As Everyone Expected, Presidents Do Not Have Immunity from Criminal Prosecution

The United States Court of Appeals for the District of Columbia Circuit decided an interlocutory appeal on Tuesday, Feb. 6, in United States v. Trump, in which it held that former Presidents of the United States do not have immunity from criminal prosecution for actions taken while in office and that the prohibition on double jeopardy does not prohibit bringing criminal charges against a former official who was acquitted in an impeachment trial about related conduct. The outcome was almost a foregone conclusion — no serious observers thought that there was a substantial chance that the Court of Appeals would accept former President Trump’s arguments, which ranged from bad to terrible, that he could not be tried for his role in the effort to overturn the 2020 election. Nonetheless, the decision is still of great importance because it is essential to preserving the rule of law and preventing the President from becoming an elected despot. I don’t think the court’s handling of this appeal was particularly good — the court acted without adequate alacrity in light of the circumstances of the litigation and the opinion is of mediocre quality, although I agree with the outcome of the appeal. In this essay, I will discuss and critique the court’s opinion and process.

A United, Bipartisan Panel of Judges

As with the overwhelming majority of court of appeals proceedings, this appeal was heard by a three-judge panel of the D.C. Circuit. The decision was written per curiam — literally “by the court” and meaning that the opinion was not signed as authored by a specific member of the three-judge panel — and joined by all three judges without any concurrences or dissents. The most typical use of a per curiam opinion is deciding a case of little consequence, often with a brief “unpublished” opinion that is not binding precedent in other cases. This case, of course, does not fit that pattern in any regard — this is an important (even though easy) case, the opinion is not brief (54 pages), and it is a published, fully precedential decision. In this case, my assumption is that they divided up the opinion between the three chambers and then merged the work into a single document to speed the process of issuing the decision. They may also have been motivated by a desire to avoid focusing the danger of politically motivated violence from having a single target; the district court judge in this matter has been the victim of swatting, and similar or worse violence is definitely possible.

The panel consisted of one Republican judge, Karen L. Henderson (appointed to the district court by Ronald Reagan and raised to the court of appeals by George H.W. Bush), and two Democratic judges, J. Michelle Childs (appointed to first the district court by Barack Obama and later to the court of appeals by Joe Biden), and Florence Y. Pan (appointed to both the district court and the court of appeals by Joe Biden). Coincidentally, both Judge Henderson and Judge Childs happen to be from South Carolina, although they are otherwise very different.

I do not have a high opinion of Judge Henderson; I was part of a large team of lawyers who litigated a major case before a three-judge district court panel that she sat on, and I was unimpressed with her work product. I view her as quite far over to the “political hack” side of the judicial spectrum. She has also been widely criticized for hiring almost exclusively men for her law clerks. Clerking for a circuit court judge, especially on a high prestige circuit like the DC Circuit is a major jumping off point for a successful legal career, and in light of a world in which roughly half of law students and roughly half of the best law students are women, it’s almost impossible to imagine that a judge could hire almost exclusively men without engaging in direct, intentional discrimination.

I do not have any personal awareness of Judge Childs or Judge Pan. Judge Childs was on Biden’s short list of candidates to fill the Supreme Court seat ultimately filled by Justice Ketanji Brown Jackson. Judge Pan is a recent addition to the court of appeals, having only served on the court for about a year and a half, but has sterling credentials and an impressive career before joining the court.

I’m certain that all three judges understood the importance of a unified, unanimous decision in this case. Had they split in any major way, even if they agreed on outcome, that would have been perceived as an obvious sign of partisanship and run the risk of damaging the court’s reputation. I don’t think that shaped the outcome of the case — as I’ve said, I think the outcome was certain — but I think it may have made them extra careful to avoid a situation where there was any sign of disunity.

The Major Process Failure: Failing to Act Quickly Enough

The most important failure of this panel was how slowly it acted on this matter. We’re in the middle of a presidential election campaign, and the court was considering an appeal about whether an individual who is very likely to be one of the two major party nominees can be tried in a criminal matter. The D.C. Circuit expedited consideration of this appeal to a substantial degree and acted much more quickly than in the ordinary course of consideration, but even so, the court has pushed the start of any possible trial until after the primaries will functionally end and raised a substantial risk that no trial court verdict will be reached until after the Republican convention is over or possibly even until after the general election.

The district court rejected Trump’s immunity argument on December 7, 2023. Four days later, the United States asked the Supreme Court to grant a writ of certiorari to resolve conclusively any appellate process with regard to the immunity argument. Skipping the intermediate court of appeals is highly unusual, but the Supreme Court’s rules do permit it for exceptional cases. However, the Court refused to grant cert in this matter on December 22. Concurrently with the Supreme Court’s consideration of the cert petition, the D.C. Circuit issued an expedited briefing schedule on December 13 and set oral argument for January 7. The last briefs by the parties were filed on January 2, with a handful of amicus briefs filed a few days later.

