Sovereign Immunity, Equity, and the USAID Temporary Restraining Order

Nicholas Bagley & Samuel L. Bray

Earlier today, in Department of State v. AIDS Vaccine Advocacy Coalition, the U.S. Supreme Court denied an application to vacate or stay a district court’s order to disburse about $2 billion owed by the federal government for work already completed by various contracting parties. That order was an enforcement of a temporary restraining order previously issued by the district court. Justice Alito, joined by Justices Thomas, Gorsuch, and Kavanaugh, dissented from the Court’s denial of the application to vacate or stay. Here are a few observations and a few questions:

Observations

1. The Chief Justice’s administrative stay and the Court’s denial of the application had the salutary effect of avoiding the Court being forced to decide—or to tip its hand about a decision regarding—some major legal questions. It would not be good, for example, for the Court to determine the interplay between sovereign immunity, equity, and the disbursement of federal funds on an application for a stay of an order enforcing a temporary restraining order. That emergency posture is not conducive to deliberate decisionmaking.

2. The dissenting justices were correct to say that the TRO granted below was overbroad, since it should have run only to the protection of the plaintiffs and those represented by the plaintiffs.

3. The district court was correct to emphasize the need for learning and updating as the case continued from the TRO stage to the preliminary injunction stage, both with respect to factual development and to the arguments made by the government. SeeAIDS Vaccine Advocacy Coalition v. U.S. Department of State, 2025 WL 485324, *5, 6 & nn.3, 4 (D.D.C. Feb. 13, 2025).

4. One issue at the heart of these various requests for emergency orders is what their purpose is. Is it primarily to preserve the efficacy of the court’s ultimate remedial options, or is to accelerate the decision of the case? That matters for how central the merits should be in the analysis at each stage of the case. (For more on this question, see The Purpose of the Preliminary Injunction and Preliminary Injunctions on a Blank Slate.)

5. One way to think about the preceding observation is to ask how the assimilation should work between the preliminary injunction, the TRO, and the administrative stay. Should the preliminary injunction move toward the TRO and the administrative stay (more focused on status quo preservation, less on merits previewing or acceleration)? Or should the TRO and administrative stay become more like what the preliminary injunction has increasingly become (i.e., heavily focused on the merits)?

6. The jurisdictional fight at the heart of the case—is this a routine APA suit or is it a claim for “money damages” under the Tucker Act?—will likely prove quite consequential. If it’s an APA suit, an order setting aside the funding freeze as to the parties may well be appropriate, perhaps backed up by an injunction if the Trump administration is recalcitrant. The courts have made a practice of entering preliminary injunctive relief in anticipation of such an outcome (though we doubt that practice is sound). If it’s a Tucker Act suit, in contrast, the relief will be money damages down the line, and immediate injunctive relief is probably off the table. The eventual resolution of the jurisdictional question may, indirectly, supply a gauge of the Supreme Court’s willingness to police President Trump’s assertion of authority to impound appropriated funds.

Questions

1. The dissenting justices note that the APA’s waiver of sovereign immunity does not extend to “money damages.” They then analogize the order requiring payments to monetary damages: “the relief here more closely resembles a compensatory money judgment rather than an order for specific relief that might have been available in equity.” Department of State v. AIDS Vaccine Advocacy Coalition, No. 24A831, slip opinion at 5 (March 5, 2025) (Alito, J., dissenting from the denial of the application to vacate order). But that is not inevitable. The relief could just as easily be characterized as relief from an agency action freezing funds that agencies are required by law to pay out. If it seems artificial to say that the case is not really about money damages—after all, the plaintiffs do want money—it is at least as artificial to insist that the case is just a routine dispute over, say, the terms of a contract for military equipment.

2. Another question raised by the APA argument is whether the APA’s waiver of sovereign immunity is narrower than scope of sovereign immunity articulated in Edelman v. Jordan, 415 U.S. 651 (1974). Edelman asserted a broad sovereign immunity over claims to monetary relief against the government, including claims for equitable monetary remedies. If we assume arguendo that Edelman is correct, it is not obvious that the APA’s waiver of sovereign immunity for certain actions “seeking relief other than money damages” has an exception as large as Edelman. Three kinds of monetary relief that are not money damages are: accounting for profits (an equitable remedy), equitable compensation (an equitable remedy that roughly but not exactly coincides with damages), and legal restitution (also called quasi-contract). All of those would seemingly be forbidden by Edelman, but Edelman’s sovereign immunity can be waived by statute, and the wording of the APA exception may well do that.

