Over at the Volokh Conspiracy, Sam Bray has this fascinating and timely post on why the Administrative Procedure Act (APA) does not allow for nationwide injunctions:

Sometimes the question is asked whether the Administrative Procedure Act authorizes courts to give national injunctions, because it says that a “reviewing court shall . . . hold unlawful and set aside agency action, findings, and conclusions found to be . . . arbitrary . . .” (5 USC § 706). In this post I’ll offer a quick reminder of why the APA does not authorize national injunctions.

First, when the APA was enacted the expectation was that agencies would make policy primarily through adjudication, not through general rulemaking. The leading source for this is Reuel Schiller’s excellent article “Rulemaking’s Promise: Administrative Law and Legal Culture in the 1960s and 1970s.”

Second, “set aside” was a technical term for reversing judgments. This can be seen in Morgan v. Daniels, 153 U.S. 120, 124 (1894). This understanding is also consistent with statutes that appear to treat “enjoin” and “set aside” as distinct actions, such as 28 USC § 2112(a). Note that the case so far is cumulative: “set aside” as a term for reversing judgments, not for giving national injunctions, is exactly what we would expect if Congress were anticipating a norm of agency policymaking through adjudication.

Third, it would be very odd to think “set aside” means “enjoin enforcement of against anyone” given the full set of objects for the verb in Section 706. Those objects are: “agency action, findings, and conclusions.” Findings and conclusions are not enjoined. But they may very well be set aside by a reviewing court. At this point the idea that the APA authorizes or even requires national injunctions hangs by a thread—one has to postulate a statutory zeugma. As wonderful as a good zeugma is—and who doesn’t love a good zeugma?—it seems like a strained reading.

Fourth, the complete absence of national injunctions in the decades before and after the APA makes it highly unlikely that the text was understood by Congress to authorize or require national injunctions. (For evidence, see “Multiple Chancellors: Reforming the National Injunction.”)

So, there are two options. You could think that Congress, never having seen a national injunction and not expecting a lot of agency rulemaking, decided it would authorize national injunctions through an idiosyncratic use of a technical term (“set aside”) embedded in an extremely rare figure of speech (zeugma) but no one understood this was happening until the 1960s. Or you could think that the APA does not authorize national injunctions.

Bray’s position obviously cuts against conventional wisdom in administrative law. My initial reaction is that I’m still not convinced, though I need to think more carefully about the argument.

After all, a final rule is an “agency action” under the APA, so it seems it can be “set aside.” The effect of that is the invalidation of the final rule—in essence, a nationwide injunction (though courts have approved of a narrower remedy of remand without vacatur in some circumstances).

Bray’s arguments about how Congress expected agencies to make policy primarily through adjudication and how nationwide injunctions were not common shortly before or after the enactment of the APA don’t really move me precisely because Congress did not expect agencies to make policy through rulemaking and, indeed, agencies generally did not make policy through rulemaking during the time period in question. So one wouldn’t expect to find evidence of courts invalidating final rules then, and the 1946 Congress that enacted the APA would have had no reason to focus on the practical effect of setting aside a final rule. Now that many agencies primarily make policy through rulemaking, courts have begun to set aside invalid final rules. That remedy strikes me as the most natural reading of the ordinary meaning of “set aside” a final rule.

What gives me pause, however, is Bray’s argument that “set aside” is a term of art that refers only to reversing judgments. If Congress was indeed using “set aside” in this technical manner, then I might need to rethink my current stance against remedial restraint in administrative law—at least as a matter of statutory interpretation.