What all this meant was that the appeals court had plenty of time to prepare a draft opinion before the oral argument on January 7. The court moved very quickly relative to standard practice to get the case heard by January 7 — requiring the parties to fully brief the case over about three weeks that included the Christmas and New Years holidays. But then, rather than deciding the matter promptly and allowing the district court to potentially keep the March trial dates that had been scheduled, the court of appeals sat on the matter for nearly a month.

Part of the expectation of lawyers is that they will research and write material very quickly when necessary. A standard move of a trial lawyer when an important oral motion during a trial looks like it will be denied by the judge is to ask permission to file a written brief — and it’s not unusual for the judge to grant permission and to ask for the brief to be filed the next morning. Writing substantial, important legal documents under deadlines of a few days is part of the profession. In fact, I’ve even encountered a situation where a court gave the legal team I was part of literally about three hours to file a reply letter brief — a time table so tight that we were literally editing the file entering changes to the early pages of the brief while the later pages of the draft were still printing. That’s exceptional, but expecting the court to handle a matter of grave urgency in a few days is perfectly reasonable. If it had taken 72 hours after oral argument, I would have no complaints about the court’s process.

Instead, here we are, a month later, and the decision has finally been issued. The mandate — the official command of the court, in this case returning the case to the district court — would not issue until February 13, at the earliest. Because Trump filed a motion on February 12 in the Supreme Court asking for a stay in the case pending resolution of a cert petition and pending consideration for rehearing by the en banc D.C. Circuit, the case will remain stayed until the Supreme Court acts on that motion — likely until the Court decides whether to grant cert. While the Supreme Court could act almost immediately, whether to grant or to deny cert, Chief Justice Roberts (who serves as the circuit justice for the D.C. Circuit) gave the special counsel until February 20th to respond to the stay application. The special counsel in fact filed the United States’s response two days later, on February 14, demonstrating the speed with which competent lawyers can act when necessary, requesting that the Court deny the stay or in alternative immediately grant certiorari and schedule the case for resolution this term (i.e. before the end of June). It is likely that the Court would not deny cert until the end of February at the earliest. So even if the Supreme Court decides, sensibly in my opinion, that it wants no part of this, the district court proceedings won’t likely resume until March. If the Supreme Court grants cert, even if they expedited the case to the same degree as the court of appeals, we would be looking at oral argument in mid to late March at the earliest, with a decision likely not issued until around the end of March at the earliest and mid to late April entirely plausible.

The district court had about 3 months of proceedings scheduled before trial under the old schedule, before this appeal derailed that, and then budgeted substantial time for the trial. If the court maintained that schedule, the trial might begin around the beginning of June if the Supreme Court denies cert quickly. If the Court dawdles, or if the Court grants cert and then affirms, that might put us on a path to a trial beginning in July or August. All of that means that, even without additional delays (and Trump’s lawyers have sought delay at every opportunity), trial would begin in the summer, during the height of the presidential election campaign and after the primaries are entirely settled. If anything delays matters further, the trial could begin in the fall.

This is all far too close to the presidential election for comfort. Imagine early voting taking place before a guilty verdict sends Trump to jail. Or for that matter, imagine a conviction days before the Republican convention. Would the convention still nominate Trump for President? Would they even have time to adequately consider whether this changes matters, let alone who else they might nominate? And the prospect of a conviction immediately after the general election is no better, when many polls show that a conviction would have a substantial influence on some voters’ decisions. How would the system even cope with a President-elect who is also a newly convicted criminal? The prospect of someone needing to report to a federal prison while preparing for inauguration feels preposterous. The problem isn’t limited to the prospects of a conviction — if Trump is acquitted, there are still major issues of fairness in not having the decision rendered well before the general election and with a major party nominee needing to balance time campaigning with time spent on trial.

Trump wants delay, of course — he has argued that the case should be dismissed because of immunity, but that even if it is to go forward, any trial shouldn’t take place until after the election. Of course, were he to win the election, he would then immediately argue that any trial should be postponed until after he had left office. Sufficient delay may be equivalent to immunity if the voters return Trump to office.

If the Supreme Court correctly affirms the Colorado Supreme Court’s Section 3 disqualification decision, all of this becomes moot. If Trump can’t be re-elected, then delay matters no more and no less than in any other criminal trial. However, if the Supreme Court reverses the Colorado Supreme Court’s decision, the panel’s failure to treat this case with the urgency it deserved may end up creating a disaster.

But at least the panel had the time to spend on making the opinion as strong as they could… They got the right outcome — an outcome that nearly everyone agreed was inevitable. Did they at least produce an expertly crafted decision worthy of respect? Not really. They spent a month after oral arguments to write a shoddy opinion.

The Panel’s Opinion

The panel’s decision begins with a fairly extensive discussion of the factual background of the January 6 charges against former President Trump and the procedural history up to this point. There are some nice rhetorical flourishes in the introduction about how after the end of his term, Trump went from President Trump to being “citizen Trump, with all of the defenses of any other criminal defendant.” Op. at 1. The rest of the introductory and background materials are largely not noteworthy, except perhaps for some statements noting that at this stage in the proceedings the court cannot consider whether adequate evidence exists to support the indictment’s claims. That’s obvious to a lawyer, but with a decision like this that will be read by many nonlawyers, it’s worth underscoring, for the same reason that lawyers get annoyed when nonlawyers treat the allegations in a civil complaint as accurately describing what actually happened. An indictment is more than just allegations, but not much more, and indictments and civil complaints include many statements that turn out to be not quite true, even if the allegations turn out to be mostly true. It’s important that people be reminded not to treat an indictments claims as proven, even if a court needs to take them as true for specific pretrial motions.

