VI. Judicial Misconduct

Canon 3A(6) of the Code of Conduct for United States Judges requires federal judges to "avoid public comment on the merits of [ ] pending or impending" cases. Canon 2 tells judges to "avoid impropriety and the appearance of impropriety in all activities," on the bench and off. Canon 3A(4) forbids judges to initiate or consider ex parte communications on the merits of pending or impending proceedings. Section 455(a) of the Judicial Code requires judges to recuse themselves when their "impartiality might reasonably be questioned." 28 U.S.C. s 455(a).

All indications are that the District Judge violated each of these ethical precepts by talking about the case with reporters. The violations were deliberate, repeated, egregious, and flagrant. The only serious question is what consequences should follow. Microsoft urges us to disqualify the District Judge, vacate the judgment in its entirety and toss out the findings of fact, and remand for a new trial before a different District Judge. At the other extreme, plaintiffs ask us to do nothing. We agree with neither position.

A. The District Judge's Communications with the Press

Immediately after the District Judge entered final judgment on June 7, 2000, accounts of interviews with him began appearing in the press. Some of the interviews were held after he entered final judgment. See Peter Spiegel, Microsoft Judge Defends Post-trial Comments, Fin. Times (London), Oct. 7, 2000, at 4; John R. Wilke, For Antitrust Judge, Trust, or Lack of It, Really Was the Issue--In an Interview, Jackson Says Microsoft Did the Damage to Its Credibility in Court, Wall St. J., June 8, 2000, at A1. The District Judge also aired his views about the case to larger audiences, giving speeches at a college and at an antitrust seminar. See James V. Grimaldi, Microsoft Judge Says Ruling at Risk; Every Trial Decision Called 'Vulnerable', Wash. Post, Sept. 29, 2000, at E1; Alison Schmauch, Microsoft Judge Shares Experiences, The Dartmouth Online, Oct. 3, 2000.

From the published accounts, it is apparent that the Judge also had been giving secret interviews to select reporters before entering final judgment--in some instances long before. The earliest interviews we know of began in September 1999, shortly after the parties finished presenting evidence but two months before the court issued its Findings of Fact. See Joel Brinkley & Steve Lohr, U.S. vs. Microsoft: Pursuing a Giant; Retracing the Missteps in the Microsoft Defense, N.Y. Times, June 9, 2000, at A1. Interviews with reporters from the New York Times and Ken Auletta, another reporter who later wrote a book on the Microsoft case, continued throughout late 1999 and the first half of 2000, during which time the Judge issued his Findings of Fact, Conclusions of Law, and Final Judgment. See id.; Ken Auletta, Final Offer, The New Yorker, Jan. 15, 2001, at 40. The Judge "embargoed" these interviews; that is, he insisted that the fact and content of the interviews remain secret until he issued the Final Judgment.

Before we recount the statements attributed to the District Judge, we need to say a few words about the state of the record. All we have are the published accounts and what the reporters say the Judge said. Those accounts were not admitted in evidence. They may be hearsay. See Fed. R. Evid. 801(c); Metro. Council of NAACP Branches v. FCC, 46 F.3d 1154, 1165 (D.C. Cir. 1995) ("We seriously question whether a New York Times article is admissible evidence of the truthfulness of its contents.").

We are of course concerned about granting a request to disqualify a federal judge when the material supporting it has not been admitted in evidence. Disqualification is never taken lightly. In the wrong hands, a disqualification motion is a procedural weapon to harass opponents and delay proceedings. If supported only by rumor, speculation, or innuendo, it is also a means to tarnish the reputation of a federal judge.

But the circumstances of this case are most unusual. By placing an embargo on the interviews, the District Judge ensured that the full extent of his actions would not be revealed until this case was on appeal. Plaintiffs, in defending the judgment, do not dispute the statements attributed to him in the press; they do not request an evidentiary hearing; and they do not argue that Microsoft should have filed a motion in the District Court before raising the matter on appeal. At oral argument, plaintiffs all but conceded that the Judge violated ethical restrictions by discussing the case in public: "On behalf of the governments, I have no brief to defend the District Judge's decision to discuss this case publicly while it was pending on appeal, and I have no brief to defend the judge's decision to discuss the case with reporters while the trial was proceeding, even given the embargo on any reporting concerning those conversations until after the trial." 02/27/01 Ct. Appeals Tr. at 326.

We must consider too that the federal disqualification provisions reflect a strong federal policy to preserve the actual and apparent impartiality of the federal judiciary. Judicial misconduct may implicate that policy regardless of the means by which it is disclosed to the public. Cf. The Washington Post v. Robinson, 935 F.2d 282, 291 (D.C. Cir. 1991) (taking judicial notice of newspaper articles to ascertain whether a fact was within public knowledge). Also, in our analysis of the arguments presented by the parties, the specifics of particular conversations are less important than their cumulative effect.

