass tort actions. If you talk to trial attorneys and to trial judges, that seems to be a very confident position on their part. And, likewise, if you think of the Supreme Court as a sort of reform agency, it's bound to be a very poor monitor of its own reforms over time.

The Judicial Conference is sometimes suspected of having a bias to reduce transaction costs and move cases along and settle them come hell or high water, move these cases out of the system and get on with business. The Congress is frequently perceived as rather clueless on these matters. And it's also the institution that is most likely to cut an interest group bargain on class action reform that has little to do with the underlying merits. And while the FTC in its current composition is probably likely to take a sensible view on the settlements and proceedings that it comments on, one wonders whether the same might be true under President John Edwards.

So the long and short of it is that the potential or would-be class action reform confronts the dual task of finding somebody who will listen, and then to fight the institutional biases that might derail this train once it gets underway, if ever. And that's the subject of our second panel.

I now leave the first panel in the capable hands of Frank Buckley. Professor Buckley has taught at George Mason University Law School since 1989, I believe. He's the author of numerous books and articles in very prestigious journals, and we're very grateful to him for agreeing to serve as the moderator, and he'll introduce the combatants or participants.

Thank you.

MR. BUCKLEY: Thanks very much, Michael.

Well, I'm delighted to be here. I'm also the Director of the George Mason Law and Economics Center, and while I'm here and as we've mentioned George Mason, let me recognize my former colleague, Henry Butler, back there in the back row. I'm happy to see you, Henry. Welcome back.

I'm also happy that Michael mentioned privateers. My family got started in the privateer business sailing out of Quebec City and preying on fat Yankee merchantmen sailing out of Boston. I guess there is an analogy to class actions, but, frankly, our family saw this as a very good incentivizing device. So it is about incentivizing devices that we're gathered here today to talk about. The presenters are both strongly within the law and economics tradition of applying rational debate to vexed questions of public policy, and as a consequence, the differences are actually not as great as one would have thought. I mean, that is one of the virtues of law and economics, it seems to me, to civilize the debate.

By way of format, we'll have the presenters talk for about 20 minutes each. Then they'll have 10 minutes or so to respond to each other, and afterwards we'll open it up to the floor to questions.

We'll start first with David Rosenberg, who's been teaching at Harvard Law School for over 30 years and who is the author of this splendid new book. Let me hawk it. This is what moderators should do, is hawk books. "Making Tort Law," available today from, let's see, AEI. I think I know that place. So, David, why don't you start us off?

MR. ROSENBERG: Buy it.

[Laughter.]

MR. ROSENBERG: Let me just say--I should turn this on. Let me just say that we are at a very important moment in the debate of class actions because of what AEI has done in convening this session. It has been my experience over the years that, in a sense, the first principles and the basic purposes of the civil liability system have been largely neglected in the debate over whether we should have class actions or whether we should reform tort law to control liability.

We are now asked to think about why we have the system as a prelude to discussing how that system ought to be designed. So this is a moment that's been long in coming, and I thank AEI for opening the door to basic questions, a discussion of basic questions.

Now, let me say that because I lost my glasses, I tried to make things readable to me, my notes readable to me, so I've produced this slide show.

MR. : It says securities regulations rule 10(b)(5) [inaudible].

[Laughter.]

MR. ROSENBERG: No, no, Richard. It doesn't matter. So I'm going to try to follow this outline, whether I can see it or not.

Let me begin with this simple proposition: that when people talk about sprawling class actions, for example, I think the Supreme Court did use that term in connection with the asbestos settlement case, Amchem, and the problematic use of Rule 23(b)(3), those kinds of comments and the conclusions people, lots of people have drawn that (b)(3) class actions of this nature are unworkable, and everybody thinks so, to a large degree those kinds of conclusions stem from a failure to think from first principles. A very efficient, a very effective class action model can be easily developed and easily applied once we start from first principles, and that's going to be my effort to show.

I'm going to start with a very quick statement of what I consider to be the first principles, and I've styled it general normative theory of civil liability, and it's basically something as astounding as the proposition that the law ought to make people better off, ought to aim at enhancing individual well-being.

Now, if I can figure out how to get to that slide, if we start from this, in a sense, diantological(?) position, since you can't prove it but it seems quite sensible, once you start from that proposition, virtually everything else I'm going to say flows logically and necessarily.

What do people want? How do people make themselves better off in a world of risk and scarce resources? I think those are empirically testable propositions. What do they want? What do we observe from their behavior, including hundreds of billions of dollars, maybe trillions of dollars invested annually, net dollars invested annually out of their pockets? What they want is control of risk, regulation of risk, and insurance. And they want the two in as optimal a combination as society can deliver it.

We know that that's what people want. In a sense, the mission of civil liability flows directly from the proposition that the society wants what I'll call optimal risk regulation and allocation. Civil liability could provide it. We often call it optimal deterrence and optimal insurance. And I'll use the term "insurance" and not the term "compensation" because insurance is what people pay to receive, and people are paying quite a bit for the tort system, for example, or for the civil liability system to deliver payments or replacement for their losses in the event of--I'll use the term "accident," but I'm using it very capaciously to include any type of risk that civil liability might address, for example, the risk of fraud, the risk of racial discrimination, and so forth.

People want optimal risk regulation in combination with optimal insurance. We know that. A $200-plus-billion system of civil liability under the heading of tort alone should be devoted to what people want and not just what lawyers want. So the assumption is--not the assumption. The necessary conclusion is civil liability ought to contribute to the overall effort by the legal system to deliver what people want, and what people want is to minimize the sum of accident cost through optimal risk regulation, optimal insurance, and putting it strictly in terms of civil liability, optimal deterrence and optimal insurance.

Now, I want to say right now under (d), I assume that there is some utility to civil liability operated by courts. I assume it in a very strong way; that is, I don't want to take for granted the idea that courts, judges, lawyers, and, to the extent juries are used, jurors know what they're doing when it comes to the formation and execution of the public policy that's necessary to achieve these social objectives, optimal deterrence and optimal insurance. In fact, I'm pretty well convinced that our current system of civil liability can't achieve those results very effectively. The courts and the lawyers and so forth are not competent to deliver it. They don't have the resources to upgrade their capabilities. And as a result, if we're going to count on courts to deliver these goods, optimal deterrence and optimal insurance through civil liability, we're going to have to reconceive how people are educated and prepared to undertake the task for running society through the judicial system.

So let me pass on to the next stage of the argument I want to make. Normatively, the law ought to make people better off, and it does it according to what people want, and we know what they want, with some great degree of assurance. They want risk management at an optimal--as optimally as we can get it. And that means optimal regulation of risk, deterrence, optimal insurance of that part of the risk that can't otherwise be eliminated or shouldn't be eliminated. Theory applied.

Now, what I've stated is civil liability applied across the board without limitation. However, when you talk about class actions, there's a necessary context. The necessary context is the management of mass production risks, because basically what class actions are addressing is mass production decisionmaking by so-called private enterprise and by government enterprise. That's where it's all coming from. It's mass production decisions.

Now, mass production decisions have a very special characteristic about them. They don't have any individualizing characteristic to them at all. They're aggregate, average, and largely statistical. You can't break them apart with respect to any individual who might be affected. In fact, no individual would want them broken apart in that way because it would destroy the scale benefits of mass production. The very benefits that we receive from these mass-produced processes and goods is derived by uniform treatment of people on average, statistically. There is no individuality to this world, and if it were injected, it would destroy the very goods that people, in fact, want to maximize their well-being.

So we are thinking about an area of life called mass production. You can give it any other name you want, but I think that captures the idea. What's being done is that we're producing collective goods. These are goods that don't have any individualizing characteristic to them at all. Or I'm saying that in principle--at all. And the only means by which you can produce those kinds of collective goods is through collective production, unifying--sorry, uniform design, uniform design of processes, uniform design of the products, uniform design of the services. And I'm being very capacious in my view, very broad in my view of what these goods are.

Now, it turns out, because of what I've just said, that the secondary goods--I call them primary and secondary just for purposes of division, not in any attempt to give them any priority other than for my outline purposes. Secondary goods I'll call the goods that are produced by the legal system, deterrence and insurance. We've talked about those. It turns out to produce the--to safeguard the collective goods that are primary, the processes and products and services of mass production, to safeguard those the regulatory system and the insurance system that the law operates has to be collective as well. It can't be disaggregating, it can't be individualizing, without destroying the very things that we prize for mass production. So deterrence has to be a collectivized product, and insurance has to be.

Now, when I say this, let me bring it down to some reality. Deterrence is a common good. If you deter unreasonable risk-taking for one person, you've deterred it for everybody who might be affected by a mass production risk. So, in effect, this good can only be produced by some collectivizing means that gives the producer of the good the proprietary interest to produce it for everyone concerned.

