Oral Argument Summary: In re: Michael Flynn

The U.S. Court of Appeals for the D.C. Circuit today, June 12, heard the case of Michael Flynn, the man who oh-so-briefly served as President Trump’s national security adviser at the beginning of the president’s term. The court was not hearing Flynn’s criminal appeal; though Flynn pleaded guilty to lying to the FBI, the government has moved to drop the case against him.

Nor was the court considering, at least not directly, the question of whether the government should be allowed to drop the Flynn case or whether District Court Judge Emmet G. Sullivan has authority to block the dismissal of the case under Rule 48(a) of the Federal Rules of Criminal Procedure. Judge Sullivan has not yet ruled on that question—though it came up repeatedly throughout the argument anyway.

No, the case before the D.C. Circuit’s remote hearing concerned whether the appeals court should preemptively order Sullivan to dismiss the case on the theory that he has no authority to do anything else. Flynn’s lawyer, Sidney Powell, has asked the court for a writ of mandamus to compel Sullivan to grant the government’s motion to dismiss, and the government is supporting Flynn. In other words, today’s case pitted both the prosecutor and defendant against—get this—the district court itself. The respondent in today’s case was none other than Sullivan, who is trying to defend his authority to hold a hearing on the motion before him and to seek the input of a court-appointed amicus whom he named to argue against what both prosecutor and defendant want him to do, and to weigh in on the appropriateness of contempt charges against Flynn. Sullivan’s hearing is scheduled for July 16.

Readers can listen to the full hearing here. Below is a summary of what happened.

Powell is up first, representing Flynn before a panel consisting of Judges Karen LeCraft Henderson, Robert Wilkins and Neomi Rao. Powell is not a particularly polished advocate and lapses frequently into rhetoric more suited for a Fox News television hit than for an appellate argument. But her point is clear enough.

The district judge’s appointment of an amicus and desire to hold a hearing on the dismissal motion, she argues, is an impermissible intrusion into the sole power of the executive branch to dismiss a case. In this case there is an “extensive and thoroughly documented case by government” in support of dismissal, based on “extraordinary and exculpatory evidence” on behalf of her client. This case thus presents the “quintessential case of mandamus” and the “judge has no authority to do anything further” in it. The parties have decided, and the government has quit, she says. The judge has no authority to look further into the motives of the government.

Judge Wilkins points out that in Rinaldi v. United States, the Supreme Court was reviewing denial of a Rule 48(a) motion, and the court did an “independent evaluation of the unusual circumstances.” If the Supreme Court thought that there was a role for the court to play in that case, doesn’t the district court have that same role here?

No, says Powell. The presumption of regularity applies to everything that the attorney general has done. There is nothing further for the court to do here. There is no indication or clear evidence to go behind the government’s determination to drop this case.

Wilkins pushes: The court said in Rinaldi that it wouldn’t presume bad faith, he points out. In other words, the court applied a presumption of regularity then too—but it still did an independent investigation.

Powell responds that Rinaldi was different, because there was no factual background investigation; no amicus was appointed. She then looks to United States v. Fokker Services B.V., a D.C. Circuit case from a few years back. Fokker, she says, stands for the proposition that there is no authority for the court to second-guess the prosecution’s decision.

Wilkins tries again: What if the district court had issued a minute order saying it would do an “independent investigation” and issue an opinion in due course. Would that be an error?

Powell says it would be. In Rinaldi, the court did not go further than assessing the government’s arguments in doing its investigation. The court can’t make the government pursue this criminal case if the executive branch doesn’t want to, she argues.

Wilkins goes on: In Thompson v. United States, he points out, the Supreme Court was reviewing the denial of a Rule 48(a) motion and it did an independent investigation of the record despite the solicitor general suggesting it should dismiss the case.

But all that means, Powell responds, is that the court reviewed the record before it and the documents that the government had provided. The court wasn’t trying to add on perjury or contempt charges there or do more investigating, as Sullivan is contemplating here.

Wilkins turns to the fact that Sullivan hasn’t really done anything yet. All he has done, after all, is appoint an amicus and schedule a hearing—and not yet granted the motion to dismiss. Is Powell really arguing that the judge not yet granting the motion is sufficient grounds to justify mandamus?

She is, and she cites Fokker in support of the notion.

But Fokker, Wilkins responds, wasn’t even a Rule 48(a) case; rather, it concerned a deferred prosecution agreement. Wouldn’t interpreting Fokker to require dismissal without examination undermine what the Supreme Court itself did in Thompson—that is, conduct an independent investigation when there was no claim of bad faith?

Powell again stresses that she doesn’t object to an independent review of the record, provided that it is just that: a review of the existing record. And the record here is well documented in showing prosecutorial misconduct and suppression of Brady evidence, she insists. The judge does not have the authority to appoint an amicus. Of course he can look at the materials before him, but that’s really all he can do.