For these reasons we have decided to adjudicate Microsoft's disqualification request notwithstanding the state of the record. The same reasons also warrant a departure from our usual practice of declining to address issues raised for the first time on appeal: the "matter of what questions may be taken up and resolved for the first time on appeal is one left primarily to the discretion of the courts of appeals, to be exercised on the facts of individual cases." Singleton v. Wulff, 428 U.S. 106, 121 (1976); accord Hormel v. Helvering, 312 U.S. 552, 556-57 (1941); Nat'l Ass'n of Mfrs. v. Dep't of Labor, 159 F.3d 597, 605-06 (D.C. Cir. 1998). We will assume the truth of the press accounts and not send the case back for an evidentiary hearing on this subject. We reach no judgment on whether the details of the interviews were accurately recounted.

The published accounts indicate that the District Judge discussed numerous topics relating to the case. Among them was his distaste for the defense of technological integration--one of the central issues in the lawsuit. In September 1999, two months before his Findings of Fact and six months before his Conclusions of Law, and in remarks that were kept secret until after the Final Judgment, the Judge told reporters from the New York Times that he questioned Microsoft's integration of a web browser into Windows. Stating that he was "not a fan of integration," he drew an analogy to a 35-millimeter camera with an integrated light meter that in his view should also be offered separately: "You like the convenience of having a light meter built in, integrated, so all you have to do is press a button to get a reading. But do you think camera makers should also serve photographers who want to use a separate light meter, so they can hold it up, move it around?" Joel Brinkley & Steve Lohr, U.S. v. Microsoft 263 (2001). In other remarks, the Judge commented on the integration at the heart of the case: "[I]t was quite clear to me that the motive of Microsoft in bundling the Internet browser was not one of consumer convenience. The evidence that this was done for the consumer was not credible.... The evidence was so compelling that there was an ulterior motive." Wilke, Wall St. J. As for tying law in general, he criticized this court's ruling in the consent decree case, saying it "was wrongheaded on several counts" and would exempt the software industry from the antitrust laws. Brinkley & Lohr, U.S. v. Microsoft 78, 295; Brinkley & Lohr, N.Y. Times.

Reports of the interviews have the District Judge describing Microsoft's conduct, with particular emphasis on what he regarded as the company's prevarication, hubris, and impenitence. In some of his secret meetings with reporters, the Judge offered his contemporaneous impressions of testimony. He permitted at least one reporter to see an entry concerning Bill Gates in his "oversized green notebook." Ken Auletta, World War 3.0, at 112 (2001). He also provided numerous after-the-fact credibility assessments. He told reporters that Bill Gates' "testimony is inherently without credibility" and "[i]f you can't believe this guy, who else can you believe?" Brinkley & Lohr, U.S. v. Microsoft 278; Brinkley & Lohr, N.Y. Times; see also Auletta, The New Yorker, at 40. As for the company's other witnesses, the Judge is reported as saying that there "were times when I became impatient with Microsoft witnesses who were giving speeches." "[T]hey were telling me things I just flatly could not credit." Brinkley & Lohr, N.Y. Times. In an interview given the day he entered the break-up order, he summed things up: "Falsus in uno, falsus in omnibus": "Untrue in one thing, untrue in everything." "I don't subscribe to that as absolutely true. But it does lead one to suspicion. It's a universal human experience. If someone lies to you once, how much else can you credit as the truth?" Wilke, Wall St. J.

According to reporter Auletta, the District Judge told him in private that, "I thought they [Microsoft and its executives] didn't think they were regarded as adult members of the community. I thought they would learn." Auletta, World War 3.0, at 14. The Judge told a college audience that "Bill Gates is an ingenious engineer, but I don't think he is that adept at business ethics. He has not yet come to realise things he did (when Microsoft was smaller) he should not have done when he became a monopoly." Spiegel, Fin. Times. Characterizing Gates' and his company's "crime" as hubris, the Judge stated that "[i]f I were able to propose a remedy of my devising, I'd require Mr. Gates to write a book report" on Napoleon Bonaparte, "[b]ecause I think [Gates] has a Napoleonic concept of himself and his company, an arrogance that derives from power and unalloyed success, with no leavening hard experience, no reverses." Auletta, The New Yorker, at 41; see also Auletta, World War 3.0, at 397. The Judge apparently became, in Auletta's words, "increasingly troubled by what he learned about Bill Gates and couldn't get out of his mind the group picture he had seen of Bill Gates and Paul Allen and their shaggy-haired first employees at Microsoft." The reporter wrote that the Judge said he saw in the picture "a smart-mouthed young kid who has extraordinary ability and needs a little discipline. I've often said to colleagues that Gates would be better off if he had finished Harvard." Auletta, World War 3.0, at 168-69; see also Auletta, The New Yorker, at 46 (reporting the District Judge's statement that "they [Microsoft and its executives] don't act like grownups!" "[T]o this day they continue to deny they did anything wrong.").