The same with insurance. Unless there's universal risk pooling, the law of large numbers can't be achieved to provide the basis for the insurance system to operate effectively, whether it's social or commercially based. And you can't have people opting out of the pool without potentially destroying it.

Now, there's another basis on which--and I think it's related to what I've just said. Scale economy and investment requires collectivization. Now, what do I mean by that, collectivization in the production of secondary goods? What I mean by that is because we're dealing with mass production risks and because the central questions of regulation and insurance relate to a central design that affects every potential person at risk, there are scale benefits to be achieved through united effort in managing the legal system to achieve regulation or deterrence. You can invest once on a claim and invest on all potential claims. I'm using the word "claim" here--it's almost begging the question. You invest once to achieve regulation, and you can achieve regulation across the board. Scale economies.

But there's another feature of scale that's often forgotten. It's called or I've called it scale investment. If you invest in all of the benefit, achieving all of the benefit from regulation and insurance, you're likely to invest more productively, more effectively, and qualitatively better than if you're only investing in a fraction of that benefit. Very simply, if you're hiring an expert witness and you have one claim worth $100, the limit on what you might pay that expert will be something less than $100. The expected value of the claim would be something less than $100. You'll get an expert of some quality, but it won't be particularly good. And the chance and the probability of success or the probability of effective law enforcement will drop.

But if there are a thousand claims, each for $100, and you own them all, $100,000 expected value, the investment in the expert is likely to go up. And the way things work, probably the quality and probably the effectiveness of enforcement that results.

What we have in our current system is a basic skew in the investment opportunity--scale investment opportunities that the parties have in litigation. On the mass producer side, the mass producer owns the value of its defense in every single case, every claim, and will invest accordingly. As I've put it, the mass producer has a de facto class action, advantage in investment.

Compare that to a so-called plaintiff's attorney investor who might have only a fraction of the claims on the plaintiff side. That investor is not going to optimize the scale opportunities of having--that the defendant can. You're going to have a lopsided, skewed, biased system of investment opportunity. On average, that's going to distort the results for optimal--for achieving the objectives of optimal deterrence and insurance.

So when I talk about collective means, I'm talking about collective means not in any conceptual sense, but in a purely functional sense. Let me say one more word about the secondary goods--I've put it up here--called autonomy. You could call it individual claim ownership. You could call individual justice. You could call it fairness. You could call it anything you like. It's total nonsense. It's total nonsense because, as I've suggested, to begin with, in these kinds of cases no individual has anything to say about his or her relationship to the mass producer of the risk. It was a unified, uniform, inseparable, statistically based, aggregate decision that didn't even take account of--didn't necessarily take account of any given individual, and probably shouldn't. And in the end, to do the work of insurance and deterrence, we don't even have to pay attention to what happened. All you have to do is look back at what the level of investment in precautions should have been, assess the aggregate sanction that is appropriate if the investment level was inadequate, and then distribute the funds according to insurance needs.

I'm told I have 20 minutes left.

[Laughter.]

MR. : Just because you speak [inaudible].

MR. ROSENBERG: All right. So, very quickly, the class action ought to be mandatory, it ought to be automatic, and it ought to be universal. There should be absolutely no question that class action should be certified. There should be no questions of predominance. There should be no questions of homogeneity versus heterogeneity in the claims. It's all hogwash if the objectives are deterrence and insurance. Insurance doesn't want to know anything about any individual claimant's causal relationship to the harm if that issue doesn't affect the level or severity of the loss that that individual has suffered. Insurance doesn't want to pay any attention to that, and we shouldn't.

I just want to pre-empt Richard and everybody else by saying that you can design a class action to achieve any number of things you want consistent with the basic social objective, which would be give priority to deterrence over all other interests. So I can design--and this is just to say that I'll be heard from again with respect to problems that somebody might raise. I can design a class action that can allow plaintiffs to recover as they do now--a mandatory class action, no opt-out. They can recover as they do now. They can play the wealth-maximizing game to their heart's content under the banner of autonomy. I can design a class action where that can go on, but before that happens, optimal deterrence is achieved in the aggregate by assessing the sanction that should be imposed for the tortious harm.

Blackmail settlements are easily solved. Sweetheart settlement problems are easily solved. Nuisance value settlements where a class claim is brought simply for the price of--something short of the price of throwing it out, that can be easily solved. And so I just want you to know that it can be done.

MR. BUCKLEY: Well, you can do a lot. Glendonor (ph) said, "I can call creatures out of a starry deep," and Hos-(?) said, "Ah, you can call them, but will they come?"

Our next speaker is Richard Epstein. In the tradition of plugging books, let me plug "Skepticism and Freedom," the last in a trilogy of books on the philosophical foundations of law and economics. And Richard's a splendid person. I will keep him strictly within 20 minutes, I swear to God.

MR. EPSTEIN: I never go over time. I'll just speak fast.

I'm going to just read one sentence from David's book because I think it sort of sets the nature of the discussion that we're having. He says, "If we were working from scratch, tort law would probably not be in the picture, certainly would have nothing to do with accident compensation." And, in effect, what's happened here is we are seeing the results of that.

What David has done is something which I am not going to do. What he has done, in effect, is he's asked the question: If we were going to start to build the system on harm, what ought we to do in order to make it work well and would have the idea that compensation is always handled by insurance on the first-party basis and that deterrence is always handled by some kind of collective action, probably by a fine or by some sort of administrative procedure, so you wouldn't have the tort law at all. And there is some sense to that, I suppose, in a certain way.

But the discussion we're having here is not, I think, about the foundations of the legal system, writ large. It's about the question of how it is that we do class action law. And in order to understand what's going on, I'm going to go back to an old distinction which said that when you were dealing with these various issues, that you would treat the substantive law as essentially the primary law, and procedural law was essentially called adjectival. And what that meant, in effect, is that if you figure out what it is that the substantive law believes as its major goals and objectives and then whether you agree with those particular goals or not, the procedural task is to try to find a way to realize the substantive objectives that have been created by the legal system.

And this means, in effect, that the law of class actions, like any portion of the law of civil procedure, is not meant nor should it be used to try and achieve some sort of a covert reversal of what the applicable substantive principles are going to be.

If it turned out that we think those principles are wrong, what we ought to do is to not fiddle around with the stuff under Rule 23, but we, rather, ought to go back to our state capitals or to our Congress and to argue that we have to reverse this.

Now, if one actually looks at the structure of the rules and so forth, I think that you'll find that that is, in fact, the way in which we've done it. Lee Rosenthal is sitting here. She's running the Rule 23 Committee, and I've actually attended one or two sessions of it. And every time the question comes up whether or not you make a change in substantive law, the issue always comes back that the rule's enabling act only allows us to deal with matters of procedure, and those particular kinds of issues, as important as they may be, are simply beyond the purview of our particular results.

If you actually look at Rule 23 and the way in which the thing is organized, it is remarkably faithful to that particular conception, and what it does is try to figure out under various circumstances what sort of class actions should be allowed and why they ought to be allowed or denied and so forth.

So essentially the way in which I view a class action is it has exactly the opposite function of what David wants to say. It's designed to preserve as much as you possibly can the substantive law, be it for good or for evil, as it exists outside the class action framework. And so let's just go back briefly and indicate what that substantive law starts with, and I think that you will see what the sources of the difference between us are.

Essentially if you want to go back to the traditions of substantive law, it starts with what has been termed "hogwash." It starts with the general principles that individual autonomy means an enormous amount in this world, that autonomy is a substantive matter, means that each person ought to be entitled to figure out how to lead his or her own life within the confines that they respect the like liberties of other individuals. It turns out that that principle is translated to if somebody invades your space by the use of force or by fraud, if somebody engages in actions which breach a contract that they've made with you, you are now declared to be a victim, a holder of a particular right, and the legal system is designed to vindicate the rights that you had in your autonomy to the extent that those rights have been violated by other individuals.

So in the ordinary civil action, the general question is how it is that you find the proper plaintiff, i.e., the injured party, to get a mechanism which allows you to sort out the particular claims. This means, in effect, that you're desperately trying to find a situation in which the accurate state of affairs as it existed in the primary world can be reflected in litigation. It means that you have to figure out how you plead a case, how you organize a substantive law, how you present evidence on that law, so as to realize the substantive vision that essentially has animated our law, for better or for worse, about how it is that individual autonomy is going to be protected under various circumstances.

Now, what we understand quickly about this system is that, however noble the ideals that we have as a matter of first principle, it turns out that the greatest difficulty that every legal system has to face at every stage in which it wants to operate can be summarized in one word: friction. Or if you want to get a Nobel Prize, you will call it transactions cost and become Ronald Coates (ph).

And what you're trying to do, in effect, is to see how many of these substantive claims that you can vindicate in this world, when friction seems to occupy at every turn. And for this, of course, it means that in many cases the story of substantive issues will become less important in day-to-day operations than the story of friction, which constantly interferes with its operation.