Wilkins turns to whether mandamus is the appropriate mechanism to examine this question. It seems, he says, that Flynn has an alternative avenue of review. If the judge denies the motion to dismiss, he can just come back here on appeal, right?

Powell does not want to wait. We would simply be delaying the inevitable in doing that, she says. Sullivan doesn’t have the authority to do this. The briefs Flynn will have to respond to if Sullivan is allowed to go forward are long and have a lot of exhibits. And such a process takes a huge toll on the defendant. “This is the most impressive motion to dismiss I’ve ever seen in decades of practice,” she says, and the most well documented. There’s nothing more to talk about.

Judge Henderson steps in: If Sullivan had just kept this motion waiting and languishing, that would be one thing, she suggests. But he has set a hearing for mid-July. For all we know, he may just grant the motion to dismiss then. Isn’t this discussion premature? This is the first of several moments in the hearing in which Henderson wonders if mandamus may be the wrong remedy and if the process should be allowed to play out in the regular order.

But Powell gives no ground. Sullivan doesn’t have the authority to conduct that hearing, she insists. He has appointed this amicus to go well beyond the current record and look into the government’s motives for dismissing the case.

Wilkins responds that the courts have said the judge is not a rubber stamp in the context of a Rule 48 motion. What is the authority that suggests that Sullivan can’t even hold a hearing?

Powell responds with something of a rant about how Sullivan is going to drag this case out for another six months. It’s not going to be just another hearing in mid-July. The judge clearly wants to give Flynn the maximum sentence, she insists, and by dragging this out, the “government is just wasting resources out the wazoo on this.”

Why can’t we hold this matter in abeyance and see whether that happens, asks Henderson.

Because the damage continues to accrue as the case drags on, Powell says. The judge has no jurisdiction over the matter, as there is no case or controversy before him. He didn’t have authority to appoint an amicus, she argues.

Wilkins stops her with a hypothetical about evidence gathered from a drone looking in a bedroom window. If amici wanted to participate in support of a motion to dismiss in that case, he asks, are you saying there is no authority for the district court to grant the motions of amici to file briefs?

Powell responds that there is no authority to grant amici permission to do the job that the government would have done had it chosen to continue the prosecution. If the court wants to hear from an amicus in support of one side or another, that’s a different thing. But the court can’t appoint someone to take the place of the attorney general or to act like the executive branch itself. The judge can’t go outside of his lane to appoint someone to do that.

Wilkins isn’t buying it: The order appointing the amicus was for him to present arguments in opposition to the government's motion to dismiss, he says. How is that encroaching on Article II?

Because the government had already made the decision to stop the case, says Powell, and the government is the only one that can make that decision. “The government has quit and it’s time to leave the field,” she argues.

What about appointing an amicus for the contempt charges, Wilkins asks. The Supreme Court said in Young v. United States ex rel. Vuitton et Fils that the court can appoint an amicus to prosecute contempt charges. Would Powell’s argument against appointing amici apply there?

Powell responds that contempt is not available for perjury in these cases.

Judge Rao speaks up for the first time: But that goes to the merits of whether contempt would be appropriate, she says. What about the question of appointing an amicus?

There is no basis for that, Powell insists. It’s not within the judge’s power to investigate.

Rao reiterates Wilkins’s earlier question: Isn’t that inconsistent with Young, where the Supreme Court said a private party could be appointed?

Powell doesn’t seem to have a general answer to that question, so she goes narrow: In the circumstances of this case, she says, contempt is not available.

Rao then turns to the jurisdictional question. With respect to Powell’s argument that there is no case or controversy, she points out, in Thompson, the solicitor general suggested that the case be dismissed under Rule 48(a). But the Supreme Court did not say there was no case or controversy. Instead, it remanded the case to the court of appeals for reconsideration in light of the government’s current position. In other words, the Supreme Court didn’t take the fact that the government had filed a Rule 48(a) motion as ending the case and thus leaving no case or controversy.

Powell’s response to this question is a bit of a ramble. Thompson was a case of mandamus for mandamus, she says. And the Supreme Court decided that the appellate court needed to address an additional issue. That’s not the situation here. No appellate court in the country has refused to affirm a Rule 48(a) motion. The government is the only entity that can prosecute this case, and it has declined to do so for a number of reasons. There was suppression of exculpatory evidence; the FBI agents made up statements in their 302s; there were discussions about trying to get the defendant to lie; the prosecution wasn’t going to show Flynn the evidence; it didn’t warn him about Section 1001. (There is little evidence for many of these assertions.) It is a “travesty of justice” that Flynn has been dragged through this for three years on a case that was “absolutely concocted by FBI agents.” The only thing left to do is for the judge to order it dismissed. Rao’s question about Thompson gets lost quickly in Powell’s litany of FBI villainy.

