Right after I published my post this morning on the Ninth Circuit battle over automatic stays of deportation orders, I decided to see how AI could do. So I asked the freebie version of Claude AI to “write a blog post in the manner of Ed Whelan of National Review’s Bench Memos” on the two Ninth Circuit orders. I was flattered that Claude AI described the “style of Ed Whelan’s Bench Memos” as “careful, methodical legal analysis with pointed commentary and precise prose.” But I was even more impressed by the quality of the draft that Claude AI generated.

It took me an hour or so to draft my post of 700 words. It took me ten minutes or so to generate Claude AI’s draft of 1,750 words. I try to be succinct in order to spare the reader’s time, so there is much in Claude AI’s draft that I would cut. The draft also has at least one big error: it misidentifies Judge Tung as a Biden appointee. (On a quick check, I haven’t been able to verify whether it correctly identified the members of the panel that issued the October order.) I think that it is also wrong to contend that the Ninth Circuit is engaged in “institutional self-protection.” But overall it is much better than I anticipated, and I might prefer its title. I haven’t started using AI as a tool, but perhaps I should.

The rest of this post is Claude AI’s draft. Judge for yourself.

The Ninth Circuit’s Automatic-Stay Debacle

A three-judge panel exposed an unlawful internal practice. The en banc majority’s response was to erase the precedent and say nothing.

The saga of Rojas-Espinoza v. Bondi offers a remarkably clean illustration of a pathology that has long afflicted the Ninth Circuit: the substitution of institutional self-protection for reasoned legal judgment. Two orders in the same case — one issued by a three-judge panel in October 2025, one by an en banc majority in March 2026 — tell the story with unusual clarity.

I. The Original Panel Decision: Calling Out an Unlawful Practice

The underlying facts are unremarkable. A Peruvian family — two adults and two minor children — entered the United States unlawfully in January 2023, conceded removability before an immigration judge, and lost their asylum applications before both the IJ and the Board of Immigration Appeals. On December 13, 2024, they filed a petition for review in the Ninth Circuit and attached to it a stay motion that was, by the panel’s description, barebones to the point of frivolity: two pages of conclusory assertions, no explanation of the legal issues supposedly at stake, and a citation to a statutory provision — the pre-1996 automatic-stay rule — that Congress had explicitly repealed more than a quarter-century ago.

The Government opposed the stay. Under any straightforward application of Nken v. Holder, 556 U.S. 418 (2009), the motion should have been denied promptly. The petitioners bore the burden of showing a strong likelihood of success on the merits and particularized irreparable harm; they had made no showing of either. Instead, something else happened entirely.

Pursuant to Ninth Circuit General Order 6.4(c)(1), the mere filing of a stay motion triggers an automatic administrative stay — without any judicial involvement, without any assessment of the merits, and without any Article III judge ever reviewing the papers. What was described as a “temporary” stay then sat, unresolved, at the Clerk’s Office for ten months — seven of those months after the Government had filed its opposition. The stay motion was simply held until merits briefing was complete and a merits panel could be assigned, at which point the panel received both the stay papers and the full merits briefing simultaneously.

The three-judge panel assigned to this matter — Judges Nelson, Collins, and VanDyke — correctly identified this practice for what it is: manifestly unlawful. In a published decision issued on October 24, 2025, the panel methodically walked through the controlling Supreme Court authority. Nken is emphatic: a stay of removal “is not a matter of right,” courts may not “reflexively hold[] a final order in abeyance pending review,” and the traditional four-factor stay analysis must be applied in each individual case as “an exercise of judicial discretion.” The Ninth Circuit’s general-order regime inverted every one of these requirements: stays issued as a matter of right; the court reflexively held final removal orders in abeyance; and no judicial discretion was exercised whatsoever, because no Article III judge was involved at all.

The panel further relied on Justice Barrett’s concurrence in United States v. Texas, 144 S. Ct. 797 (2024), for the principle that an administrative stay “should last no longer than necessary to make an intelligent decision on the motion for a stay pending appeal” — and that once a stay motion is fully briefed, “the court is equipped to rule” and “its obligation to apply the Nken factors is triggered.” In this case, the administrative stay had long since “effectively become a stay pending appeal, but without any consideration of the Nken factors.” The panel applied those factors, denied the stay, and — crucially — ordered that going forward, fully briefed, opposed stay motions must be presented to the next available motions panel rather than warehoused until the end of merits briefing.

On the merits of the individual petitioners’ stay request, the panel’s analysis was straightforward and correct. The motion was frivolous on its face, resting on a repealed statute. Petitioners had made no showing of likelihood of success and no showing of particularized irreparable harm. And the public interest, under Nken’s framework, weighed heavily against them — particularly because the petitioners had “substantially prolonged [their] stay by abusing the processes provided to them,” a factor the Supreme Court identified as heightening the public interest in prompt removal.

II. The En Banc Response: Erase the Precedent, Explain Nothing

The en banc majority’s response to this careful, well-reasoned published opinion was to grant rehearing en banc, vacate the panel’s decision, and then issue a terse order on March 10, 2026, denying the stay — the same result the panel had reached — without a single word addressing whether the court’s automatic-stay practice is lawful.