The opinion then turns to three major legal issues: (1) Is this the sort of matter that can be appealed prior to a final judgment? (2) Do Presidents have immunity against criminal prosecution for acts taken while in office, and if so does that immunity apply here? and (3) Does a criminal trial after an acquittal in an impeachment trial violate concepts of double jeopardy?

The first issue is actually the most challenging one in the abstract, because of a specific Supreme Court decision, though it ends up being settled by clear circuit precedent. The second issue is of course the most important one. The third issue is the one that the opinion handles best.

Is an Interlocutory Appeal Appropriate for a Presidential Immunity Claim?

The basic rule in litigation is that you can only appeal from a final judgment, not from adverse decisions that a court makes along the way before reaching a final judgment. Lawyers creatively refer to this as the “final judgment rule.” The primary rationale for the final judgment rule is that it serves judicial efficiency. Rather than hear a series of appeals about every issue that comes up in a trial court, a single appeal can handle all of them at once. If cases routinely bounced back and forth between trial and appellate courts, they would also last forever, and the courts would have to refamiliarize themselves with the case repeatedly for each separate appeal. Also, many issues seem important during a trial but turn out to be unimportant after a trial — for example, it’s common for judges to work to “appeal-proof” their decisions by deciding marginal issues in favor of the litigant who is losing the case overall. It may seem very important to a plaintiff that the judge is excluding a piece of evidence that the plaintiff views as really powerful until the judge rules in the plaintiff’s favor without that piece of evidence because the rest of the case was so strong. Allowing the plaintiff to appeal that evidentiary decision would simply waste everyone’s time — even if the plaintiff is correct that the judge got it wrong, it doesn’t matter.

While the final judgment rule is the basic standard, there are exceptions in which an “interlocutory” appeal can be brought — an appeal before the trial court has reached a final judgment in the case. An interlocutory appeal simply means an appeal that isn’t from a final judgment. Interlocutory appeals are fairly common in certain civil litigation contexts — for example, when a court issues a preliminary injunction (an order to do or not do something until after trial), an interlocutory appeal of the injunction is proper before the trial court holds a full trial and issues a permanent injunction. However, they are highly disfavored and quite rare in the criminal context.

The general understanding in the criminal context is that errors in a trial court proceeding can be adequately resolved by an appeal after a final judgment. For example, if a defendant argues that a piece of evidence was taken in violation of the Fourth Amendment and ought to be suppressed, and the trial court disagrees, the defendant can appeal a conviction and have the conviction overturned if the appeals court agrees that the evidence should have been suppressed and would have made a difference. Of course, if the defendant is acquitted, no appeal of the issue is ever necessary (and if the evidence is excluded, the prosecution generally can’t appeal anyway), so this greatly serves interests in judicial efficiency and in providing criminal defendants speedy trials.

However, some rights cannot be vindicated by an appeal after a final judgment. The most obvious example is the protection against Double Jeopardy, which protects a defendant from needing to go on trial for an offense for which they have already been charged and acquitted. If a defendant has to go through the entire difficult process of standing trial (and perhaps being jailed pending trial) for an offense for which they have already been tried, an appeal after final judgment cannot remedy the harm of the trial. And while many Double Jeopardy claims are straightforward to resolve, such that error by a trial court is unlikely, complicated edge cases always exist. As a result, courts will permit interlocutory appeals of decisions concluding that a new trial would not violate Double Jeopardy, without requiring the defendant to wait for a verdict and final judgment.

At first glance, a right to an interlocutory appeal in this case seems clear. Presidential immunity, if it exists, would be an immunity from trial, not merely an immunity from punishment. That’s how other forms of immunity work, whether that’s the absolute immunity that judges have from civil liability for their official acts or the qualified immunity that courts have made up to protect government officials who have violated peoples rights if courts haven’t specifically said that that particular violation is illegal. And indeed, both the prosecution and the defense agree that this particular interlocutory appeal can be properly heard by the Court of Appeals.

What complicates matters is that an amicus argued that this interlocutory appeal is not permitted, and because that question goes to the jurisdiction of the court (whether the court has the power to rule on this question), the court must satisfy itself that it has jurisdiction even if all the parties agree that it does. The standard test for whether an interlocutory appeal is proper is three parts: (1) does the decision completely settle the issue under appeal; (2) is the issue under appeal completely separate from the merits of the case; and (2) would the issue be effectively unreviewable after a decision on the merits. As I noted, claims of immunity fall in the heartland of issues where an interlocutory appeal is proper. However, with regard to interlocutory appeals about criminal law matters specifically, the Supreme Court has said that the exceptions to the final judgment doctrine are resolved with the “utmost strictness in criminal cases” and that interlocutory appeals based on a right not to be tried must be based on “an explicit statutory or constitutional guarantee that trial will not occur.” Op. at 10, quoting Midland Asphalt Corp. v. United States, 489 U.S. 794 at 799 and 801 (1989) (emphasis added by the panel opinion).