Henderson raises an additional question about Powell’s contention that it was improper for Sullivan to appoint an amicus. In Fokker, she points out, the D.C. Circuit appointed an amicus. Was this improper too?

Powell clarifies that the court can appoint an amicus as long as that amicus is addressing an issue within the case, not just adding in charges the judge wants to investigate. What the D.C. Circuit did in Fokker is the same as what Beth Wilkinson, the lawyer representing Sullivan, is doing in this case. Appellate courts often allow amicus participation. But district courts don’t do it in criminal cases. There’s no provision for that in a criminal case in a district court, she argues.

Wilkins jumps back in, noting that he asked earlier if an amicus could file a brief in support of a motion to suppress. What’s your answer?

Powell acknowledges that there is authority for that.

Based on what rule, Wilkins asks. Powell responds that she doesn’t think there’s a rule for it, but she believes there have been amici in other cases in support of a position of one of the parties, just for a judge trying to gin up charges himself.

So with respect to a Rule 48(a) motion, Wilkins probes, if for some reason Flynn opposed the government’s motion, it would be okay for the court to appoint an amicus to give briefs in support of Flynn?

This would probably be okay, says Powell, because there would still be a case or controversy. It would be up to the district court to do that and there’s not a rule, but it might be okay. But Sullivan essentially appointed his amicus—former U.S. district judge John Gleeson—as what amounts to a special prosecutor, which he doesn’t have authority to do. If he had asked Gleeson to weigh in on a side that existed in the case, that might be permissible. I might argue against it, Powell says, but I wouldn’t have sought a writ of mandamus.

Rao clarifies: Isn’t your argument also that it’s impermissible to have amicus briefs in criminal cases under the rules of the court?

There’s no rule allowing this, Powell responds, but she has seen it done as long as it is in support of one of the parties. It is not okay when the judge is acting as an independent prosecutor.

And with that, Powell concludes.

Principal Deputy Solicitor General Jeffrey Wall takes the virtual podium for the government. Wall is a dramatically better oral advocate than Powell. He is also taking a somewhat more modest position; he is not generally challenging, for example, the district court’s authority to appoint amici in criminal matters.

In recent mandamus cases, he says, the court looked to whether there was a clear and indisputable case on the merits, and then it has turned to the mandamus factors. If the government is right that Fokker applies to an unopposed Rule 48(a) motion—and thus controls Sullivan’s consideration of the Flynn motion—then it becomes evident why it is so harmful to let this case proceed, he argues.

On the merits, Wall argues as follows: First, he cites precedent for the notion that Rule 48(a) applies at all stages of a case. No matter when the government wants to drop a case, it gets to do so. Second, he relies on Fokker for the substantive standard that requires Sullivan to dismiss the Flynn case.

Wilkins pushes him on this latter claim. Fokker, after all, was not a Rule 48(a) case. It concerned a deferred prosecution agreement.

Wall acknowledges the point but says that Chief Judge Sri Srinivasan’s opinion in Fokker said that the deferred prosecution agreement and a Rule 48(a) motion are analogous. That was central to the court’s reasoning, he argues. So Fokker is suggesting that dismissing is clearly within prosecutorial discretion, and there is no oversight power in the judiciary and no substantial role for courts to second-guess the dropping of a case. If the government is right on this point, he goes on, then it is a question of what the point of further proceedings is if the district court is required to dismiss the case.

Do you believe that Fokker stands for the proposition that the district court can’t do an independent investigation of the record, asks Wilkins, invoking again the language of the Supreme Court?

Wall responds that he understands Fokker to mean that if, tomorrow, the district court said it was going to set up a process for deciding whether to grant this deferred prosecution agreement, and the district court did all of that, it would be a straightforward violation of Fokker. The harm is not just that there will be some briefs; the point of Sullivan’s appointment of an amicus is basically to investigate prosecutorial motives.

Wilkins responds that the order under review is limited: that the amicus is to present arguments in response to the government’s motion to dismiss, and there’s going to be a hearing. It doesn’t say anything about factual development or anything else.

But Wall disputes this point. We are here on mandamus, he says, not on review of any order by the district court. Rather, Powell and the Justice Department are asking that the district court be directed to grant a motion. The respondent—which is to say the judge—wants to inquire into prosecutorial motives and opinions. His amicus has flat out alleged that there was misconduct by the attorney general on behalf of the president. We’re going to have to respond to these things in proceedings going forward, Wall says, and those are the systemic costs that this court laid out in Fokker. To prevent this, this court would at least have to issue a limited mandamus and take off the table evidentiary motions and contempt and just permit Sullivan to consider the 48(a) motion.

So, Wilkins asks, the district court doesn’t even have authority to appoint an amicus to advise it on whether it should issue an order on contempt?

