Most legal scholars agree that securities fraud class actions do little to compensate investors. Most investors are well diversified and thus are just as likely to sell an overpriced stock as to buy one. Moreover, since the defendant company ultimately pays in a successful class action, holders effectively pay buyers. Although this circularity is widely recognized, few have noted that because of the anticipated payout, the prospect of a class action causes stock price to decline by more than it otherwise would, thus generating additional (feedback) loss for both buyers and holders. In this article, I describe a method by which one can measure the net effect of class actions on fund investors who are both buyers and holders of a fraud-affected stock. Since an index fund almost always holds more shares than it buys during the fraud period, an index fund almost always loses more than it gains. Thus, class actions systematically penalize rational index fund investors for the benefit of irrational undiversified stock-picking investors. Accordingly, index funds should oppose class actions as contrary to the best interest of investors. To be sure, one possible problem is that in the absence of the deterrent effect of class actions, there might be more securities fraud. The answer is that whenever there is a meritorious class claim, the corporation itself will also have a claim – against the individual wrongdoers – for any increase in cost of capital resulting from reputational harm and any direct expenses relating to enforcement proceedings. In a class action, these elements of loss are imbedded in the price decrease that occurs when the fraud is discovered. But these losses are in fact suffered by the corporation and should be the subject of a derivative action for the benefit of the corporation – and thus all of the stockholders – not a class action for the benefit only of those who bought during the fraud period. Although the corporation claim may be smaller than the class claim in the aggregate, it is likely to be quite substantial from the point of view of individual wrongdoers and thus to constitute a significant deterrent to fraud. Happily, the rules of civil procedure provide a clear fix for the problem. First, the law is clear that a claim that can be handled as a derivative claim must be handled as a derivative claim and that a derivative claim must be resolved first before any class claim may be addressed. Second, no class action may proceed unless the court certifies it as a proper class action. And no action may be so certified if there is any other equally good way to litigate the issues (as by means of a derivative action). But someone make the argument. It is puzzling that no one has done so, especially because derivative actions eliminate feedback losses and serve to restore stock price. There are several possible explanations. One is that insurance does not cover derivative claims as it does class claims. Another is that attorney fees are likely to be higher in class actions than in derivative actions. These factors may incline plaintiff lawyers to favor class actions even though investors would be better served by derivative actions. On the other hand, until now no one has quantified the costs and benefits of class actions for real world investors. As shown here, index funds almost always lose more than they gain and thus should oppose class actions in favor of derivative actions. Indeed, index funds owe a duty to their investors to do so.