What do we do when we start to deal with friction? Well, in many cases, it turns out it clearly goes back and it influences the way in which we think about the substantive law. There are, for example, doctrines, like the mere fright doctrine, which say although you may have been frightened by somebody else, if you don't have any lasting physiological complications associated with that fright, we're not going to compensate you for that injury, mainly because we think the frequency of suits would be so extraordinarily high and the cost of bringing individual suits would be so extraordinarily great that what we just do is rule these things out altogether.

Somebody then will start to come back and say, But, you know, if you really have large numbers of small diffuse harms for which individual suits cannot be brought, then, in effect, what's happened is the defendant is allowed to do something which is wrong as you would perceive it, and he could get away with it. What could we do in order to stop that particular kind of situation?

And the usual answers are two: One of them is you could say we could give up on the ideal of compensation because it's just too complicated to handle, and we could have some sort of a system of civil fines that will deal with the wrongdoer without making any effort for compensation to the individuals on the grounds that it's too costly. And this is, in fact, perfectly in line with what David wanted to say, although it comes with a slightly different source. And there what you do is you say, Look, if we have the deterrent effect, all individuals would benefit in their individual capacity by virtue of the reduction of the number of suits that we have or the number of wrongs that we have, and, in effect, that the money when it goes into the public treasury could be spent on something which benefits the public at large, so there's a kind of indirect compensation.

Now, whether or not this system would work depends upon the question of whether or not you've introduced a second set of frictions at the administrative law level to displace those that you have with respect to the primary conduct. And in many cases, it turns out that that's true. If you're trying to figure out how OSHA, for example, handles these low-level harms, you would come, I think, quickly to the conclusion that doing nothing is sometimes the course of wisdom, mainly because the administrative distortions that you get in trying to deal with these mass production wrongs are, in fact, great in their own way.

In other cases, however, what we decide to do is to aggregate the individual claims and to allow them to be sued through a class action. And the way in which I understand this particular transaction is rather different from the way in which David understands it. What I understand it to be is as follows: We've already made this substantive commitment to hogwash, to individual autonomy. Now what's going to happen is to run the class action or to run the group action, we have to find ways to get people together.

One of the ways in which you could do it is through voluntary joinder, and, in fact, this is extremely important in many areas and works quite well, or quite disastrously, depending on your point of view. One should remember, for example, that the asbestos cases for the most part are not brought as class actions. They are brought as permissive joinder cases where, in fact, all the problems of voluntary coordination on the plaintiff side are solved by ingenious agreements, many of which involve the use of intermediaries like unions, for example, and so forth, who bring people together. The plaintiffs' bar may start out in a fragmented way, but they're terrific at dealing with one another and having various kinds of cooperative arrangements across the lines, capital funds and so forth. So that what they do is they are essentially, like good libertarians, able to use all sorts of ingenious devices of one kind or another to achieve a voluntary form of amalgamation. And before one starts to say that you've just got to go to the class action in order to make sure you can get people together, I think it actually pays a little bit of attention to look to see the way in which this has been done.

Now, what are the conditions under which these voluntary agreements take place? Well, the general view is quite invariant with respect to this. You cannot get these large assemblages of individuals to come together unless everybody in them is left better off than he was by staying outside of the agreement. And it turns out typically what happens through these voluntary agreements is there is a certain amount of averaging that takes place precisely for the reason that David said. I don't want to spend all my time individuating my claim if the net effect of it means that I have to stay outside the pool and recover nothing. And so the standard voluntary plaintiffs pool gives the attorneys the rights to allocate proceeds across a bunch of individual plaintiffs in what one wants to call a kind of a batch settlement. And that's the terms that they use in the trade to figure out what has happened.

Now, I think, in effect, there is nothing that any good consistent libertarian could say against this form of ingenuity. The problem comes on the other end, which is whether or not the underlying merits of the suit turn out to be sound once you aggregate or don't aggregate the claim. And so one of the reasons why I find it so difficult to deal with the class action question generally is though I am sure that aggregation leads to effective magnification, I am not sure that it leads to a social improvement. The truth about the matter is when we start to deal with mass production wrongs of one form or another, we don't know how to deal with them particularly well. And in the asbestos-type situations, there are just huge questions of system design, one of which asks whether not the workmen's compensation system ought to displace the tort system entirely, which was the rule in the United States before the Burrell (ph) case in 1973. Another has to do with the nature and the quality of the medical evidence in these cases where it turns out that there seems to be rather systematic overcompensation with respect to certain kinds of wrongs. And if you're trying to figure out what the optimal level of investment is, you have to figure out what the anticipated harm is. And if you systematically miscalibrate certain kinds of injuries, well, it turns out that you're going to get that wrong in any individual case. And if you get it wrong in an individual case, you're going to get it wrong in the aggregate. And if you're doing aggregates, it's going to be much more dangerous because the errors will never cancel themselves out. They're only going to cumulate in one form or another.

But, anyhow, we live with that. Whether there's a ,problem or not I think we could answer quite clearly. You look down in Washington with respect to what's happening in the asbestos case, and you've got defendants and their insurers who are willing to pony up something close to $100 billion to be rid of the class action system, and you get the plaintiffs' guys saying, well, hey, this isn't big enough but we're willing to go with this so long as you can make a few other adjustments, like if you don't pay we go right back to the tort system again.

This is a huge battle. It's not taking place because everybody's thrilled with the class action. But what I think it's taking place with or it's not facing up fully is that if you actually look at the underlying substantive rules, nobody, at least I think no independent observer, thinks that you're getting the optimal level of deterrence because they think that there's systematic overcompensation of claims because of a miscalibration of what happens on the medical side. And before you go into any form of aggregation, be it a class action or a permissive joinder, remember this: If you're making a single mistake in one case, that's a problem. You make the same mistake in a billion cases or a million cases, it's a million times bigger problem than you had. And so you just have to be very, very sure of what you're doing when you move this way.

Now, what happens with class actions is they don't work in all cases by these permissive joinders. Sometimes there will be holdouts or whatever it is, and so we go to the class action. And the model of the class action is to use coercion to emulate the result that you get with permissive joinder. And that means, in effect, that at least if the darn thing is working particularly well, everybody is told that they have to surrender their individual claim to the class in a world in which there's no opt-out, and in exchange, they receive something from the class which is equal or greater than that which they've put in.

Are there cases where this particular condition works reasonably well? Well, the answer is it probably works pretty well in the (b)(1) and the (b)(2) kinds of class actions, that is, those types of situations where what you're looking for is collective relief. And that's the situation where, for example, somebody raids the corporation, makes a phony deal, and gives the money away, and you bring a derivative class action, the purpose of which is to restore to the corporation that would has been taken from him. And the reason you allow the class action there is because that is a real public good in the sense that you cannot restore the money to the corporation for the benefit of one shareholder unless you restore it to the benefit of the corporation for every other shareholder. And so what you have to do is to tax the entire firm, to wit, the corporation, so as to make sure that when the restorations take place, you eliminate the issue of free riders.

Oddly enough, when you start dealing with these individual tort actions and people want to opt out, the free rider problem, it seems to me, is much less severe under these circumstances; rather, what you're saying, in effect, is class actions involve all sorts of tactical and important business decisions, I don't like the business decisions that are being made by the class representative, I think I can do these things better on my own. And if the case is large enough to make a difference, it seems to me that we ought to follow the general proposition, which I think is the current law, that if somebody is entitled to bring an individual claim, you don't want that claim to be snuffed out because somebody else wants to bring a class action. Opt-out then becomes an important way in which you can preserve that autonomy and to discipline the choice of class action representatives.

Indeed, I am told by some very good plaintiffs' class action lawyers that the opt-out, in fact, is extremely important because one of the things that it does is it means that when the judges are starting to choose who's going to be the class representatives, they're not going to be able to choose weak lawyers in order to do that result, which they might do under other circumstances for a whole variety of reasons, because the moment the weak lawyer gets chosen, the other class action lawyers in competition try to go after the clients in there and to form classes out of the original class on a permissive joinder basis.

Why are they doing this? Because essentially they don't have any confidence in the sort of non-diversified, monolithic production that you get when you say there's a mandatory class action. And what David has to be able to solve and everybody has to be able to solve is how it is that you pick the right jockey for the horse when you decide that you're not just riding a single horse, but you're riding a whole miserable herd of horses, and if you got the wrong guy on that thing, it's an enormous disservice with respect to the way in which the class is running. And opting-out becomes extremely important.

So, in my view, just to sum up on how all this thing ought to be done, I think that there are lots of difficulties that are associated here with trying to make the class action system work, even after you understand what its operations are. And these include all the problems of getting the right class of representatives, trying to make sure that you get the right substantive theories and so forth.