Wall stresses that the government, unlike Powell, has not generally challenged the district court’s authority to appoint an amicus. But this amicus appointment was improper, and it would be improper for the court to proceed under the same reasoning that Fokker had about deferred prosecution agreements. This is not a general argument, but there were problems here with this particular appointment of an amicus. Once these harms are identified, there is no reason not to take this final step of issuing a writ of mandamus. This case has already become a public spectacle, and it threatens to harm not just the executive but the judiciary as well, Wall says.

Rao enters the fray: We have repeatedly declined to grant mandamus when the court is addressing abstract separation of powers violations she says. Can you be more concrete about what the violation is? What precisely is the infringement here?

Separation of powers harms can be subtle, Wall says, but “they are as stark and concrete here as they come.” What the district court is contemplating is a fact-intensive inquiry into a variety of questions, for example the uncharged conduct on Turkey statements. It is an intrusive process that will harm the executive, and it is playing out in a politicized context. There was a “70 page almost polemic” that the court-appointed amicus filed—referencing the brief that Gleeson filed recently.

Wilkins comes back to his original point: But all of the Rule 48(a) opinions from every court have said that the court has some role—in that the court needs to make sure there’s not something being done that is clearly contrary to the public interest. There must be some case where even if it’s unfortunate that there is a clash between the two branches, the “leave of court” language in Rule 48(a) allows the court to step in.

Wall responds that “leave of court” can do some work where the court has an opposed motion. And even where a motion is unopposed, the court can ascertain whether it reflects the considered position of the parties or, for example, whether the prosecutor was bribed. But when the parties agree and the judge has their considered opinions, the court is required to grant the motion in light of the constitutional requirements the court discussed in Fokker.

Moreover, Wall goes on, everyone accepts this prior point before a plea agreement is reached. Sullivan is trying to argue that Fokker is only about the pre-plea situation, not the situation after the plea is accepted. That can’t be right for several reasons: The constitutional issues are the same at any stage, Wall says. Indeed, the United States can dismiss even after judgment or trial; there simply is no magical line that is drawn at the plea. What’s more, Fokker rejected this distinction. Put simply, dismissing is not a judicial act. It is in deference to the executive’s exercise of prosecutorial discretion. Once the court accepts this, then the merits collapse and the appeals court is back to the question of whether mandamus is appropriate.

But Fokker also made clear, Wilkins responds, that there are different considerations at different stages of the case, even if the court accepts that Fokker applies. Here, there are two different district court judges who made findings with respect to materiality, Wilkins says—referencing the fact that Flynn’s case was before Judge Rudolph Contreras before it was transferred to Sullivan. And so the government’s motion doesn’t just implicate the government’s decision. It also implicates those rulings by judges as well. The case isn’t in the same posture as it would be prior to a plea agreement in that respect.

Wall agrees that there are different considerations at different stages of a case. But he disagrees that the plea is a magic line, after which everything that follows changes the court’s role. Fokker says that it is not true, he says. And in In re United States, there was a sentence already, he notes, and the government wanted to dismiss some of the charges because it no longer wanted to proceed with them. The district court wouldn’t allow it. The U.S. Court of Appeals for the Seventh Circuit said that it didn’t matter if the prosecutor was trying to get around the court’s sentencing authority, because the government is the master of its own case. A Rule 11 plea agreement would be different if that was what was before the court—and Fokker is clear about that, Wall acknowledges—but when Sullivan grants this motion, he is just acknowledging a co-equal branch’s exercise of its power. It’s not a formal judicial action signing off on something.

Wilkins now turns to a different angle: Is it your position, he asks, that the government doesn’t have to state all of its reasons in support of dismissing a case but, rather, only those that it chooses to share with the court?

Wall acknowledges that this is the government’s position, but he says that he doesn’t think anything turns on the point. The government didn’t have to give any reasons, and district courts routinely grant motions to dismiss based on nothing more substantial than a one-page motion to dismiss with no background material or argumentation. In this case, he says, the government went beyond what it was obligated to do in terms of giving reasons.

Wilkins offers a hypothetical to highlight his concerns about this position: Imagine a case in which a white federal law enforcement officer pleaded guilty to criminal civil rights charges for using excessive force against a black victim, he says—no doubt thinking of recent news. Now imagine that the government says that it uncovered Brady evidence and is moving to dismiss under Rule 48(a). But what if part of the reason that the government thought it couldn’t try the case was that it didn’t think the jury would believe the black victim over the white officer without corroborating evidence—yet didn’t include that in its motion to dismiss because it knows that argument wouldn’t play well with the judge? So it just said instead that the exculpatory evidence was its reason to dismiss. Would that be proper?