None of the claims of immunity that Trump advances in this case are based on an explicit statutory or constitutional guarantee not to be tried. Nowhere in the Constitution does it say that a former President has immunity from criminal prosecution for actions taken while in office, and there is likewise no explicit statutory grant of immunity. If any such immunity existed, it would be grounded in the overall structure and policy considerations in the Constitution and based on negative inferences from the clause of the Constitution permitting criminal trial of officials after conviction in an impeachment trial. In other words, any Presidential immunity here would be based on an implicit constitutional guarantee not to be tried, not an explicit one. Likewise, while Double Jeopardy immunity does have an explicit basis in the Constitutional text (“… nor shall any person be subject for the same offense to be put in jeopardy of life or limb, …” U.S. Constitution, Amendment 5) and can be the basis of an interlocutory appeal, Trump’s claim is based on a broader conception of double jeopardy than the Double Jeopardy Clause itself covers, and thus again is an effort to use a claimed implicit constitutional immunity. Taken literally, Asphalt Midland would thus be binding precedent requiring the panel to dismiss this interlocutory appeal for lack of jurisdiction.

The panel opinion then goes on a long exegesis of Asphalt Midland in which it argues that this argument puts far too much weight on the word “explicit,” which appears only once in the Supreme Court’s decision, when the overall structure of Asphalt Midland suggests that the Supreme Court was actually focused on distinguishing between a right to not be tried, on the one hand, and a right to have charges dismissed, on the other. The panel argued that the Supreme Court has raised the possibility in a later case that the use of the word “explicit” was a suggestion, not mandatory, and that that was dicta (a statement not necessary to the reasoning behind the court’s conclusion, and therefore not binding precedent, although still persuasive). The panel also raised policy arguments, suggesting that not allowing Presidential immunity, if it exists, to block a criminal trial would impair the dignity of the presidential office, and likewise the lack of a principled distinction between immunities that have an explicit basis in text and immunities that courts have recognized that don’t have an explicit basis in text.

The panel’s discussion presents some good arguments for why Midland Asphalt’s reference to “explicit” constitutional or statutory immunity from trial should be read as dicta, against which would have to be weighed the requirement that lower courts take the Supreme Court at its word and apply its decisions, even if the Supreme Court would not continue to apply the statement. After all, while dicta is not binding, a statement that constitutes part of the holding of the court (the reasoning that the court used to determine the result) is still binding even if the holding could have been written in a way that didn’t include that statement. In any event, all of this bloviating and careful parsing of Midland Asphalt ends up being entirely unnecessary to the panel’s decision.

The panel cites circuit precedent interpreting and applying Midland Asphalt and allowing implicit immunities to be raised in interlocutory appeals even in criminal cases. The D.C. Circuit decided a case in which a senator argued that, analogously to the Speech and Debate Clause that protects senatorial debate from being the basis for civil or criminal suits, the constitutional provision creating the Senate’s ability to set its own rules creates an implicit immunity from criminal cases that would require interpreting the Senate’s rules. United States v. Durenberger, 48 F.3d 1249 (D.C. Cir. 1995). In that case, the court described this claim as “colorable,” which I generally describe as lawyer-speak for “dumb and wrong, but not so dumb and wrong that it’s frivolous” and said that was enough to permit an interlocutory appeal of a denial of immunity, notwithstanding Midland Asphalt.

A basic principle of how the courts of appeal operate is that they are bound by the precedent of previous panels in the same circuit, unless and until that precedent is overturned either by an en banc (sitting of all the active judges of the court together) decision in that circuit or by new Supreme Court precedent. Durenberger is, as the panel acknowledges, directly controlling circuit precedent. That’s it! That settles the issue fully. There’s no need to analyze whether Durenberger was correct to treat the word “explicit” in Midland Asphalt as more of a guideline than a rule — this panel is bound by Durenberger’s holding even if this panel thinks that Durenberger’s interpretation of Midland Asphalt was indefensible. Normally, it would merely be poor style to spend unnecessary time and text working through the details of why apparently binding Supreme Court precedent does not control, even though clearly binding circuit precedent settles the issue. A fulsome discussion can even be defended as an effort to harmonize away the apparent contradiction and improve the further development of the law. In this context — where no party raised the issue, time is of the essence, and the opinion would be long and have taken too long to write even without this section — the waste of judicial resources is highly disappointing. This entire section of the opinion could have been three paragraphs: lay out Midland Asphalt and the amicus’s argument, point out that that can and has been read as dicta and that the real focus was on the distinction between protection against trial and a ground to dismiss the case, and then cite the controlling circuit precedent. In fact, it’s not even strictly necessary to include this section at all — once the court concludes that there is controlling circuit precedent, it could have handled the jurisdictional question with a footnote or a single sentence stating the D.C. Circuit’s position with a string cite to the prior circuit opinions. But in a case where the panel took a month to issue an opinion that it should have gotten out in 72 hours or less, spending 10 pages giving a jurisdictional issue controlled by circuit precedent every possible consideration represents a serious failure of judgment.

