The real question is what one has as a baseline for efficiency. If the "efficient" result is a dispute-resolution system that both (a) properly internalizes all externalities and (b) costs less—in total, in transaction costs, and in translation costs—than any other reasonable alternative, it's pretty clear that the "objector" system is normal to the curve, for a very simple reason. The reality is that successful class actions are a "last resort" after all less-costly methods have been exhausted; so this is merely fine-tuning an "unacceptably" expensive method of regulation. Then, too, there's the "reporting problem." Outrage at class actions is ordinarily outrage at outlier cases not representative of class actions as a whole—and usually not representative of "regulatory" actions at all, but of post hoc attempts to obtain compensation for harms from events of dubious repeatability. Consider, instead, alternatives like this one. Without disclosing any confidential details, the combined attorney's fees on both sides were roughly three orders of magniture lower than those claimed in the MasterCard/Visa litigation. Gibson, however, has been quietly effective in actually changing some behavior. A well-informed cynic would say that used car dealers just hide their "finance managers'" shenanigans more effectively and in different places now; but the same can be said for any antitrust case. So, on this basis, which result was more "efficient"—and was there a doctrinal distinction between the two?
I argue instead that the actual baseline is the ever-elusive "costless voluntary compliance," not competition among statistically invalid subsets of instances attempting to reach any kind of compliance at all. To put it another way, measure the "invalid" transactional cost result—and that is exactly what class counsel's fees are—against the systemic objective, not against the immediate financial interests of the plaintiff class and defendant(s). The real test, then, is not the result of the particular case, but the systemic balance achieved some time later considering all of the actors. Again, part of the difficulty here is the "large number problem" I alluded to last week. When one looks at a $220 million attorney's fee request, one needs to think for a moment about the context. What is the proper context? Remember that this is a one-time, and not continuing, fee. Given that the profits of the MasterCard and Visa networks between the date the lawsuit was filed and the filing of the fee petition exceeded $600 billion, from a systemic perspective it's almost unmeasurable.
It's one thing to complain about large costs in terms of how those costs relate to Joe Lunchbox's annual salary. It's another entirely to look at them in the context of the subsystem to which those costs relate. Sometimes that glass will be half-full; sometimes half-empty; but the size of the glass seems to be a much more important consideration. Notice that I've carefully avoided questioning the purity of the water, since then we'd be talking about mass torts, which would obscure the point rather successfully. Which, of course, is the real premise of too many advocates of "tort reform": whether or not the cost is properly externalized, since it's in their immediate best interest to do so, the entire system should do so. What? Me cynical about "enlightened self-interest"? Never!