Wall responds that Fokker is clear that whether the government provides some reason or no reason at all—as long as it is not an unconstitutional reason—then it can dismiss. But if the court disagrees that this point is clear, Wall says, it can try to leave that question open and say that whatever the standard is on the merits, the government has met it here through its fulsome explanation of its decision. I hope the government hasn't ever filed a brief like that, he says. But if it has, then the court should still have to grant the motion because the government no longer wants to proceed with the case.

In any event, Wall says, if the government did such a thing, the remedy is that other defendants would walk into court and make claims under United States v. Armstrong, saying that the government is making racially biased prosecutions, and that would be a strong case. So there are remedies available other than distorting Rule 48(a).

But you can’t have remedies if you don’t know it’s happening, Wilkins responds. And you won’t know it’s happening if the government isn’t giving all its reasons and the district court isn’t required to ask for reasons.

But that would be equally true in Fokker, Wall says. Any court might think that a U.S. attorney cut some sweetheart deal with a defendant, and the point of Fokker is that it is not for the courts to police whether the prosecutors have the right motives. There are other ways to do that in the political arena—through dismissal of corrupt officials or even through impeachment. If the court could do the kind of thing Wilkins is talking about, then Fokker is a dead letter because a court could always inquire into a deferred prosecution agreement or a Rule 48(a). And that’s contrary to Fokker.

Rao asks about the presumption of regularity. What about when a district court thinks that the presumption is overcome on the face of evidence presented by the government?

Wall responds that even if the district court thinks that the prosecutor had a bad-faith motive in declining to bring a case, no one can force the prosecutor to bring it. The same logic applies for maintaining a case. Courts are very unlikely to second-guess a prosecutor’s decision in the absence of an unconstitutional motive. So the government would thus still be entitled to mandamus. There must be clear evidence of an unconstitutional motive to rebut the presumption under Armstrong, Wall says. If the considered decision of the executive branch, no matter its motives, is not to proceed, then there’s no case or controversy.

Even if the prosecutor is dismissing the case because he didn’t think that a white officer should have to answer for a crime against a black citizen, asks Wilkins?

Wall responds: I don’t think that the court can force the executive to keep a case alive in the absence of case or controversy. Even if you disagree with me, the reason that hypothetical has force is because it is an unconstitutional motive. And there is nothing like that here.

But then, Wilkins asks, what work at all does the “leave of court” language in Rule 48(a) do?

Wall says that it does work with an opposed motion to dismiss. And with unopposed motions, the rule can be used to make sure that the decision is the considered position of the executive and the defendant, meaning that they weren’t bribed or poorly counseled.

Wilkins returns to his hypothetical: What if the government makes a considered but racist decision? Why isn’t it the case that the district court can deny the motion and let the political chips fall where they may? The district court can’t itself force the government to prosecute, but maybe other pressures or operations of the legislative branch can lead to the appointment of a new prosecutor and the case proceeds. Isn’t that what “leave of court” is intended to do?

Wall responds that the courts have no power to make the executive move forward to trial if it doesn’t want to.

But while the case is in limbo, Wilkins wonders, pressure could be put on the government to bring the case, right?

Wall reiterates that the remedy in this situation would be to dismiss other cases, but not to refuse to dismiss this case in particular. If the government had clear evidence of unconstitutional motive, there could then be an Armstrong claim, Wall says, but we don’t have that here. No one is arguing that there is an unconstitutional motive on the face of the motion to dismiss that the government filed here.

Henderson returns to her theme that the matter is premature. Going ahead and mandating the granting of the motion to dismiss would harm the regular order, she says. Mandamus is a drastic remedy. Unless Wall thinks that the setting of a hearing is ultra vires, that leaves the appointment of the amicus within the judge’s power. Granted, Sullivan may have chosen an intemperate amicus. But that doesn’t mean he is necessarily going to deny this motion. So, Henderson says, there is no precedent that allows the appeals court to move without an order, and given how drastic mandamus is, why should it not observe the regular order and allow Sullivan to rule first?

Wall acknowledges that mandamus is an extraordinary writ, but he says that this is an extraordinary case—a point with which Henderson agrees. We’re well past the regular order, he argues. If the goal is just to get back to the regular order, this court should issue a more limited mandamus that takes certain things off the table that probe the government’s motives. Respondents seem to want some kind of evidentiary or discovery process, and that’s too much. At a minimum, the government needs the court to decide this with dispatch. But even if Sullivan grants the Rule 48(a) motion eventually, it’s not the case that there is no harm. There are harms to Flynn and to the government. There are lots of allegations being lobbed at the executive branch, and the government is going to have to answer those claims in a public forum. This chills law enforcement, and it invades the government’s deliberative process.

Henderson does not seem convinced. Sullivan is “an old hand, an excellent trial judge,” she says. He may ignore the advice of the amicus and grant the motion to dismiss. Why shouldn’t he be allowed to take that route if he wants to?

