Proving Harm That Money Cannot Repair

Irreparable harm is a comparison, not a magnitude. The question is whether an award of damages entered after trial would put the moving party back where it would have been, and if it would, the harm is reparable no matter how large the number. That framing surprises litigants who arrive with real and quantified losses, and it explains why the factor defeats more applications than the merits do.

What makes an injury irreparable

Three features recur in injuries courts treat as beyond repair. The loss is difficult to measure, so a jury could not calculate it without speculating. The loss is compounding, so each day adds injury the eventual award will not capture. Or the loss destroys something that cannot be reconstituted, such as a confidential document once published or a business once wound up. Any of the three can carry the factor; none of them is satisfied by asserting that the loss is severe.

The Supreme Court added a second requirement that is easy to overlook. The harm must be likely. Several circuits had allowed a strong showing on the merits to reduce the harm requirement to a possibility of injury, and the Court rejected that as too lenient. A movant must therefore establish both that the injury would be beyond repair and that it is probable, which means the declarations have to describe a mechanism rather than a fear.

The categories, and their limits

Goodwill is the most frequently invoked and the most frequently rejected. Courts accept that reputational damage can be irreparable, but they want to see how the damage occurs. A declaration naming customers who have raised the issue, or describing a distribution channel about to be lost, does work that a general statement about brand harm does not. The same is true of employee departures, which count when they take institutional knowledge and not when they are ordinary turnover.

Confidentiality cases are stronger because the mechanism is obvious. Once information is disclosed the secrecy is gone, and no damages award recreates it. The contested question is usually whether the information was confidential at all, which pushes the analysis back onto the merits. Cases about unique property behave similarly: the difficulty is proving uniqueness rather than proving that a unique thing cannot be replaced.

The economic-loss cases are where the doctrine does most of its filtering. A plaintiff losing revenue to a competitor is suffering exactly the injury damages exist to remedy, so the argument has to be that something makes this loss different. Insolvency of the defendant, imminent collapse of the plaintiff, or a statutory bar on recovering money from the particular defendant are the recognized routes. Each is provable with documents, and each fails when supported only by an assertion that the sums involved are large.

Delay, and what a court reads into it

Nothing damages an application more reliably than an unexplained gap between discovery of the conduct and the filing. The inference is simple: a party that lived with the injury for months has demonstrated that the injury is survivable. Courts apply the inference to the emergency posture in particular, because a movant asking for relief within hours has to explain why the same urgency did not exist earlier.

Delay is rebuttable. Time spent investigating, negotiating, exhausting a contractual dispute procedure or waiting for the harm to become concrete is generally accepted where it is documented. What fails is silence in the affidavits about the interval, which leaves the court to supply its own explanation. The chronology also bears on the certification supporting relief without notice, discussed in the record of efforts made to give notice.

Presumptions, abandoned and restored

Courts once presumed irreparable harm from a finding of likely success in intellectual property cases. That practice largely ended when the Supreme Court insisted that equitable relief in patent cases follow the ordinary principles of equity, and lower courts extended the reasoning to copyright and trademark. The result was that plaintiffs in those fields had to prove harm like everyone else.

Congress then intervened for one of them. The Lanham Act now provides a rebuttable presumption of irreparable harm on a finding of likely success on the merits for a preliminary injunction in trademark cases. The presumption is rebuttable, so the defendant can meet it with evidence, but the burden shifts. Patent law contains no equivalent provision, which is why the two fields now diverge on a question they once shared.

Statutory schemes elsewhere achieve something similar without using the language of presumption. Where a statute authorizes an agency or a private party to seek an injunction on proof of a violation, courts often read the provision as reflecting a legislative judgment that the violation itself is the harm. The reading is not automatic, and it depends on how the enforcement provision is drafted, but it means the harm inquiry in a regulatory case can look quite different from the same inquiry between two private parties.

Building the record in the time available

Because the factor turns on prediction, the evidence is almost always declarations plus documents. Useful declarations come from people with operational knowledge: the person who manages the customer relationship, the engineer who knows what the disputed file contains, the finance officer who can show the cash position. Expert declarations help where measurement is genuinely contested but rarely substitute for a witness who can describe what is happening.

The opposing side's most effective response is often not denial but reframing, showing that the injury is measurable and therefore payable. That is why the harm showing and the equities showing tend to be argued together, and why the material supporting the weighing of harm to each side should be assembled at the same time. The same declarations also feed the court's view of what security is proper, a question treated in the rule on the security a moving party must post.