In a recent essay in The Yale Law Journal, Douglas Lichtman argues that courts considering preliminary injunctions should account for irreparable benefits in addition to irreparable harms. This is a provocative idea. If a preliminary injunction harms one party but benefits the other, and if both effects are equally difficult to subsequently undo, why focus on one effect (harm) and ignore the other (benefit)? There is a compelling geometric validity to this symmetry observation. But is this a valuable “flipping” exercise? Does it shed a new light and provide useful insight into the law of injunctions?

In this Response I want to suggest that the case for irreparable benefits is valid, but perhaps overstated. I have no quibble with the claim that if there are irreparable benefits, and if they are indeed overlooked, the legal doctrine ought to be corrected. But I am less sure that the existence of irreparable benefits is prevalent, and I am even less sure that they are systematically overlooked. Let me explain.

I. Are There Irreparable Benefits?

Consider this scenario: the plaintiff has a patent, the defendant allegedly infringes it, and the plaintiff seeks a preliminary injunction. The plaintiff’s harm if the court wrongfully denies the injunction is the profit he might lose when the defendant competes with him. That harm is speculative because its measurement requires a counterfactual—a court in the future will have to figure out how much more the plaintiff would have sold but for the competition. On the other hand, the defendant’s benefit if the injunction is temporarily denied is the profit he will in fact enjoy by competing. The profit is not based on any counterfactual. It is ascertained on the basis of the actual sales that the defendant enjoyed in the interim. This is a standard restitution, or disgorgement-of-benefit remedy. Lichtman too recognizes that “[s]ales forsaken, for instance, are hard to measure because it is difficult to run the necessary counterfactual. Sales unlawfully made, by contrast, can often be easily disgorged, namely by taking away the undeserved cash.” In such cases, therefore, the defendant’s benefits from an erroneous denial of preliminary injunctive relief are not irreparable.

Sometimes, it is true, courts cannot measure benefits so easily. Lichtmanuses the example of the benefit enjoyed by an infringer who saves the cost of inventing around a patent. That defendant’s benefit cannot be measured by “the undeserved cash” he earns from infringement because some of that cash is legitimate profit. Instead, the benefit is equal to the increment of the infringer’s profit that he could not have earned by other, non-infringing means. In this example, a purely benefit-based recovery would be difficult to assess, and the benefits would therefore be irreparable.

This is a valid analytical point, but does it really come into play in the context of preliminary relief? At the preliminary relief stage, the benefit that the defendant enjoys is the continued activity during the period of trial, whether measured by cash from sales or by freedom to continue to develop a product. And those interim-stage activities are distinct from the costs of invention inputs, choice of products, and the like, which accrued prior to the dispute. The defendant will squander these invention and product-development costs anyway, once a permanent injunction is issued. Allowing the defendant to continue with the activity in the interim stage may create some additional short term benefits for the defendant, but it is hard for me to imagine that this incremental gain would have a fundamental ex ante effect on the defendant’s innovation incentives.

My first point, then, is that the asymmetry in existing law between harms and benefits may be a product of a systematic difference in measurement difficulty. Harms are more important because they are generally harder to measure after an erroneous decision regarding preliminary relief.

II. Are Benefits Overlooked?

Even assuming that courts should consider irreparable benefits, I believe they already do consider them, despite the seemingly asymmetric, harm-focused doctrine. This is an objection that Lichtman anticipates but does not find convincing. So let me try to articulate it more persuasively.

If an injunction is issued, the defendant is harmed. By how much? By however much he is worse off relative to the situation in which the injunction is denied. That is, the defendant is injured by the amount of benefit that he would otherwise enjoy, if he were allowed to continue with the activity enjoined. The harm, then, is the forgone benefit. Why count it twice?

The same is true for the plaintiff: if an injunction is denied, the plaintiff’s injury is the benefit he would have enjoyed had an injunction been granted—a benefit that is now forgone. For each side, then, the harm is defined as the difference between two states: injunction and no injunction. Since all that is measured is this difference, it need only be measured once. As a matter of definition, benefits are already accounted for.

To illustrate, suppose a plaintiff-patentee will earn $100 from sales if an injunction is granted (maintaining his monopoly) but only $40 if an injunction is denied and the defendant is allowed to compete. Suppose also that if allowed to compete, the defendant will earn $25 (but $0 otherwise, if an injunction is issued). Thus, an injunction would benefit plaintiff by $60 (raising his profit from $40 to $100) and would harm defendant by $25. A denial of the injunction would benefit the defendant by $25 and harm the plaintiff by $60. Under the traditional practice of comparing harms, we need to compare the plaintiff’s harm of $60 with the defendant’s harm of $25 (and check to what extent these harms are irreparable). I am worried that the proposal for counting benefits amounts to making this exact comparison twice.

Lichtman, as I said, recognizes this double-counting objection, but he argues that it is answered by a more complete understanding of the effects of errors. A denial of injunction, if erroneous, may cause irreparable harm to the plaintiff-patentee and afford an irreparable benefit to the defendant-infringer. The most important irreparable harm is the reduced gain from the plaintiff’s innovation, measured by the added profits that the plaintiff would have made in the interim phase. This incremental diminution in payoff affects the plaintiff’s incentive to innovate. The most important irreparable benefits of an erroneous denial are the profit the defendant now enjoys or the freedom he has to continue his infringing research. This payoff also has an effect on the defendant’s choice of research projects because it reduces his need to invent around a patent.

These are important observations that sometimes elude courts, namely, that the distributive effect of an injunction-or-denial decision, perhaps even in the interim stage, might have ex ante effects, such as on the incentives to innovate and to choose projects. The ex ante effects of post-injunction (or post-denial) reallocations surely ought to factor in to the decision whether to grant or deny an injunction. But the way to factor them in is not by counting benefits. Rather, it is by conceptually understanding harm to embody these more subtle precautionary actions that parties might take to avoid an even greater harm. In the example above, if the harm of an injunction to the defendant is the denial of a benefit of $25, then the defendant might, in anticipation of that harm, invent around and develop a non-infringing product with a benefit of $15. The harm, then, is not the full $25, but only the unavoidable $10 (the difference between $25 and $15).

The point is that recognizing the existence of incentive effects does not mean that benefits ought to be counted twice. Once we identify how a party’s payoff changes when an injunction is granted, we are set to evaluate whether this payoff is avoidable through self-help or other incentives, and whether the unavoidable part of the harm is irreparable. But all we need to know is this change in payoff, and the “irreparable injury” methodology provides the necessary traction to make that calculation.

Conclusion

Despite these critiques, I like Lichtman’s symmetry observation because it focuses attention on a factor that might otherwise be overlooked. I argued that courts don’t need the language of irreparable benefits to manage the benefit side, but are they careful to do so? Do they truly measure harms relative to the forgone-benefit benchmark? The important lesson I take from Lichtman’s work is, therefore, “internal” to the irreparable injury rule. What makes an injury irreparable is not just the difficulty in measuring the loss; it is also the difficulty in identifying the benchmark from which the loss ought to be assessed.

Omri Ben-Shahar is Kirkland & Ellis Professor of Law and Economics at the University of Michigan Law School.

Preferred Citation: Omri Ben-Shahar, Against Irreparable Benefits, 116 Yale L.J. Pocket Part 381 (2007), http://yalelawjournal.org/forum/against-irreparable-benefits.