Because we are past regular order, Wall says; we are now over the mandamus threshold. We have briefs raising questions about motives and the amicus raising questions and making allegations, and the court is apparently contemplating that we will defend ourselves. This is going to become the kind of public spectacle that mandamus is intended to foreclose. These unnecessary proceedings are going to do damage to the executive branch.

Wilkins, at this point, raises an uncomfortable point: It was Flynn, not the government, who filed the petition for the writ of mandamus. Are the government’s claimed harms even at issue here?

Wall, in response, tries to merge the Flynn petition with the government’s concerns. There was uncertainty in the district court about what the government was going to do, he says vaguely. Flynn filed right away with his mandamus petition, and the government had to decide whether to support his mandamus petition or to file a duplicative one of its own. The government decided to support Flynn’s. If the court really cares about this distinction, the government should be given the opportunity to file an additional mandamus petition as there are “serious, indeed grave” separation of powers concerns here.

Wilkins points out that it’s far from regular order to give the government leave to file a mandamus petition here. Wall is making an argument about harms to the government even though the court doesn’t have that petition before it, Wilkins says.

Wall seems a bit thrown by this argument, but he recovers nimbly. The fact that we didn't file isn’t legally relevant in any way, he says. We support other people’s motions all the time and were invited to give our views in this case as we are in the Supreme Court all the time. To get back to regular order, we need mandamus to take certain things off of the table.

After a brief exchange with Henderson about her concern that the case should wait until Sullivan actually rules, Rao raises a different issue. She’s concerned about this government fallback position that the court should grant some kind of more limited mandamus relief. Wouldn’t that require the court to articulate more legal standards about what is on and off the table? It seems much less clean than issuing a writ of mandamus in full. Do you really think partial would be better?

Wall reassures her that he does not prefer partial relief. The cleanest way to resolve the matter under Fokker is to grant the writ, he says. But if the court has these concerns, it could grant a more limited mandamus.

Henderson once more returns to what is fast becoming her hobbyhorse. You still have to say why there is no adequate remedy at law, she chides. Why is there none available if Sullivan can just grant the motion to dismiss on July 16?

Wall concludes by arguing that even if a month or two from now the court grants the Rule 48(a) motion, there will be a proceeding that forces the government to explain itself. The district court is very careful not to say what it envisions in this proceeding, but it has left room for documents, witnesses and other production. This will cause all the harms to the executive branch that are laid out in Part 2(a) of Fokker. Impugning the motives of the attorney general is a real harm to the executive branch—even if after that spectacle, the district court grants the motion. If ever the court of appeals were going to say that a district court has to grant the motion, the Flynn case would be the case that justifies the move.

And with that, Wall turns it over to Beth Wilkinson, who argues the case for Sullivan.

Wilkinson begins by noting that Flynn is asking the court to grant an extraordinary remedy. She argues that the writ should be denied for three reasons: First, because the government’s motion is still pending and could be granted; second, because the law does not clearly and indisputably foreclose the district court’s consideration of the government’s motion; and third, because it is inappropriate to grant mandamus in a case with open questions where the government is raising novel constitutional questions that were not raised below. No one is saying that the district court can second-guess prosecutors because it disagrees with their decision not to pursue a case, she says. The issue here is whether a federal district court judge can set an expedited briefing schedule and appoint an amicus to give adversarial arguments before ruling.

Rao jumps in: In a case like this, isn’t the appointment of an amicus creating a case or controversy where there isn't one?

Wilkinson argues that there is clearly a case or controversy here. The government and respondent are both asking for the case to be dismissed with prejudice. By definition, there cannot be a ruling to dismiss with prejudice if the court doesn’t have jurisdiction.

Rao pursues the point: But if the district court were to determine that the motion to dismiss should be denied, then what happens? Does the court go on to sentence Flynn even though the executive is no longer pressing its prosecution?

The question is premature, Wilkinson contends. There is no reason to fear at this point that the district court is going to deny the motion to dismiss. But if the answers at the hearing give some basis to do that, then the parties can file another motion for mandamus at that point. This happened in Aiken County. In that case, the court denied the first writ because the government hadn’t had the chance to act—and that gave the government years of time. The court granted the writ only when the agency in question said it wouldn’t rule on the relevant issue. In this case, by contrast, a hearing time has been set and there is no suggestion that the court is going to engage in this parade of horribles that has been presented by the respondent and government in terms of asking for a ton of information. All the court is doing is asking for advice.

Rao turns to another subject: What standard is an Article III judge supposed to apply in this context to assess the motion to dismiss? Rule 48(a) just says there must be “leave of court.”

The governing law is United States v. Ammidown, Wilkinson says, which is still good law and was cited in Fokker. Under Ammidown, the court must consider whether the presumption of regularity may stand or whether there was a clear violation of the public interest for the court to seriously consider whether it can deny a motion to dismiss. That will guide the court’s inquiry after receiving the briefs and the answers of the parties.