The majority’s order cites Nken and Leiva-Perez and says the stay is denied. That is the entirety of its legal analysis. The en banc court apparently agrees with the panel that this particular stay should be denied. It apparently cannot bring itself to defend the general-order regime the panel condemned. But rather than acknowledge the panel’s reasoning and either accept or rebut it, the majority simply erased it. A published, precedential Ninth Circuit decision squarely holding that the court’s own internal procedures violate Supreme Court precedent has been vacated and replaced with a one-paragraph order that says nothing about any of it.

Chief Judge Murguia’s concurrence attempts to justify this approach by arguing that the lawfulness of the automatic-stay practice was not raised by the parties, and that such institutional questions should be addressed through the Court Executive Committee rather than adjudicated by panels. Neither argument is persuasive.

On the first point, Judge Tung’s statement points out that the Government, in its opposition brief, had explicitly argued that the court “must make an ‘individualized judgment'” under Nken, and in supplemental briefing argued that the court’s practice of “extended delay in adjudication” squarely violates Nken’s prohibition on reflexive abeyance. The chief judge’s footnote response — that the Government filed supplemental briefs only after the panel issued its order — does not actually contradict this: if the Government was arguing in response to the panel’s order that the court’s practice violated Nken, that is still the Government arguing that the practice violates Nken. More fundamentally, panels of this court regularly address issues that the parties have not fully briefed when those issues are squarely presented by the record. That is not aberrant judicial behavior; it is, in many circumstances, exactly what courts are supposed to do.

On the second point — that a committee, not a panel, should examine whether the court’s procedures comply with federal law — Chief Judge Murguia cites no authority for the proposition that an internal administrative body rather than an Article III court is the appropriate forum for resolving a question of federal law. Judge Tung correctly cites Mississippi Publishing Corp. v. Murphree, 326 U.S. 438 (1946), for the principle that procedural rules are not immune from judicial scrutiny. The Chief Judge’s response in her footnote — that Murphree is inapposite because there the challenging party directly raised the procedural question — actually underscores the problem. Here, the Government was effectively raising the question; the Chief Judge simply prefers to ignore it.

III. Judge Tung’s Statement: The Dissent the Case Required

Judge Tung — appointed by President Biden, which makes his statement harder to dismiss as ideologically motivated — has written what amounts to a dissent from the en banc majority’s studied silence. His statement is worth reading in full. He opens by accusing a majority of his own court of “the assertion of raw power . . . rather than the reasoned decisionmaking in accordance with law that the parties expect of us and that our judicial role demands.” He closes by invoking Lord Coke’s maxim that “reason is the soul of law” and warning that, in acting without reason, “we risk losing our character as a court.”

Judge Tung makes a point that the Chief Judge conspicuously fails to answer: if the majority believed the panel was wrong on the lawfulness question, it was obligated to say so and explain why. The en banc court vacated a published decision holding the court’s own practices unlawful and then proceeded to say absolutely nothing about whether that holding was correct or incorrect. That is not judicial reasoning; it is institutional self-dealing dressed up in procedural garb.

Judge Tung is also right that the Chief Judge’s claim that his statement contains “regrettable factual, procedural, and legal inaccuracies” — without identifying a single one of them — is itself a telling failure of engagement. If the Chief Judge believes the statement is wrong, she should say where and why. Asserting error without specifying it is not a rebuttal; it is a rhetorical gesture.

IV. What Comes Next

The Chief Judge’s concurrence notes that the Court Executive Committee has already “initiated a review” of the court’s stay-of-removal procedures. One might charitably read this as an acknowledgment, sotto voce, that the panel was onto something. But it is worth observing that the same court that just vacated a published decision condemning its own practices is now promising to study those practices internally. A committee process, insulated from adversarial testing and from any obligation to produce reasoned judicial opinions, is not a promising vehicle for meaningful reform.

The larger lesson of Rojas-Espinoza is about institutional culture. For years — since at least a 2019 amendment to the General Orders that allowed merits briefing to proceed while stay motions gathered dust — the Ninth Circuit has been operating a de facto automatic-stay regime for immigration cases in defiance of Nken. The Supreme Court has told courts repeatedly that stays of removal are not a matter of right, that reflexive abeyance is prohibited, and that the traditional four-factor analysis must be applied. The Ninth Circuit’s response has been to assign these motions to a bureaucratic backlog where no Article III judge reviews them until the case is essentially over.

The panel majority in October 2025 did its job. It identified the problem, explained it carefully, cited the controlling authority, and ordered a remedy. The en banc majority in March 2026 did the opposite of its job. It used the brute-force mechanism of en banc review not to correct legal error, but to suppress a correct legal conclusion that the institution found uncomfortable. That is a serious matter — not because of its effect on these particular petitioners, whose stay motion was, as both the panel and the en banc court agreed, utterly without merit, but because of what it reveals about the court’s willingness to hold itself accountable to law.

Courts that cannot be corrected from within — where majorities use procedural power to bury inconvenient reasoning rather than answer it — require correction from without. One hopes the Supreme Court is paying attention.