The only defense of this can be that citing circuit precedent without harmonizing that with the language of the Supreme Court’s decision in Midland Asphalt might make the Supreme Court more likely to grant certiorari to consider whether the circuit precedent is good law. But this extended argument doesn’t really resolve that issue — at the end of the day, the panel is still left with “we’re bound by circuit precedent,” and the Supreme Court, which isn’t, could still choose to reopen the question. So even if this was an effort to prevent a grant of certiorari, it doesn’t really accomplish much — especially when neither party will seek cert on this ground, and an amicus cannot seek cert.

Rejecting Trump’s All-Encompassing Claim of Presidential Immunity

Turning to the substantive analysis: Trump argues that he is entitled to immunity from criminal liability for all actions that fall within the outermost ambit of the concept of official duties of the President. This theory has no basis in the constitutional text, in history, in court precedent, and would make the President literally above the law and able to take almost any action with impunity. Combined with the pardon power, which allows the President to unilaterally protect any of his subordinates from federal criminal liability, this would allow the President to use the awesome might of the federal government to do whatever he wished. “Will no one rid me of this meddlesome [Supreme Court Justice, Senator, rival political candidate, whoever]?” Indeed, at oral argument, Trump’s lawyer argued that the President would remain free of criminal liability even if he ordered the military to assassinate his opponent in an upcoming election — an unbelievable assertion that demonstrates how crazy Trump’s position is, and should have been (but wasn’t) cited in the opinion to demonstrate how much Trump’s position would transform the president into a king who is beyond the reach of the law.

Under existing precedent, Presidents have absolute civil immunity for their official acts, with “official acts” understood extremely broadly. They remain civilly liable for private acts taken while President. The question this case poses is whether Presidents have a similar immunity from criminal prosecution. The panel considered and rejected three arguments in favor of Presidential immunity: (1) Separation of Powers concerns; (2) policy justifications in terms of ensuring effective functioning of the executive; and (3) a negative implication from the Impeachment Judgment Clauses statement that the punishment for conviction in an impeachment process extends only to removal from office and a prohibition from holding office in the future but ordinary criminal prosecution is still permitted, from which Trump draws the idea that ordinary criminal prosecution of a former President is only permitted if the President has been first impeached and convicted.

The panel begins by examining the separation of powers arguments. It recites some standard statements about the unique importance and specialness of the President, and then turns to distinguishing a quote from Marbury v. Madison that the President’s actions “can never by examinable in court,” 5 U.S. 137, 166 (1803). The panel spends a long time distinguishing between discretionary acts, which cannot be reviewed by courts, and ministerial acts, which can. In essence, the panel tries to define this in terms of what acts represent political judgments (and thus are beyond the ability of the courts to question) and what acts are required or forbidden by law, and thus are amenable to judicial review. It then traces judicial review of Presidential power through a series of major cases, noting that Presidential power is at its maximum when it acts to enforce an affirmative act of Congress and at its minimum when legislation specifically restricts Presidential action. All of this is really besides the point — these are cases about whether the President’s attempts to take specific actions in office are within the scope of their power and whether a court can enjoin their actions (or the actions of subordinate executive officials carrying out their actions). None of this really deals with the core issue in this case: is there some license for the President to break criminal law with impunity? Eventually, the panel decision does get to those issues, quoting sweeping Supreme Court language about how “No man is so high in this country that he is above the law.” Op. at 24, quoting United States v. Lee, 106 U.S. 196, 222 (1884). I find it a little sickening that the panel quoted this heavily gendered language from nearly a century and a half ago without either alteration or even a comment, but except for the use of “man” and “he,” the point is sound and gets directly at Trump’s claim. The panel then notes that courts have held that even a sitting President is subject to criminal process, such as subpoenas to compel testimony. This, too, is highly relevant to this question, unlike the strange foray into questions of judicial review more broadly.

Only after having wandered through paragraph after paragraph of discussion of hoary old cases that are only marginally relevant does the panel turn to one of the best guides for the scope of Presidential immunity: comparing it to the immunity that other government officials enjoy. Judges and legislators are absolutely immune from civil liability for their official actions. A judge can wrongfully, maliciously, and in complete bad faith ruin someone’s business, take their home, and leave them destitute, and yet no civil action for damages will be recognized by a court. (It’s almost like the rules about judicial immunity were created by judges…) Similarly, a state legislator can enact a law that obviously violates the Constitution, fully aware of the unconstitutionality, and still be immune to civil suit from someone whose property is destroyed before the law can be enjoined. And of course the Speech and Debate Clause provides members of Congress absolute immunity from suit for their statements in Congressional proceedings, no matter how libelous and knowingly false those statements may be. But none of those immunities, save only the Speech and Debate Clause, provides any protection against criminal prosecution for wrongful acts.