If I had more time, I would start to indicate to you what I think to be some of the real dangers of a class action, which not only include sweetheart settlements and the like, but the inveterate tendency on the part of judges to change the substantive law to facilitate the class action so that they systematically make things easier than they ought to be for plaintiffs, generally speaking, and, therefore, introduce a bias through aggregation. David's view, in effect, is that the law remains relatively fixed in its content when you put it into a class action format, so it's the same body of rules that get better enforcement rather than none. My own experience, having worked with these particular cases, is that that is indeed not the case, that judges, once they commit themselves to the class action as a form, will now let the tail wag the dog and will change the substantive law underneath them so as to facilitate settlement in ways that -- [tape ends].

-- make all the problems that we've experienced for years and years and years disappear by a wave of the wand is wishful thinking. The reason why these problems are so difficult is that everybody who wants to play these games is looking out for himself. And when you have this complicated system of coerced interaction, loaded by a heavy dose of powerful individual self-interest, the one thing that you can be sure of is mayhem. And nothing that we do as an accurate--as an intellectual law, as an abstract matter is going to solve that problem at the ground level. Rules only get you so far. But in the end, the slippage is enormous, and that's just a general thing. When I talk in the next 10 minutes or so at the end, I'll try to explain what's wrong with the mass production logic, both as a matter of industrial organization and as a matter of class actions.

Thank you.

MR. BUCKLEY: Thank you, Richard.

I think for the response, I had said 10 minutes, if we could actually keep it, let's say, 5 to 10, and if you keep it closer to 5, we'll have more time for questions.

MR. ROSENBERG: [inaudible] and respond to only one point that Richard raised, recognizing that he made many very good points.

The one thing I want to start with is the notion that somehow tort law is founded on these individual claims about a person--what a person's space means and is worth. There's no reason to believe that that is the source of the concepts of tort claims. Tort law, after all, was a means of achieving the ends of criminal law to begin with. And it's not obvious at all that there was this dominant concept that Richard has.

But, in any event, the motivating concept for mass production civil liability stems from judicial decisions in the state courts that are conceiving of the function of tort liability not in any atomistic sense of individual rights and duties, but in the terms that I've suggested. If you go back to Trainor's (ph) animating a decision in the Escola (ph) case, which I think launches our modern era of product liability, at least catalyzed it, you can hear exactly the same claims about what civil liability ought to be doing that I was making, deterrence and insurance.

Now, there's no talk about an individual's claim. There's the conception of a system of civil liability delivering the goods of regulation of risk and insurance of those risks that can't be otherwise eliminated.

What I would say about Trainor is that he, unfortunately, didn't have access to modern theories of public choice, of rational choice, deterrence theory, law enforcement, and insurance. He did the best he could. We know far better now. Insurance is delivered much better by first-party social and commercial suppliers, and if the tort system can do anything, it's deterrence. That's not to say that I think it can do it effectively. So the rest of Richard's or many of Richard's complaints about distortions that result from what judges do are accepted by me as a starting point for further analysis. I don't claim the tort system is going to be a wonder at achieving deterrence. But if it's got on goal to--if it's got anything to do, it's deterrence and that requires the collective investment that only class action on a mandatory basis can supply.

MR. BUCKLEY: Thank you, David.

Richard?

MR. EPSTEIN: The Trainor opinion is quite remarkable for a number of reasons. The chief and the most important blunder that it makes is it doesn't understand that most of these problems can be solved by contractual solutions, and quite the opposite of any sensible situation, what it did and what the later Trainor decisions did, and the similar decisions in Henningson (ph), is they killed the market for product warranties by making them all illegal. And if you're trying to figure out where you start with and you're beginning with this better-off universe that David refers to, the first principle of that better-off universe is that contract dominates tort, not the other way around. And that's the fundamental mistake that Trainor made.

Even within the framework of the tort, the opinion in fact, has two parts. The famous part is the first 95 percent of the opinion. The last sentence of it, of course, goes right back to traditional conception, and what it says is that all this product liability presupposes that the product is in the individual condition that it left the manufacturer's hands in and is subject to damage only in normal and proper use. And what has happened is essentially all of these limitations have been eliminated by subsequent decisions so that the stabilization features which made his strict liability rule very effective and effective in change. Downstream risks of product modification now fall to the manufacturer. They don't fall, in effect, to the consumer, to the middle party. So that what you do is you have a substantive body of rule which systematically seeks out the worst cost avoider and holds him responsible for the majority of the injuries. And one of the dangers that you get with respect to class actions is when you get a body of law which is so utterly ignorant, that is, he didn't know any economics and he got it tolerably well. The guys who were supposed to know it, the modern judges, they get it so far wrong that it's really painful.

So beware of this when you want to talk about mass investment in the torts, that you're getting a rummy body of law out there which you're trying to deify.

So that I think is--

[Laughter.]

MR. EPSTEIN: --the real problem, is that the tort law in product liability is just horrible from the point of view of what David thinks the system ought to do. Let's get ourselves back to the class action again. Mass production is a misnomer. That used to be the way in which things were produced in 1890. It's not the way things are produced today. If somebody is serious, what you realize is that mass production in effect is going to be an inefficient mode of production to the extent that there's heterogeneity within the consumer class. And so what everybody tries to do in this mass production universe is to allow for individuation to take place within the confines of a product lot.

And so what you do is you see products that start having menus, and what the menus say is, well, you want to buy this on your PC, you could have this kind of a hard drive, this or this, this kind or that, there are 27 different variables, each of which have 14 different choices. It's all mass produced because the moment you press the button, they could figure out which of those products you're going to get out with. If you're trying to figure out what the greatest advances in drugs will be, it is surely in the next generation, not that everybody gets the same tablet of aureomycin. It's that you can figure out a way to make the biologics person-specific by having the same kind of menu protocol.

Now, once you understand that mass production today means how it is that you want large industries to produce individuated goods that respond to the particular variations of individual consumers, put that model against the class action and ask yourself, if it turns out that there are 79 different kinds of tires that you can put on your pick-up truck, do you want to have a class action which covers all 79 different types of tires? And for somebody who doesn't believe that this is a serious problem, read the Easterbrook opinion in the Firestone-Bridgestone case where what he does is he goes through what mass production in America looks like today--different tires, different warranties, different cars. You put this tire on that car, it's going to have a different effect than putting it on the other car. And when somebody says, well, just go after Firestone, what? I mean, what? What are you supposed to say? I mean, what's the optimal level of investment that you're looking to?

The facts of that particular case just defeat the class action, and what the right thing to do is if you're playing this is ask yourself in the case of an individual tire, Did it fail? And if you start giving these crazy class actions for refunds or for damages for people who weren't hurt, what happens is you're doing a game in which you're looking at only one decision made by a firm, the question of whether or not we have tires or something of that sort, and are using that to turn thousands upon millions of dollars' worth of liability. Essentially what happens is when you try to aggregate these claims, you make sure that all of them are completely unrecognized.

And there is no doubt that if you want to regulate tires and you want to have some government agency which says that the tread has to be this thick or it has to withstand that stuff, do it on an ex ante basis. But the correct institutional solution is you set the standards by statute, and anyone who deviates from the standard can be sued individually for the wrongs in question.

But what happens in America today is if you meet the standard, somebody could dream up after the fact the theory of liability of what you ought to have done. And when you have a system of substantive law that's this crazy, you cannot expect to have a system of mass production class action law superimposed upon it in ways that are going to make things work out together. It just can't happen. You've got to start from scratch--David's right about that--and rethink it. But the place to rethink is the interface between regulation and private litigation, which we've gotten wrong in the United States since at least 1970.

MR. BUCKLEY: Thank you very much, Richard.

We now have an opportunity for questions from the floor. If you want to ask a question, raise your hand. We'll come around with a mike to you. And I'd ask you to identify yourself.

Questions?

MR. EPSTEIN: Somebody. What a shy bunch.

MR. BUCKLEY: Well, okay. We have one over there. I can ask a question, but I'd like you guys to.

MR. : [inaudible] for Dr. Rosenberg. [inaudible, off mike]. And that would seem to be a problem in class action because usually that would involve some trade-off between the amount of money [inaudible].

MR. ROSENBERG: Good. Yes, I'm glad you raise it. The only reason I focus on use of monetary sanctions is because that's a fairly standard mode of civil liability. But your point is absolutely correct. If the civil liability system is going to do regulatory work, then it has to do it as effectively as it can be done. And if that requires a mix of ex ante standard setting, ex post fines, and so forth, and even ex post future controls, then that's what the courts will have to learn how to do.

Now, again, I'm not going to let Richard put me in this box. I start with the assumption that there's a need for judicial regulation because of some failure in the regulatory system, and there are public choice arguments why that might be a wise way to view the civil liability system as a backstop.

But it's a proviso. It simply says if we've got to have the courts do something, then this is--then the courts will have to fill the gap the way a regulatory agency would if we could have gotten one to do it. And the courts have to be educated and lawyers have to be educated to run that civil liability system accordingly.