Rao is unsatisfied. The public interest is not a standard that is mentioned in the rules, she says. And in the U.S. system, isn’t the public interest one that is committed firmly under Article II to the politically accountable executive branch, not to an Article III court?

Generally yes, says Wilkinson, but in Rinaldi the Supreme Court specifically held out that standard in Footnote 15 in declining to adopt an abuse of discretion standard; the court said it had not decided whether a district court could consider a Rule 48(a) motion in light of the public interest. In fact, the dissent led by Justice William Rehnquist said it was clearly an independent standard. In any event, that law is not clear, which means there is no clear and indisputable standard for this court to issue a writ of mandamus.

It is unclear as to how to determine whether there is leave of court, Wilkinson continues. But as to Fokker and Ammidown, both allowed for a review by the court of the government’s motion and allowed for questioning of the parties, including the government, in open court. Judge Richard Leon questioned the parties in Fokker and when it was decided, no one said in Fokker that he was wrong to have a hearing or to pose questions.

But the government, Rao argues, contends that Fokker necessarily rebuked that approach by the district court.

That’s not what Fokker says, Wilkinson responds. Fokker says that the district court judge abused his discretion when he stated that he disagreed with the government’s basis for its prosecutorial decisions. Fokker does not deal with Rule 48(a). It certainly doesn’t say that the court can’t engage in consideration or scrutiny or that the court has no right to ask questions. Fokker talks about scrutiny and the court’s power being circumscribed, but it doesn’t say there is no right to ask questions. All Sullivan is doing is receiving briefing and asking questions.

Rao takes another turn. How does the presumption of regularity apply in a situation like this? Before asking questions and appointing an amicus, doesn’t the court have to determine that the presumption of regularity was overcome?

Wilkinson turns this question around. The court can’t determine whether the presumption is overcome without at least questioning the government about its motion, she responds. The court has been clear that there is a presumption of regularity, so it is a long hill to climb to overcome that presumption, but there is nothing in Fokker that says the judge can’t question the submission. The government answers these questions all the time. In fact, in Rinaldi, the judge in Florida called the parties in and asked questions, and no one said that was improper.

But there is more here, Rao pushes: Appointing an amicus to oppose the motion to dismiss is not an everyday occurrence.

Wilkinson agrees but notes that, normally, the parties are opposed. In this unusual circumstance, both the government and the defendant agree—so all the district court did was appoint an amicus to respond to the government’s motion to dismiss. And the amicus filed that brief without doing any fact finding or questioning. It may well be that Sullivan will read both sides, ask some questions, and then grant the motion to dismiss.

Who is the amicus representing here, asks Rao? Who is he arguing on behalf of?

Wilkinson responds that the amicus is arguing on behalf of the adversarial position, just like this court and the Supreme Court often have appointed amici to do.

Rao asks: Is there some kind of judicial right or power that the amicus is representing?

There is an inherent power for the district court to appoint an amicus, and here it is important because the court needs adversarial briefing, Wilkinson responds.

Rao is back on her jurisdictional point: In a criminal case, the adversarial process requires that the government advocate against a criminal defendant. What does it mean to have an adversary when both sides agree? Is this coming from some kind of inherent authority of the court?

Wilkinson contends that it’s the court’s inherent authority to understand what the law is. Just like Wilkinson is doing here, the district court brought someone in to represent the other side. This doesn’t mean that the district court thinks the motion should be denied. The district court judge appointed an amicus to understand what his authority looks like.

But then, Rao says, the amicus isn’t an adversary; he is just providing information about the law to the district court, right?

He is an adversary, Wilkinson says, in that he is intended to take positions opposing the government’s motion to dismiss. The court can ignore the amicus, or it can consider some things he says, ask him questions and then make its ruling.

Rao is not done. There is a core Article II power over prosecution, she says. Even the weakest understandings of Article II include control of the prosecution. So can Wilkinson articulate the countervailing Article III issue at play? In United States v. Nixon, there was a conflict between two branches of government, and there was an Article III interest in protecting independent liberty in a criminal trial. What is the countervailing Article III interest here?

Wilkinson does not seem to want to engage on the basis of high theory. Fokker said that courts are allowed to question the presumption of regularity even when the parties agree, she says. Even Nixon recognized that the Framers didn’t intend for separate powers to operate with absolute independence. As Fokker says, the court has exercised its coercive power without a trial and has thus stepped into the shoes of the jury—and that’s important for Article III power. And now the government is asking the court to vacate those charges, and it is unclear what they want the court to do about the actual plea and what the effect is. The executive branch is asking the court to vacate its findings. The court has the right, duty and language under Rule 48(a)’s reference to “leave of court” to ask the government questions. No case says that the “leave of court” language is meaningless when the parties agree. That doesn’t make sense under any of the cases starting with Rinaldi. And every panel has at least considered whether the court below abused its discretion in denying the motion to dismiss—thus acknowledging that it had some discretion.

