The first potential problem is that this places an even greater premium on "first to file" than the current system. I have seen too many ill-thought-out class complaints in my time, particularly in non-mass-tort contexts, that were slapped together to get to the courthouse first. Perhaps this can be controlled by having the court set the proportion of compensation to the "original" counsel based on "how close" that initial complaint is to the eventual winning theory of the case. <SARCASM> We really, really need another opportunity for satellite litigation over fees! </SARCASM>
The corollary is where things begin to get really messy. Why should this concept be limited to single class actions, aside from administrative convenience? For example, at this time Ellison v. Robertson remains a live matter. (The statute of limitations has expired for a new action on the operative facts, so this really is a purely hypothetical example.) Shouldn't I get compensation if BigClassActionLawFirm files a "piggyback" class action relying upon the four years of work I've put into this case? Then there's the can of worms opened by "competing" class actions. In my former life as a class action lawyer on consumer finance issues we had to deal with intricate issues of interlocking, overlapping, and contradictory class actions filed both across the country and within the same district. (In Gibson, the Court of Appeals even expressed concern that the actions had not been consolidated by the trial court.)
What this perhaps points out more than anything else is that the means used to pay lawyers in the US seriously distorts litigation. But that is for another time. I don't like to be the critic who bitches that something is broken without offering a workable alternative that does not have similar, or even more significant, flaws.