3. Is the best way to characterize the district court’s order enforcing the temporary restraining order “equitable compensation,” roughly analogous here to compensatory civil contempt? For more on equitable compensation, read Fiduciary Remedies; for more on contempt against government officials and agencies, read The Endgame of Administrative Law: Governmental Disobedience and the Judicial Contempt Power.

4. How far does Bowen v. Massachusetts, 487 U.S. 789 (1988), go? The Court offered a narrow construction of the phrase “money damages” in the APA, reasoning that a challenge to the legality of an agency action precluding the payment of federal dollars was not the same as a backward-looking remedy seeking damages for past wrongs. But Bowen was a case about the violation of a law governing payment rates, not the violation of a contract. Does that distinction matter when the underlying claim is the direct violation of Congress’s general command to spend certain moneys pursuant to contracts?

5. Great-West treats an injunction to pay money as if it were damages. But the notorious error in Great-West was its treatment of non-proprietary restitutionary relief in equity—essentially erasing the category, even though it includes the vast reach of accounting for profits. This error was authoritatively demolished—hit, sunk, battleship—by John Langbein. Great-West also quotes the erroneous statement from Mertens that mandamus is equitable. The point is that Great-West, though sound in its basic insistence that “equitable relief” in ERISA does not refer to any relief that is just and its general invocation of “the days of the divided bench” (rather than 1789), is not a fully reliable guide to the line between legal and equitable relief. Beyond that, Great-West did not involve a temporary restraining order or preliminary injunction. The scope and availability of those interim measures raise distinct questions not present when considering a final injunction.

6. A key weakness in the district court’s enforcement order is that it is mandatory (i.e., an order that requires action, rather than one that prohibits action). Mandatory injunctions are disfavored for good reasons, and those reasons especially hold for mandatory preliminary injunctions and temporary restraining orders. Nevertheless, interim injunctive relief sometimes must be mandatory to preserve the efficacy of the court’s remedial options. But a critical question is how willing the court should be to go beyond what it would do in a temporary restraining order when that temporary restraining order is not complied with. That is essentially what happened here, and equity does go to whatever lengths necessary to ensure compliance with its orders, especially if a defendant shows a propensity to disobey. (Those orders must, however, be issued by a court that has personal jurisdiction over, and has given notice to, the defendants.) How should that critical question be resolved in this case?

7. Relatedly, to what extent should it matter that the underlying claim arguably involves an agency’s failure to act? In Norton v. SUWA, 542 U.S. 55 (2004), the Supreme Court emphasized the need to avoid “broad programmatic attacks” on agency inaction for fear of “injecting the judge into day-to-day agency management.” The Trump administration criticizes the TRO for so inserting the judge into USAID affairs, emphasizing that each contract is its own special snowflake requiring individualized review. At the same time, Norton v. SUWA blesses claims that the “agency failed to take a discrete agency action that it is required to take.” That arguably covers following through on congressionally mandated, contractually specified obligations. Beyond that, it should probably make a difference that it is the Trump administration’s adoption of a blunderbuss spending freeze—a discrete agency action—that is the source of the inaction that it now claims as a defense to judicial review.

8. The dissenting justices are right that in the end the determination will be either that the agency had the authority to stop the disbursements or it did not. If the agency loses at the temporary restraining order stage but ultimately wins the case, then it will be hard to get back the money paid out. If the plaintiffs lose at the temporary restraining order stage but ultimately win the case, then they (and their fired employees) will have suffered extreme hardship in the meantime. So it’s hard here for a temporary restraining order (or preliminary injunction) to serve its office of preserving the efficacy of the court’s remedial options. But that could be mitigated somewhat by creative conditions (and equitable relief is always subject to conditions). The district court could, for example, require payments by the government on condition that the recipients agree to pay them back if ordered by the court. Or the district court could deny a temporary restraining order or preliminary injunction on the condition that the government agree that if it loses it will pay the money with a premium (essentially the cost of borrowing, which could be covered from the Judgment Fund if pursuant to a court order). Or the court, acting with the authority approved in the All Writs Act, could in aid of its jurisdiction compel the payment of the money into the court (comparable to sequestration), to ensure that the money would be payable to the plaintiffs, notwithstanding agency resistance, should they ultimately prevail; and also not lost to the agency, should it ultimately prevail.

The difficulty of these questions confirms the wisdom of not trying to resolve them on an emergency stay of an order enforcing a temporary restraining order.