Okay, Rao concedes, perhaps “leave of court” may not be a rubber stamp. But it’s hard to fit into it the ability of the court to keep sentencing a defendant when the government wants to drop the case. “Leave of court” cannot mean that much, even if it has some power.

It’s a circumscribed review and a limited inquiry, Wilkinson responds, but it’s not no review and no questions. It’s not an irreparable harm to ask questions—and if it is, that harm went on in Fokker and Rinaldi and is happening in district courts all the time. Normally, the parties answer the questions—and the government knows how to not answer questions if it thinks they’re not appropriate. Here, though, Powell and the government are end-running the district court and coming to the appeals court.

Wilkins comes back to his race hypothetical. What if the government says we are moving to dismiss because we don’t want to prosecute a white defendant? Then would the court still be compelled to grant the motion to dismiss and then have the remedy be for defendants in other cases to file an Armstrong motion? What’s Wilkinson’s response to the government’s view on this?

Wilkinson takes strong issue with Wall on this point. Case law starting with Rinaldi gives examples of when a court can deny a motion, she says, including in cases of bribery. That’s not a constitutional violation. It also wouldn’t be in the pleadings if the prosecutor was bribed. It would have to come out through questions. Yet it could be the basis for a court to deny a motion to dismiss. Similarly, if racism were uncovered through questioning, that could be a basis to refuse to dismiss too. What could the court do then? If it was an early decision in the case and the defendant hadn’t pleaded guilty, there would be more limited options, but there is still public pressure. But if the defendant has already pleaded guilty, then the Article III court has supervisory power over the defendant, as Fokker says. So if that police officer had pleaded guilty and the government moved to dismissal afterward, could the court sentence the defendant? That’s a hard question, but there would certainly be a basis for the judge to deny the motion to dismiss.

There is a long pause while Wilkinson waits for further questions. Hearing none, she concludes: The district court is not acting as a prosecutor. The judge is hearing all the arguments and will do what he is called on to do on a daily basis: rule on a motion. There is no justification for the court to enter the fray now and prevent an Article III judge from carrying out his responsibilities.

Henderson, who is presiding, gives both Powell and Wall two minutes for rebuttal. Powell uses hers for a rant. There were no valid proceedings in this case to take Flynn’s guilty plea, she says—a point that is almost wholly irrelevant. The initial judge should have recused immediately, not several days after the plea. And the second guilty plea colloquy with the new judge was not a colloquy at all. Everything is now refuted by the extraordinary exculpatory evidence that was hidden from Flynn for three years. (Again, there is little evidence for this.) The court doesn’t have the authority to question why the government is acting as it is. There is no case or controversy here, and the government can’t be forced to carry out a prosecution. The motion for writ of mandamus should be granted “on all counts.”

Wall’s rebuttal is better focused. The respondents have really hung their hat on the idea that the mandamus motion is too early in time, he says—in what is clearly an effort to address Henderson. There are real harms that will come from the kinds of questions they want to ask. It seems like they want to ask pretty extensive questions, and the manner in which the government would have to answer those questions—in terms of showing its deliberative processes—would be intrusive. It’s pretty clear why Fokker says that the courts aren’t supposed to go down this road. If everyone knows what has to happen, then the district court should just be directed to do it.

And all of a sudden, Wall seems to have made headway. Henderson, whose prior questions were mostly on this timeliness point, is suddenly speaking his language. Your briefs talk about the benefit of self-correction for Article II, she says. If there was bad faith in this case, it occurred in the original prosecution. So shouldn’t we allow the Article II branch to self-correct?

Having been thrown a grapefruit-sized softball to close out arguments, Wall takes a big swing. Absolutely, he says. We gave a fairly full explanation of the government’s reasons even though we didn’t need to for seeking dismissal. And while two of those reasons were challenged on a legal basis, no one has said a word about the third: that the attorney general has determined that, given the circumstances of the FBI’s interview of Flynn, it is not in the interest of justice to proceed with the case. Nobody has questioned that, because it is in the core of the Article II power to decide such matters.

And with that, the case is submitted.

Counting votes in this case is a little bit tricky. It’s pretty clear that Wilkins has no sympathy for Flynn’s position or the government’s. It’s almost as clear that Rao is hung up on whether a case or controversy remains—and might therefore be expected to act aggressively in granting mandamus. Henderson, however, clearly has warring instincts. At a few points, she seemed to make clear that she believes the case must be dismissed. And she showed her hand most clearly on the merits at the end. But she also repeatedly raised the procedural question of whether Sullivan should be allowed to act before the D.C. Circuit gets involved. Look for her perhaps to try to reconcile these instincts with an opinion that makes clear where the case should end up but gives Sullivan the chance to get there on his own.