When a legislator contacts a government employee to seek to assist a constituent, they are engaged in official acts, even though not directly legislative, and civil immunity attaches. However, if the legislator is attempting to obstruct justice or engaged in a conspiracy to commit a crime when they do so, they remain subject to criminal prosecution for their actions. The allegations against Senator Menendez are a perfect example currently — many of the allegations, perhaps all of them, involve official acts, and he would likely enjoy civil immunity, but he can still be tried criminally for those same acts. Numerous decisions have upheld prosecutions for wrongful official acts in the past. Likewise, while a judge cannot be sued civilly for their official acts, they can be held criminally liable if they engage in fixing cases or other criminal activity in their official acts. The panel uses a case allowing prosecution of a state court judge for violating a criminal law prohibiting racial discrimination in selecting jurors. Ex parte Commonwealth of Virginia, 100 U.S. 339 (1879). That case, like the panel’s earlier discussion of Marbury and its progeny, tries to cram this square into the round hole of a discretionary vs. ministerial distinction, arguing that excluding people based on race violates a ministerial duty in selecting jurors. Op. at 26–27. This is nonsense, of course — selecting jurors involves a substantial amount of discretion and judicial judgment, and even the 19th century Supreme Court, a much more formalist body than modern legal practice, recognized as much and focused on the actual issue that judicial immunity does not give judges the prerogative to ignore criminal law.

The panel then marches through a series of more recent precedents that continue to hold to the same pattern: judicial and legislative immunities (other than Speech and Debate Clause immunity) provides civil immunity, not criminal immunity. Based on that, the court correctly concludes that Presidential civil immunity does not extend to criminal immunity.

This section is a mixed bag — the treatment of Separation of Powers doctrine is so inapposite that the efforts to make it relevant simply make it confusing. At the same time, there is some need to respond to the defendant’s arguments, if only to reduce the likelihood that the Supreme Court will feel that it needs to review the cases. Still, this could have been presented in a much more effective way, leading with the strong arguments for the limitation of immunity to civil immunity, then into the case law making Presidents answerable to subpoenas in criminal investigations, before explaining why the quotations from Marbury do not require a different result.

The panel then moved on to what it described as policy considerations. I don’t like the treatment of these as a separate ground — they’re really part of one overall analysis, and it’s hard to understand how the discussion of how judicial and legislative immunities work as being about “separation of powers.” There’s a slight argument for legislative immunity, but judicial immunity isn’t about separation of powers at all. It’s about a belief that it’s necessary to protect judges from tort suits for errors, even egregious ones, to make judges willing to do their duty, while still retaining potential criminal liability to address misconduct like accepting bribes or deliberately perverting justice for private purposes. (I think some form of limited immunity to civil immunity would still be adequate — requiring active bad faith and knowing misconduct — but nonetheless, judicial immunity is absolute in civil, but not in criminal, cases essentially for policy reasons.) Indeed, both judicial and legislative immunities date back to prior to the framing of the Constitution and to English practice, where separation of powers historically did not exist, with the highest court of appeals also being part of the legislature and the head of the executive (leaving aside the monarch, now a figurehead but much less so prior to the framing) as the party leader in the legislature. My best guess for why some of the policy considerations are shoe-horned into the “separation of powers” section is that each of these sections was drafted in a different chambers and then merged together without the reorganization that would have made the argument flow better. Merging together pieces of drafts written by different people is common in legal practice, although rarer in judicial practice; usually, however, one of the more senior people involved in the process carefully makes sure that the outcome doesn’t look like the work of Dr. Frankenstein.

What’s worse, the policy arguments section then makes a lot of noise about how they are limited to this specific context: a former president accused of having sought to prevent the peaceful transition of power after a lost election. Some of the limitations make sense, notably the footnote pointing out that this does not cover sitting presidents or state prosecutions. The considerations in each of those possibilities are substantially different, and courts generally tailor their holdings to the case in front of them. I also understand the reluctance to sweep all presidential action into a “no criminal immunity” general rule — consider for example a president’s use of their inherent power as commander-in-chief. Intuitively, it seems clear that many actions presidents take would be criminal if they weren’t taken as military actions (e.g. ordering the use of lethal force under circumstances permitted by the international law of armed combat but under circumstances where they could not justify deadly force under ordinary civil laws about homicide). The obvious conclusion ought to be something like presidential military orders can create criminal liability for the president if they are war crimes or violate specific criminal statutes about the use of military authority, but not if they are consistent with the laws of armed conflict but violate general criminal laws about homicide.

This court did not and should not try to demarcate all of the boundaries in the hardest edge cases, but it strikes me that this opinion would be substantially better if it announced a general rule that former president’s have no criminal immunity for actions taken in office that are not directly and purely part of their official conduct, narrowly understood. If a former president were accused of raping someone while in office — a dismayingly plausible situation, in light of the fact that I can think off-hand of at least 3 rapists who served in that office (Jefferson, Cleveland, and Trump) — the same arguments that no criminal immunity should exist would apply with the same force. Bush v. Gore style disclaimers that a case is only precedential for this particular matter are unseemly and undercut the proper and impartial functioning of the law.