So I'm not excluding the full range of regulatory devices. It needs to be brought into play.

MR. : [inaudible].

MR. ROSENBERG: Now you're bringing in the question of whether we have--well, I treat it in my book, and I gather the next panel is going to be talking to some degree about it. But, yes, I mean, this is it.

MR. : [inaudible].

MR. ROSENBERG: I've got--basically what has to be worked out are a set of default rules for when the courts--when judicial regulation is appropriate in line with what the regulatory agencies have or propose to do. So that kind of interface has also got to be worked out.

MR. EPSTEIN: Look, I mean, I'm going to put the answer in a slightly different fashion, and I don't know whether you want to count this as agreeing with one side or the other. But forget about the class action for a second. Any single individual acting alone can try to seek various kinds of injunctive relief, and you then have the terrible question: If two people or two different classes try to seek it in slightly different form when the agency itself has got a different view on this particular issue, how do you reconcile these individual inconsistencies with respect to approach?

I mean, the problem about any injunctive, it's always going to be broader than the individual plaintiff who brings it, whether you're talking about an ordinary nuisance case or anything else.

My own inclination on these types of situations is where you have a public agency which is charged with injunctive power, I would clean up the mess by denying private relief and saying, in effect, that those private parties can petition the agency for relief, and then you can have a hearing as to whether or not it ought to be granted. But I think, in effect, one of the difficulties with this discussion is that we sort of assume that courts should be plenipotentiaries on this particular matter, when, in fact, I think that private actions have to be necessarily circumscribed under circumstances in which strong regulation, say, for the prevention of force or fraud--you know, the FTC's supposed to do that, and sometimes it even does. I think that at that point you really want to switch it back into the opposite direction.

Your problem, though, is a real one. But let's put it in its simplest fashion. You win injunctive relief against a corporation that's about to waste assets in a derivatives suit. You still have to figure out how much compensation you give to the lawyer--right?--under these circumstances. And what you have to do is to figure out you value the injunction, and that's extremely difficult to do because the reason you got the injunction in the first place is that it would be very difficult to value the harm if it were allowed to ensue.

So you can't escape that particular problem, and it just shows that David's sort of global position on the world I think is going to be constantly undercut by a thousand cuts of difficulties and anticipations with respect to administration. And if you work at the middle level, it's not going to give you much guidance.

MR. BUCKLEY: Questions? Go ahead.

MR. DUFFIELD: Hi. I'm Steven Duffield (ph).

MR. ROSENBERG: Oh, Steven.

MR. DUFFIELD: Constitutional law professor here, so I've heard this many times. (?) talks about tort law and con law.

MR. ROSENBERG: And con law and tort law.

MR. DUFFIELD: My question is for Professor Epstein.

MR. EPSTEIN: He stacked the audience.

[Laughter.]

MR. ROSENBERG: I didn't even know he was here. Not only that, I didn't even recognize him for a second because he looks so much more relaxed than he did when he was a law student.

[Laughter.]

MR. DUFFIELD: And he always gets the last word.

The question is--I'd turn the tables on you, Professor Epstein, a little bit and to ask: What is the core of class actions that should be preserved and enhanced? In other words, what is the best case for a class action, you know, in the basic structure that we have it now? And the next-layer question would be: What's the best case for a (b)(3) class action, if you believe we should have one?

MR. EPSTEIN: Yes, I mean, I think that's a great question. Well, it's clear that if you go back to the history of class actions that essentially the soundness of the class action is pretty accurately mirrored by the period of historical times of their introduction. The original class actions were essentially inventions of the court of equity to deal with voluntary partnerships, corporations, charitable associations, and so forth, in which a single member was allowed to petition the court for injunctive relief or restorative relief against the misdeed that the organization did. And the theory of the class action was it was too costly on transactions costs grounds to get all of the individual shareholders and members to do this. But once you got the relief, all the other people would benefit because it's a classic public good. You cannot restore the money to the corporation for the benefit of the individual shareholder unless you restore it for all shareholders. So that seemed to me to work pretty well.

It's amazing. There are difficulties with class actions in securities business, but the precise argument there is the magnification effect. If you've got some doubts, as many people do, as to whether or not the rules on disclosure under 10(b)(5) make any sense--and I have enormous doubts about that. Then when you decide to use the class action to magnify this stuff, what you're doing is you're taking a bad substantive rule and multiplying it by, you know, 10⁴.

You know, my attitude is--I will give you my reform in the securities law. This is what I would tell the Senate. Abolish the whole thing and tell the small investor to buy a mutual fund. I mean, that essentially is the right answer for that question, and the class action stuff simply doesn't do it.

With respect to the private actions, the interesting thing about this is the area where the class action works best is the area where parens patriae suits work best as well, and it becomes very hard to choose between them. You get a low-level fraud against huge numbers of consumers, none of whom will bring a dime with respect to it. And at that particular point, a government suit or a private action will have deterrence under circumstances where compensation would never take place in practice, even if the class action were abolished.

The cases where I think it actually works the best is--I have been on the plaintiff side in antitrust cases having to do with overcharges, for example, as a result of monopolistic practices or rate cases, and there I think David's assumption is pretty good, which is that you get some guy selling grain or shipping something and it's a perfectly standardized commodity. The monopoly rate is $1, and the competitive rate is 50 cents, and what you're trying to figure out is how much of that overcharge you're going to refund and why. And I think under those circumstances the antitrust class action will work pretty well.

I think the worst cases are the anti-discrimination cases. Mass production goods, I mean, when you sue a firm with 72 different decisions on the ground that all of its employment decisions were animated by race, when no two decisions were made by the same person, and the Second Circuit in the Robertson case says, This is a class action? Well, it seems to me they're on steroid.

[Laughter.]

MR. EPSTEIN: David, I'm just curious. You agree with that, I take it, right? I mean, you would deny class action certification under Title 7 [inaudible].

MR. ROSENBERG: No.

MR. EPSTEIN: [inaudible].

MR. ROSENBERG: Of course.

[Laughter.]

MR. EPSTEIN: Are you mad?

MR. BUCKLEY: You don't have to answer that.

MR. ROSENBERG: Well, no, I'd like to.

MR. EPSTEIN: But, David, what's the common decision?

MR. ROSENBERG: First of all, they're not alleging that they are disparate decisions. They're alleging that there's a common animating illicit policy.

MR. EPSTEIN: Which is?

MR. ROSENBERG: Discriminate, and discriminate any chance you can get and any way you can get it.

MR. EPSTEIN: That's not what the allegation was. The allegation was that you delegated the decisions on employment, for better or for ill, to the individual units rather than trying to centralize them in top headquarters.

MR. ROSENBERG: And you're--

MR. EPSTEIN: And that is now a Title 7 violation.

MR. ROSENBERG: It may be a Title 7 violation not to have title--the correct anti-discrimination procedures in effect.

MR. EPSTEIN: Well, how do you know whether in an individual case it turned out that any of those defects affected any of the individual decisions?

MR. ROSENBERG: Let me say more broadly, because you're raising the same point that you made before about multiplicity of designs and individualization of products and so forth, so you get 17,000 different models of Firestone tires, each for its own road and temperature and so forth. Whether there are bases in the litigation for scale returns in economy, investment, and so forth, are beyond the capacity for any court to judge, Easterbrook included, and should be left to the investors to decide where they're going to--how they're going to proceed after investigation.

They may find that there's a common design defect in the way these tires, all of these tires were put together. They may find that there's no commonality. It wouldn't make any difference. The investor in a lawsuit is not going to waste resources investing on a common question that doesn't exist and that will be shown not to exist.

So, in a sense, you've turned over a thousand claims to an individual prosecutor, who may find scale benefits in prosecuting those claims. And if that person doesn't, then they'll either prosecute a thousand separate claims or turn them over--you know, sell them off to other people to prosecute, refer them away.

So the safest, the most efficient thing to do is to simply convene a class action. It adds no cost to the litigation because the investor will find out just what the most efficient means of prosecution is.

Let me add one more thing. There's another way to deal with Firestone. There's another way to deal with these cases of employment discrimination. And that's to simply say that we look at the entire enterprise and all its multi-factored operations and determine whether the aggregate risk was reasonable or unreasonable and then assess liability, to the extent that the substantive law allows it, vicariously against the firm.

That is an appropriate way to regulate, and it in a sense says even cases where you're assuming multiple models of the product over long periods of time, simply aggregate the risk and determine whether it's reasonable or not or apply a strict liability rule, whatever the substantive law decrees.

MR. EPSTEIN: But you've got to know what happened. I mean, suppose it turns out you have two tires, one of which breaks down in virtue of the fact that a fellow takes the tire and punctures it after he buys it, and another tire which has a latent defect which the manufacturer creates. If you're not going to look at the facts of the particular case, are you going to treat this as though there were two tire accidents and then penalize the defendant accordingly?

