Monday, April 4, 2011
Adam Liptak has an article on the NYTimes week in review entitled "When a Lawsuit is Too Big." In the piece he presents all kinds of arguments, including the argument that class certification can give the plaintiffs too much leverage in settlement negotiations. His example for this blackmail problem is an antitrust suit with millions of class members that was certified by then-judge Sotomayor and settled for $3 billion. But he misses the punchline: the named plaintiff in that suit was Wal-Mart.
A $3 billion dollar settlement is really big; it seems kind of scary and aweful and a good reason to dislike class actions. But is it too big? I don't know. It depends on the value of the underlying claims. If someone does $3 billion worth of damage, they should pay. Remember that time you broke your neighbor's window with a baseball? Same concept here, just a larger scale.
Liptak missed some other points. First, what's really the effect of these cases on defendants? Liptak should have interviewed Michael Selmi, who wrote an article back in 2003, discussing none other than the Wal-Mart v. Dukes case, called "The Price of Discrimination." In that article, Selmi shows that antidiscrimination suits rarely affect stock prices and expressed concern that damages in such suits are insufficient to really affect behavior.
Second, are these suits really unfair to the women in question? That is an important issue that needs to be explored, one that was raised by Justice Roberts in oral argument in the Wal-Mart v. Dukes, as well as by Justices Ginsburg and Sotomayor. The concern is about the ability of the courts to determine back pay on a statistical basis. I address this problem in a paper called "The Curse of Bigness." The bottom line is that a statistical model for determining back pay can be pretty accurate and is more likely to yield horizontal equity (that is, give similarly situated people the same amount) than individual suits. Furthermore, given the passage of time and the number of Wal-Mart employees over whom a given manager has oversight, the likelihood that a manager will have a specific and reliable piece of information about a given employee that is both relevant to the calculation and that is not captured in other data that Wal-Mart does keep electronically (such as tardiness, for example), is highly unlikely. If there is such a case, or many such cases, Wal-Mart can bring them to the district court's attention.
So when is a class action lawsuit too big? This is just the wrong question to ask. I know that people would really prefer to ask it, but the question reflects a misunderstanding of the law. The right question to ask is this: When are class members too heterogenious?
Consider the following simple hypothetical. Big Bank has twelve million customers. It erroenously deducted $25 from the account of each individual account. The bank is clearly in the wrong, but it refuses to return the money. Just filing a small claim action costs at least $25 in most states, so it is not economically viable for each person to file a lawsuit. Filing in federal court costs much more. In this case, all twelve million customers are identical. Wait a minute, you say, doesn't it matter that there are twelve million people in this class? No. The important question is not how many class members, but whether they are sufficiently alike. Here they are identical, so it is a good case to bring as a class action. Homogeneity, not size, is the key inquiry. It doesn't matter that there are twelve million people, just that they are all the same in the relevant way.
Now imagine that the hypothetical class of twleve million is not all the same but instead breaks down into three categories. Category A were charged $25 and suffered no other adverse consequences. Category B were charged $25 which caused them to overdraft on their account, triggering an overdraft fee. Category C were charged $25 and overdrafted on their account, but the amount by which they exceeded their balance exceeded $25 - they would have overdrafted anyway. Are they still sufficiently homogeneous? Yes. It is easily and objectively ascertainable from bank records who falls into Category A, B & C. Each of these can be placed in separate sub-classes or merely treated differently in the litigation (Category A & C gets $25, but Category B gets $25 plus their overdraft fee). They are sufficiently homogenous and the class is manageable, but the class members are not identical.
Wal-Mart v. Dukes presents a harder case than these hypotheticals. That's one reason it is before the Supreme Court whereas a run of the mill consumer class action is not. But its the same question. Ask not whether the Dukes class action is too big, but whether the Dukes class members are sufficiently alike.