Though this section is labeled as “Functional Policy Considerations,” suggesting that it is arguments about the practical effect each outcome would have that are not heavily grounded in legal doctrinal analysis, much of it reads like run of the mill legal analysis, citing precedents and quoting the language of prior opinions in related but distinct contexts to reject Trump’s claims of presidential immunity. It begins by addressing a claim to categorical immunity. One of the most telling points here is that even Trump’s lawyers conceded that any presidential immunity, if it did exist, would end upon conviction in an impeachment inquiry — a conclusion almost necessarily following from the statement in the Constitution that individuals convicted by impeachment trial remain subject to indictment under the criminal law. The panel also noted that President Ford’s pardon of former President Nixon inherently suggests that President Ford thought that Nixon was otherwise in danger of prosecution for his misconduct in connection with Watergate — an opinion shared by much of the nation. And the scarcity of prosecutions of former presidents suggests that there is little risk that criminal prosecution would become a routine form of political tussling. That point is fair enough, though the collapse of the norm that impeachment is a tool for addressing extraordinary misconduct and its use in purely political proceedings where there is no reason to believe that the impeachment has any merit, as in the impeachment of Secretary of Homeland Security Alejandro Mayorkas and the December 2023 House vote in favor of conducting an impeachment investigation about President Biden, suggests that some uniform patterns may rest on changeable norms. The panel is on stronger ground, however, when it discusses the compelling government interests in ensuring that federal law is obeyed and that the President, like any other individual, is not above the law.

After its discussion of general principles related to presidential immunity, the panel moved on to focus on the specifics of this case: the allegations that Trump attempted to prevent President Biden from lawfully taking office. The panel notes that these are fundamentally allegations that Trump tried to subvert both key portions of the Constitution, such as the provision of a four-year presidential term, and the fundamental principles of democracy. In this case, the needs to enforce criminal laws intended to protect the functioning of our democracy against an attempted autogolpe are almost at their zenith. I have no disagreement with this analysis, except that it’s gilding the lily. All that needed to be said is that a president seeking to unconstitutionally remain in office after losing an election cannot possibly be understood as engaged in official acts for purposes of criminal law, even if they might be considered official acts for purposes of civil liability, and then a general principle that there is no criminal immunity for acts beyond core official duties. This would have handled the matter more succinctly while producing unambiguously clear precedent to cover the easy cases, while leaving the hard cases — as when a president takes actions that do look like core official acts but that violate criminal law — for any later cases in which those questions arise.

The panel then launches into a long consideration of Trump’s claim that, because the Impeachment Judgment Clause provides explicitly that a president who has been convicted in an impeachment trial remains subject to indictment and criminal trial, an implicit immunity to criminal indictment exists that can only be lifted by conviction by the Senate in an impeachment trial. The panel begins by noting that an argument of negative implications of a statement about a different situation is disfavored as a textual interpretive tool. Indeed, this is more like a logical fallacy (assuming the inverse) than a standard interpretation, as the panel notes while quoting Justice Scalia. That point is surely correct.

The panel then marshals historical evidence poorly and thoughtlessly to support its conclusion. It engages in the worst sort of originalism — trying to understand a sentence that is clearly comprehensible on its own by dredging up hoary old dictionary definitions of words that still mean the same thing they mean today. “Nevertheless” still means “nevertheless,” and “liable” has been a legal term in continuous, consistent use for centuries. The panel decides that that doesn’t demonstrate enough determination to dredge up original meanings, however, so it quotes Samuel Johnson’s dictionary — a roughly contemporaneous dictionary in wide use at the time of the Constitution’s ratification, although one assembled and published an ocean away and a generation earlier, and with a much more prescriptivist approach than modern lexicographers would take. The panel quotes Johnson’s definition of “nevertheless” as meaning “notwithstanding that,” as if that provides any additional enlightenment. (They include the syllabic accent mark in nevertheless, rendered as an apostrophe, which makes it seem garbled, and then includes it again in “notwithsta’nding that” (error in opinion at 44) even though it doesn’t appear in that phrase in the original. This is baffling and errant nonsense up with which I will not put, to quote a description of prescriptivism run amok usually but probably erroneously attributed to Winston Churchill.) As if this pointless bowing at the altar of originalist lexicography wasn’t bad enough, the panel then decides that to understand “nevertheless,” which is defined as “notwithstanding that,” it should look up a contemporaneous definition of “notwithstanding” to conclude that through some insane dictionary transitive property of equality “nevertheless” has the actual meaning of “notwithstanding”’s definition, “without hindrance or obstruction from.” For years, originalists have inflicted terrible law chambers history on us to conclude that, like Lewis Carroll’s Humpty Dumpty, the Constitution’s words mean whatever they choose it to mean. Here, they supplement their bad law office history with terrible law office linguistics, even though their digression adds nothing to the plain words of the Constitution. Some of the historical evidence the panel presents is at least somewhat persuasive, though unnecessary. But the entire exercise is revealed for the sham it is when they present a much better argument: the Impeachment Judgment Clause does not apply only to the President, but also to any other civil officer of the United States as well. Numerous federal officers have been tried for criminal offenses, petty or great, without having been impeached first — indeed, impeachment is an extraordinarily rare remedy, with only a handful impeachment trials ever conducted in this country and even fewer convictions, while criminal trial and punishment of corrupt executive branch and judicial officers of the United States is much more common. Moreover, Trump argued in his second impeachment trial, and many of the Senators who voted to acquit purported to agree, that a former officer is not subject to impeachment by the House and trial by the Senate. Under Trump’s implausible interpretation, a president is immune from trial for crimes committed in office unless impeached and convicted, and immune from impeachment unless in office, so once a president leaves office, the limited immunity of unless convicted in an impeachment trial becomes an absolute immunity, regardless of what facts subsequently come out. As the panel acknowledges, this is irrational and not the law.