MR. ROSENBERG: No. You're going to make statistical estimates.

MR. EPSTEIN: But you can't make statistical--

MR. ROSENBERG: Why not?

MR. EPSTEIN: Because what happens is you're making statistical evidence when you actually have evidence which tells you that you don't need to use the statistics.

Look, it is perfectly sensible to use a class action in the case where you started all this stuff years ago in Sendell (ph), where you know that there was a manufacturer who supplied an identical fungible product to various people, you can't figure out which manufacturer supplied it, and you know that they all acted in exactly the same way with respect to the exact same class. That was the original set of factual assumptions--false in practice, by the way, but at least the ideal assumptions with respect to the DES cases.

But in these situations, you know that the products aren't fungible; you know that the use patterns aren't fungible. You know that the level of investment that was made with respect to some products may have been adequate and with other products may not have been adequate. What happens is that just everything that you're trying to aggregate is so utterly different that I don't know what aggregation means. There is certainly--let's put it this way: If you were suing for the damage associated with one Firestone tire, you would never, ever start to talk about the kinds of theories that you're doing in order to win that case. What you would do is you would look at this particular tire and you say this is the machine that made it, look at the tire, I can prove to you that there was a defect in the original composition of the rubber because there were all sorts of inclusions that ought not to have been there under sound practices.

Now, in effect, we don't do this anymore. At least I don't know what we do when we're trying these cases. Do you look at the individual tires to see if they're busted or do you not? And I guess your answer is you don't care. I can't understand how you run a system to figure out how much aggregate responsibility you want if you don't know whether they're making good tires or bad tires. And I don't know how you do that unless you look at the tires. And if you look at the tires, I don't know why you need a class action, because these are large enough damages, just bring them as ordinary tort suits.

I mean, I'm baffled by it all. I do not know what I'm supposed to do if I'm sitting there as a judge and you bring one of these mega class actions, what I'm supposed to admit by way of discovery and what I'm supposed to exclude. I couldn't begin to figure this out, David. I basically would resign my commission if I were forced to work--

[Laughter.]

MR. EPSTEIN: --under your set of rules. I mean, I understand something about economics. I mean, a little. Less after this conversation, I think, than before it, but I think I understand a bit.

MR. BUCKLEY: Let me ask a question of David. Let me offer a comparative perspective. If you compared our civil litigation system with that of the rest of the world, you would conclude that our country is vastly more pro-plaintiff than any other countries, and that across a variety of dimensions, from elected judges, which shouldn't matter but does, and punitive damages and substantive law and procedural law and even regulation, which is more burdensome here than I think in other countries. And so I have two questions.

The first question is really repeating Richard's question. If one thought that the substantive law was out of whack, wouldn't that tell you that you'd like to make a similar adjustment, a correcting adjustment with respect to the procedural law?

My second question is: You seem to have assumed that we can separate out substantive and procedural into watertight compartments. But isn't it the case, looking across the world, that they seem to be endogenous, that one laps up against another and that a more generous, shall we say, substantive law produces procedural changes which are more favorable to the plaintiff's bar? Isn't that what you'd conclude, what an econometrician would conclude?

MR. ROSENBERG: Let me say that the one thing I've tried not to do is separate procedure from substance. In fact, I think that's the virtue of this conference. Substance is for the first time, at least in a public conference, being brought in in a heavy way to the discussion about class actions.

It's quite clear to me that if you are interested in making people better off and the substantive law is bad and your only option is to manipulate the procedural law, one might manipulate the procedural law to adjust for the bad substantive law in order to come out so that the legal system does the best it can for people.

My starting point is not to do this in a sense evil-versus-offsetting-evil move. If the substantive law is bad, then I would aim to correct it. And if the procedure can be corrected so that it becomes a very efficient and simplified and, I would say, elegant platform for law enforcement, I would adopt that. And the reason I would adopt it is because we can hold that in as the baseline for all other adjustments to the substantive law to correct the substantive law.

What I'm trying to say is it's a last resort to capitulate your analysis to the notion that, gee, this doesn't work, why don't we make this work a lot less effectively to offset it. That's not the best engineering approach. Or let me put it this way: That's the engineering approach you take after you've exhausted reasonable opportunities to fix things. The class action is an effective law enforcement device. And it's simple, notwithstanding Richard's effort at complication. The substantive law may or may not be bad. We could go into discussions about successorship liability, which bothers people. But, in fact, that's just the contract system in many, many cases working its way through so that you don't have to do internal causal analysis.

We could talk about whether or not the substantive law is good or bad, but why don't we start with imagining a very efficient process and then turning our minds to specific substantive problems. I'd be glad to do it.

MR. EPSTEIN: I just have one comment. I have worked on a number of class actions, and they all make "Bleak House" seem as though they're models of expeditious judicial resolution.

In principle, you could have a class action in the sky that might be that way, but when you put it against a situation in which you're worried about jurisdictional issues and removal issues and joinder issues and class definition issues and sub-class issues and discovery issues and so forth, what happens is there's a point at which the scale economies flip over, and far from being advantageous, become disadvantageous. Anybody who has ever worked in the field on the ground instinctively recognizes this, and I would say as a general matter, if I'm for a class action, I'd be suspicious of any nationwide class actions when there are strong state substantive differences in law which might influence the outcome.

David's view on all conflicts of law or choice of law issues is average over them, just ignore them, right? But, in effect, you could have two class actions in State A and State B which work under different principles, which would be better than one class action which has both State A and B working on a set of principles that neither of them has followed. And then some judge has to figure out, if I can't use the state law of either state and I have to average it out, how do I take A plus B and divided it by two? I don't know what that means.

So this is not an efficient system. It is subject not only to economies of scale but to diseconomies of scale. And I think that in many of these cases we really have gone over by trying to bring too much into one, and it may well be that ten class actions are better than one class action, even if they come out in different ways, which they might because of either substantive differences or strategy choices by lawyers.

I mean, you give identical cases to two lawyers, and watch the way they go. It's like giving the same script to two movie directors and seeing what you come out with. I mean, it's just not going to be the same kind of thing.

MR. BUCKLEY: We have time for one last question. Yes?

JUDGE ROSENTHAL: My question is for Professor Rosenberg. My name is Lee Rosenthal, and I, too, look much more relaxed today than I did some years ago when I sat in Professor Epstein's classes.

[Laughter.]

JUDGE ROSENTHAL: But my question is for you, Professor Rosenberg, and it's this: One way, I think, to read the Supreme Court's decisions in Amchem and Ortiz is as a wholesale rejection of the model that you have proposed, that courts are indeed not administrative agencies and should not aspire to become administrative agencies, even if the scope of the harm dwarfs the ability of the single-case adjudication approach to handle.

So if you accept the premise that American law today will not admit of your solution, does your set of assumptions and insights have anything to teach on a smaller scale within the confines of existing law?

MR. ROSENBERG: Thank you, Judge. Yes, I think it does. The design options I was going to suggest--that I suggested at the end include the following and an accommodation to Amchem and Ortiz, and the reason I--and the accommodation is this: I read Amchem and Ortiz in part a rejection of my assumption that the class action can work efficiently when you have a massive case like asbestos.

But the other part of it, I think, had to do with what Richard was pushing, and that is the idea that people have an individual interest in their claims and the baseline for determining whether a class action serves that interest is what the person would have gotten if he or she had prosecuted the case outside of the class action. I call this the anti-redistribution principle. The class action shouldn't take from somebody what they otherwise would have gotten.

There is, I believe, a fairly effective rule design that could be used to accommodate that anti-redistribution principle while at the same time doing what I think most people really want, which is effective deterrence. And it's not pie in the sky in the sense that it was before the Judicial Conference as a proposal just a year ago, namely, don't allow a second opt-out. In fact, get rid of opt-out altogether.

Now, Richard raised instrumental reasons for keeping opt-out. It's a check on misbehavior by class counsel. Here I'm focusing on the anti-redistribution principle.

There is a design for class action that would be the following, and it's fairly straightforward. Deterrence is entirely a question of--entirely involves common questions if you adopt the mass production concept that I've--that we're dealing with mass production goods. You could focus it on a particular model of tire. You could focus it on Firestone in general and look at whether or not the aggregate risk from Firestone was unreasonable or not. But basically deterrence is a common question, and it's usually entirely statistical because you're asking what the firm did ex ante, and that can only be by statistical analysis of what the expected laws would be relevant to the expected investment--the appropriate investment and precautions ex ante.

You decide that. You impose an aggregate fine. I'm assuming the decisionmaker finds liability, finds liability in the aggregate, imposes an aggregate fine, puts that fine in a fund, and then allows all class members to intervene, through their own lawyers, if they want to, to sue the fund on non-common questions about how that fund should be distributed, so that in a sense you can have this mass feeding frenzy that we observe now of individuals exercising their autonomous interests to maximize their wealth at the expense of their fellow individuals who may need that money because they are out-of-pocket a great deal in terms of loss and they need the money for insurance.

There is a simple decoupling model for class actions. It's built into the rules now. You can have a class action for certain common questions and then relegate the case to non-common questions. It would solve the Amchem-Ortiz interest in preserving the distributional relationship between what party could get outside the class and what he or she could get through the class.

MR. BUCKLEY: We're out of time.

MR. : [inaudible] complicated.

[Laughter.]

MR. BUCKLEY: Last word from Richard.

Let's thank our presenters for their splendid talk.

[Applause.]

[Recess.]

MR. GREVE: Welcome back. We'll start our second panel.

Since none of the panelists has expressed a preference about the order in which they appear, we'll do this in reverse order of the order in which the panelists are listed in your program.

I should say before I introduce the panelists that, A, we're very, very grateful to have all of them appear at this event because they're all uniquely qualified to speak on the issue at hand.

I want to say one thing about--sometimes, and incomprehensibly, AEI is accused of a lack of balance. So here we have a panel about class action reform without a trial lawyer present. I suppose that's like having a conference on child abuse without an actual child abuser.

Erase that from the transcript.

Seriously, this is not a panel on class action reform yea or nay, in which case we would strive and achieve balance, I assure you. What we're here to discuss is the institutional dynamics of class action reform and the pros and cons of pushing that particular agenda in one form or another and the success or failure of the various institutions, the Judicial Conference, the Supreme Court itself, or the appellate system itself, the Congress, the Federal Trade Commission in this area.

To that end, we'll start with Judge Lee Rosenthal, who U.S. District Judge for the Southern District of Texas, in Houston. She's also been a member of the Federal Judicial Conference and particularly the Advisory Committee on Rule 23--on Civil Procedure, sorry, and the Subcommittee on Class Actions.

We'll then go to Mark Perry, who is a partner with the Washington office of Gibson, Dunn & Crutcher, a member of their litigation department, where he focuses on appellate litigation, constitutional law, and, needless to say, class actions and commercial disputes.

We'll then go to John Beisner, who's with the firm--who runs O'Melveny & Myers, very sizable class action practice here in Washington, D.C. He also has extensive experience with multidistrict litigation and with the congressional reform agenda, including the presently pending reform proposals. And we'll conclude with John Delacourt, who serves as the chief antitrust counsel in the Federal Trade Commission's Office of Policy Planning where he has helped--has played a central role in developing the Federal Trade Commission's Class Action Fairness Project.

With that, we'll start. Judge Rosenthal, please.

JUDGE ROSENTHAL: Thank you very much. It's a great pleasure to be here. I want to begin with the obligatory but sincere disclaimer. Obviously, my comments are in my own capacity and on my own behalf. I am not speaking on behalf of the Judicial Conference of the United States or of the Civil Rules Committee.

Having said that, as a member of the Civil Rules Committee I've been involved in its examination of Rule 23 now since 1996, and the committee has been actively involved in its re-examination since 1991. And the timing of this conference could not have been better because the last major reform of Rule 23, which really brings us all here today, had its birth in a meeting that was held on October 31, 1963. Halloween is when (b)(3) was born, and some might find that appropriate.

The other reason that timing is so exquisite is that in about one month from today, unless Congress acts affirmatively to prevent it, new amendments to Rule 23 will become effective. But the difference between those amendments and the (b)(3) and related 1966 amendments is what I want to talk to you about today.

To go back to 1966--and we have the transcript of the proceedings that gave birth to Rule (b)(3)--the proceedings and the faults that some find in present Rule (b)(3) did not occur because of any deficiency in the group that was present. The people who wrote this stuff were incredible. J.W. Moore and Charles Alan Wright were members of the committee. Dean Acheson was the Chair. The reporter was Professor Ben Kaplan of Harvard, later, of course, a Justice of the Massachusetts Supreme Court. These were the guys who wrote this stuff. They were pretty smart. And the issue of whether to have (b)(3) at all provoked intense debate.

They talked about mass torts, but the example used was the single-incident mass tort--the Ringling Brothers fire in the circus tent that had occurred in Hartford, Connecticut, not too long before. And Professor Moore and others were concerned that if you had a compulsory damages class, without an opt-out, as I look over towards Professor Rosenberg, you would have the ability to have what we now term the reverse auction. You would get a defendant that would rig a patsy class, arrange to have it sued, arrange to have it settled for a very low amount, and the defendant would be absolved of all liability at a very cheap price.

Judge Wyzanski had the inspiration that led to (b)(3). Would you be satisfied, he said, if the class could never include anybody who specifically protests within a given period? And the eminences around the table said, That would be helpful; if that was done, my problem would evaporate.

Voila, we have (b)(3). And we did, of course, have the wonderful committee note which says: A mass accident relating to injuries to numerous persons is ordinarily not appropriate for a class action. So much for the influence of advisory committee notes.

Members of that committee knew that what they were doing was bold. Members of that committee later described what they did as the most radical bit of rulemaking since the rules began, at which time you will all recall that one cause of action was created, and the differences between law and equity were eliminated. This was (b)(3) and the change in the class action structure, rule structure, was recognized as radical.

But Professor Kaplan also recognized that it would take about a generation to figure out just how radical it might be, at which point it would be appropriate to re-examine it. Well, that generation passed, and as everyone in this room knows, that generation saw an explosion in substantive law on all sorts of fronts, state and federal, that with the class action device available made enormous changes to the American legal landscape. No question.

In 1991, the committee is reconvened to examine class actions, specifically looking at the application of class actions to mass torts, which the committee had warned against but which the American legal world had embraced, with problems that were described in detail.

Five years later, five years of conferences, hearings, rule revisions, notice and comment, more hearings, more drafts--five years later eight proposed revisions to Rule 23 are issued, in 1996. All but one of them proposed essentially changes to certification standards: one addressed specifically to mass torts, one addressed to the negative value consumer class actions, and another addressed to settlement class actions. All of those proposed changes failed. They failed through a withering public comment process that really revealed profound differences of opinion about how mass torts ought to be handled, what the proper solution was, what the proper role of the judicial system was, and what the proper role of private litigation was in enforcing regulatory standards.

There was no consensus within the committee. There was no consensus outside of the committee.

The one rule reform that did become law was the interlocutory appeal provision, which has been enormously productive. It has produced--for the first time in the history of Rule 23 of class actions, it has begun to produce a body of appellate law that provides the backdrop for the debate we had this morning, a debate that you really couldn't have had in the same way before Rule 23(f) because you simply didn't have the appellate standards fleshed out to the degree we are now starting to see.

It has been enormously helpful, but talk about a modest and discrete reform. Rule 23(f) has to be a poster child of that kind of approach.

What that whole process in part taught us is that rulemaking itself has changed dramatically since 1966, at least in areas that involve as much money and as much power as class actions do. Radical rulemaking of the sort that occurred in a small room, Halloween, 1966, could not occur today because today rulemaking in a sunshine committee, such as the Rules Committees are, is, to use a much overused word, transparent, it is incredibly public, and it is so deliberately deliberate, strung-out, prolonged, inclusive, that it discourages changes of magnitude.

And in the case of Rule 23, as was observed by both of the panelists this morning, the difficulty of making rule changes anywhere but on the margins is exacerbated by the very great difficulty of distinguishing between substance and procedure. Because even though Rule 23(b)(3) was created within the Rules Enabling Act, it was the by-product of a procedural change. It is now widely viewed as having become so entrenched in access to and enforcement of substantive rights that any attempt to cut back on those rights or discipline those rights or impose greater procedural rigors on those rights is viewed as outside the Rules Enabling Act process. It would either expand or limit substantive rights, in effect, the argument goes, so the very Rules Committee that created this mess can't do anything meaningful to clean it up.

Is that indeed where we are? Well, the Rules Committee decided that it would continue to examine the opportunities for productive reform within the Rules Enabling Act, and it went back in 2000 to try another stab at rulemaking. And it did so on two tracks. Because of these limits and because we do learn, we have been taught a valuable set of lessons in the art of the feasible.

What we heard when we went back is that there was now--that things had changed. Things had changed very much in the last eight years since we had started this examination. Certification standards in the federal courts had become much more or perceived as much more rigorously enforced. The problem had migrated to a certain extent in that there was a huge explosion of class actions in the state courts and because of certain important decisions that in at least one case had nothing to do with class actions, had taken the form increasingly of nationwide and multistate class actions certified in the court of a single state.

At the same time, and closely related to this, you had this explosion in overlapping, duplicative, and competing class actions. These two phenomena were new. They had not been at the forefront of the complaints that had driven the proposals in the first round of rulemaking in 1996.