As with some of the earlier sections, it’s not that the panel’s conclusion is wrong — it’s obviously correct — nor even that it doesn’t eventually present good arguments for its conclusion. It’s just that the panel surrounds its good and intelligent arguments with a host of stupid and incoherent paragraphs that accomplish nothing.

Rejecting Trump’s Claim of a Quasi-Double Jeopardy Immunity

At long last, the panel turns to the question of whether subjecting someone who was acquitted by an impeachment trial to criminal trial for charges connected to the same event would violate principles related to the Fifth Amendment’s prohibition on Double Jeopardy. Here, the panel finally handles something relatively well. It begins by noting that the Double Jeopardy Clause itself does not apply here; Double Jeopardy refers to being subject to a second application of the same criminal charge for the same offense, and does not apply when someone has previously faced civil or otherwise non-criminal liability. An impeachment proceeding is not a criminal charge, such that a following criminal trial for the same offense would violate Double Jeopardy. That’s made unambiguously clear by the Constitution’s limitations on the consequence of impeachment and conviction, by the express authorization of further criminal proceedings after an impeachment proceeding, and by the fact that the limited impeachment power in the Constitution is carefully separated from the prohibited bills of attainder that were associated with impeachment in British practice. Moreover, impeachment acquittals do not involve pure determinations of fact; an impeachment is inherently a political proceeding, and Senators can and do vote on whether to convict for political considerations entirely separate from their judgment of the factual guilt or innocence of the impeached party. Indeed, even in Trump’s own case, many of the Senators argued that they were voting to acquit for reasons that had nothing to do with whether he had in fact done what the impeachment counts alleged. Because this is not covered by the Double Jeopardy Clause, the panel then considered and summarily rejected the ungrounded assertion that there is a broader principle of Double Jeopardy that applies here even though it is not within the scope of the Double Jeopardy Clause itself. Overall, I think this section was well handled and ably drafted.

The panel opinion then included an additional section in which it applied the standards that would apply if the Double Jeopardy Clause were directly implicated, even though it was not. Under settled Double Jeopardy jurisprudence, criminal prosecution for one crime does not bar subsequent criminal prosecution for other crimes committed as part of the same sequence of actions. Instead, in order to represent Double Jeopardy, one of the two crimes must include all of the elements of the other. If someone has been acquitted of murder in the first degree, it represents double jeopardy to charge them with the lesser included offense of murder in the second degree or voluntary manslaughter. Likewise, acquittal of the lesser included offense bars prosecution for the offense that requires proving additional elements. However, if someone is acquitted of murder in the first degree after allegedly burning down a home to kill someone, they could still be charged with arson for the same alleged conduct. Murder in the first degree requires proving facts (starting with the death of the victim) that are not required to prove arson, and vice versa — arson generally requires proving that the premises burned were a dwelling, while murder doesn’t require proving anything about burning at all.

Applying those rules of Double Jeopardy jurisprudence to this case, the panel concludes that both the impeachment charges and the criminal charges each included elements that were not included in the other. Therefore, even if Double Jeopardy protections applied, Trump has no entitlement to have these charges dismissed.

This section was strictly speaking unnecessary, but I believe that the opinion is stronger because it is included. First, it helps reassure the Supreme Court that there is no need to revisit this case — if any of the Justices believe that subsequent prosecution after an impeachment acquittal might implicate the Double Jeopardy Clause, they can be reassured that there is no reason to reach that question because it would still not apply. Deciding cases in the alternative can be a useful approach for judicial economy in avoiding any need for further proceedings (or for a remand if a higher court disagrees with one aspect of a lower court’s decision but the result would not change). Second, it helps demonstrate that Trump’s whole argument here is groundless — it fails in multiple ways and is not actually an argument that needs to be taken seriously. In general, I think that when there are sufficiently clear alternate grounds for a decision that can be laid out without unnecessary verbosity, it makes sense for courts to explain both lines of reasoning. Here, the panel disposed of both prongs of its analysis of the Double Jeopardy issue in a crisp six-page section, the inclusion of which surely was not a major cause of the opinion’s delay — at least there’s one section of the opinion that I consider actually well crafted.

Conclusion

The substantive outcome of this decision was correct, and generally understood as obvious. However, the panel sat on the case far longer than it needed to while working on a lengthy opinion. Unfortunately, the result of that delay, beyond the harm it threatens to do to the future of this case and to the democratic process, was an opinion that was still terrible legal work-product even if it reached the right outcome.