And at the same time, the committee was told that there were grave problems in the process by which class actions were litigated that were susceptible to improvement by judicial scrutiny, if there were better tools spelled out for litigants and judges to understand and to use.

So we did both. We did sort of a parallel rulemaking track. On the one hand, within the Rules Enabling Act, we issued proposed rule changes within four areas, targeted not at certification decisions but at the process of litigating class actions: the timing of the class action decision--the certification decision, the requirements for notice, the process and standards for court approval of class action settlements, including strengthening opt-out provisions--sorry, Professor Rosenberg--and rules for the first time explicitly addressing the standards and process for attorney appointment and attorney compensation for class actions.

These amendments got through. There was vigorous debate, but they did succeed in getting through the Standing Committee, through the Judicial Conference, through the Supreme Court, and on December 1st they will become law unless between now and then Congress decides that there's some great problem and acts affirmatively to disallow them.

And these are good amendments. They're fair, they're balanced, but they are modest. They did not attempt, except in marginal ways, which may nonetheless be very useful but are indirect, they did not attempt to address the problems of overlapping and duplicative class actions pending simultaneously in state and federal courts. We tried in a parallel rulemaking track, through a reporter's call for comment, we put to the drafting test the ability within the rules to address those problems. And our reporter, Professor Cooper of the University of Michigan Law School, issued a call for comment on three proposed rules changes that would have used preclusion and a kind of a preemption to address repeated shopping of certification decisions, that is, if it's denied once, attempts to find another court to reach a different result on essentially the same class; shopping of settlements, attempts after a settlement has been rejected to shop it somewhere else that might be a more congenial place; and attempts to give federal courts tools that would prevent cases from proceeding in other forums, specifically state courts but also other federal courts, that is, once a federal court is handling a class action and is thoroughly committed to it and involved in it, that would have given that court the tools to prevent competing and overlapping class actions seeking essentially the same relief.

We got nowhere. We tried very hard, and there was a wonderful debate over the limits of the Rules Enabling Act, the Anti-Injunction Act, the appropriate role of rulemaking, all of which led the committee itself to conclude that the rules enabling process would not permit us to get the result that many agreed was a good result through the Rules Enabling Act process.

Instead, what the advisory committee did was to adopt a resolution that large nationwide and multistate class actions involving class members from multiple states who had been injured in multiple states is the kind of national litigation consistent with the purposes of diversity jurisdiction and appropriate to jurisdiction in federal court.

Now, I cannot tell you what a huge step that was, modest as it may seem, for the Advisory Committee and the Standing Committee of the Judicial Conference to adopt, because, as you may know, some years previously the Judicial Conference had taken the position that attempts to widen diversity jurisdiction were disfavored. This was a significant step. And on March 18th, the Judicial Conference itself did adopt the following resolution: that the Judicial Conference recognized that the use of minimal diversity of citizenship may be appropriate to the maintenance of significant multistate class action litigation in the federal courts while continuing to oppose class action legislation that contains jurisdictional provisions that are similar to those in the bills introduced in the 106th and 107th Congresses. And the resolution then goes on to list the kinds of factors that Congress was encouraged to consider in deciding which class actions would be appropriate for federal court and which would appropriately be left within the state courts.

And if that language, carefully nuanced as you no doubt detected it to be, smells of compromise, you would have an accurate sense of smell. But it was an enormous step for the Conference to take, and many of us who followed the debates last week were somewhat amused to see how this carefully crafted position was somewhat garbled in the rendition on the floor of the Senate.

From the rulemaking perspective, where we are now, we are continuing to look at whether Rule 23 can meaningfully and feasibly be amended to ease many of the problems that are still before us, an examination that will no doubt be affected, both in the sense of urgency and in the specific issues, by whatever happens in Congress and when it might happen in Congress. But in the meantime, we have, I think, come to the realization that it is very much the responsibility of individual courts and the courts of appeals reviews the records that these individual courts create to avoid the excesses and abuses that class action litigation can produce without running afoul of the benefits and the incentives that the rules provide. Those policy choices have been made, and our job as courts is to implement them in a way that is fair and just, to use that word, and that is consistent with all of the constraints under which we as courts operate. But it is very much up to the case law to carry the burdens that Professor Rosenberg and Professor Epstein so dramatically talked about this morning. The rules themselves have done much of what they can do.

MR. GREVE: Thank you, Judge.

Mark Perry.

MR. PERRY: Thank you. Lee provided me a transition into my little remarks, and I thank her for that. The question that I was asked to address is: What can the federal courts, and in particular the Supreme Court, do about class actions within the existing body of law? We practicing lawyers don't have David Rosenberg's luxury of throwing everything out and starting from first principles. We have to proceed with the laws that Congress enacted, the rules that the committee and others come up with and the decisions that are out there.

In that context, and recognizing that John Beisner is going to tell us whether the context will change, I want to offer four proposals in what I think are an increasing order of controversy of things that could be presented fruitfully to the Supreme Court by the litigants in a proper case.

The first is that the Supreme Court should say in this context that the Rules Enabling Act means what it says. Richard Epstein referred to this, and stated at that level of abstraction, you're all saying, well, of course, it does, that's a dumb suggestion. But if you look at the class action cases, 95 percent of our courts don't understand that proposition. They mouth the words of Rule 23, typicality and commonality and so forth, without asking the obvious question, which is: Do the members of this class have a right that they can be--that they can enforce through this device under the substantive law on which they are suing?

And the answer to that is usually, often, frequently, that they do not, whether it is because of the lack of injury, whether it is because of the failure of each class member to meet a substantive element of the cause of action--in fraud cases, for example, which are a huge number of class actions, reliance is an element. Reliance has been an element for, Richard Epstein could tell us, I don't know, 500 years or something. Yet in class cases, reliance goes out the window.

I would submit that the Rules Enabling Act has something to say about that; that it is not enough to say that the class members have a common issue in that they've all been allegedly lied to, but that they have to prove that they have a commonality in their reliance. And in most cases, they won't be able to make that. And we could go through every tort or every securities fraud violation or securities law. Each element of the underlying offense contains issues that are just overlooked in a shockingly large number of cases at the trial level and at the appellate level that deal with these issues. The courts twist and turn regarding the Rule 23 requirements without asking that underlying question whether the class itself, whether the members of the class have a claim.

This alone would preclude the maintenance of many class actions. If you look at the mine run of cases and you actually required that each class member step up and prove or at least allege these substantive elements of the offense, we would have changed a significant thing.

That is all well and good in federal court, and you may say the real problem lies in the state courts. What could we do there? What can the Supreme Court as a matter of federal law do there?

One would be if the court were to announce the principle that I propose, which is simply that the substantive law can't be overlooked. That might have some effect on some state courts that actually pay attention. There is an encouraging trend among about half of the states to actually read Supreme Court opinions on occasion.

The other thing the Supreme Court, though, could do, should do, and I think soon will do, frankly, with respect to state court actions is to reaffirm in no uncertain terms that the Schutz (ph) decision means what it says. Schutz is one of the most important decisions ever issued by the Supreme Court in this area, and yet outside of the Seventh Circuit and Judge Rosenthal's court, it appears not to exist. It does not appear in the books. It does not appear in the briefs. It is totally overlooked. In California, where I do a lot of my practice, it has been rendered an irrelevancy.

The courts there reason--and this reasoning, if you can call it that, has been echoed throughout the country--that because California has a consumer protection statute and all states have consumer protection statutes, California may maintain actions on behalf of all consumers in all states. This overlooks the fact that in--well, actually, Louisiana doesn't have one, but in the other 49 states, 48 of those do not allow private rights of action under their consumer statutes, so California is co-opting the decisions made by the people of all those other states as to who can sue and for what.

Schutz, as written, precludes this, but courts don't seem to gather that. And for that reason, my second suggestion is that the court needs to reaffirm Schutz.

I would actually suggest that the court did reaffirm Schutz in an important way in the BMW decision, which most people read as the punitive damages case, and it is very important and valuable for that. And I'll set that aside because you can agree or disagree with the due process ability of federal courts to craft state tort law in that regard.

But there is another element of BMW which I think is noncontroversial--at least it is supported by 100 years of precedent--which is that the states may not export their rules to other states. And yet that is precisely what California and Madison County, Illinois, and certain parts of Alabama in days gone by--it has gotten a little better--and Texas in days gone by--it has gotten better. But there will be--John Beisner tells me St. Clair, Illinois, is going to be the next hot spot. Plaintiffs' lawyers out there take note--violate that principle of BMW, violate that principle of Schutz, and they do it on a daily basis.

That leads into the broader question whether there are other due process limitations on the very maintenance of state class actions. I think there are. I think they need to be explored. I think the federal courts have not done so. I don't intend, because I have not thought it through, to flesh those out today. But if we look at punitive damages and the reforms that have been wrought at the federal level, for good or for ill--we don't need to pass judgment on that--some very creat