The District Judge likened Microsoft's writing of incriminating documents to drug traffickers who "never figure out that they shouldn't be saying certain things on the phone." Brinkley & Lohr, U.S. v. Microsoft 6; Brinkley & Lohr, N.Y. Times. He invoked the drug trafficker analogy again to denounce Microsoft's protestations of innocence, this time with a reference to the notorious Newton Street Crew that terrorized parts of Washington, D.C. Reporter Auletta wrote in The New Yorker that the Judge

went as far as to compare the company's declaration of innocence to the protestations of gangland killers. He was referring to five gang members in a racketeering, drug-dealing, and murder trial that he had presided over four years earlier. In that case, the three victims had had their heads bound with duct tape before they were riddled with bullets from semi-automatic weapons. "On the day of the sentencing, the gang members maintained that they had done nothing wrong, saying that the whole case was a conspiracy by the white power structure to destroy them," Jackson recalled. "I am now under no illusions that miscreants will realize that other parts of society will view them that way."

Auletta, The New Yorker, at 40-41; Auletta, World War 3.0, at 369-70 (same); see also Auletta, The New Yorker, at 46.

The District Judge also secretly divulged to reporters his views on the remedy for Microsoft's antitrust violations. On the question whether Microsoft was entitled to any process at the remedy stage, the Judge told reporters in May 2000 that he was "not aware of any case authority that says I have to give them any due process at all. The case is over. They lost." Brinkley & Lohr, N.Y. Times. Another reporter has the Judge asking "[w]ere the Japanese allowed to propose terms of their surrender?" Spiegel, Fin. Times. The District Judge also told reporters the month before he issued his break-up order that "[a]ssuming, as I think they are, [ ] the Justice Department and the states are genuinely concerned about the public interest," "I know they have carefully studied all the possible options. This isn't a bunch of amateurs. They have consulted with some of the best minds in America over a long period of time." "I am not in a position to duplicate that and re-engineer their work. There's no way I can equip myself to do a better job than they have done." Brinkley & Lohr, N.Y. Times; cf. Final Judgment, at 62-63.

In February 2000, four months before his final order splitting the company in two, the District Judge reportedly told New York Times reporters that he was "not at all comfortable with restructuring the company," because he was unsure whether he was "competent to do that." Brinkley & Lohr, N.Y. Times; see also Brinkley & Lohr, U.S. v. Microsoft 277-78 (same); cf. Auletta, World War 3.0, at 370 (comment by the Judge in April 2000 that he was inclining toward behavioral rather than structural remedies). A few months later, he had a change of heart. He told the same reporters that "with what looks like Microsoft intransigence, a breakup is inevitable." Brinkley & Lohr, N.Y. Times; see also Brinkley & Lohr, U.S. v. Microsoft 315. The Judge recited a "North Carolina mule trainer" story to explain his change in thinking from "[i]f it ain't broken, don't try to fix it" and "I just don't think that [restructuring the company] is something I want to try to do on my own" to ordering Microsoft broken in two:

He had a trained mule who could do all kinds of wonderful tricks. One day somebody asked him: "How do you do it? How do you train the mule to do all these amazing things?" "Well," he answered, "I'll show you." He took a 2-by-4 and whopped him upside the head. The mule was reeling and fell to his knees, and the trainer said: "You just have to get his attention."

Brinkley & Lohr, U.S. v. Microsoft 278. The Judge added: "I hope I've got Microsoft's attention." Id.; see also Grimaldi, Wash. Post (comments by the Judge blaming the break-up on Microsoft's intransigence and on what he perceived to be Microsoft's responsibility for the failure of settlement talks); Spiegel, Fin. Times (the Judge blaming break-up on Microsoft's intransigence).

B. Violations of the Code of Conduct for United States Judges

The Code of Conduct for United States Judges was adopted by the Judicial Conference of the United States in 1973. It prescribes ethical norms for federal judges as a means to preserve the actual and apparent integrity of the federal judiciary. Every federal judge receives a copy of the Code, the Commentary to the Code, the Advisory Opinions of the Judicial Conference's Committee on Codes of Conduct, and digests of the Committee's informal, unpublished opinions. See II Guide to Judiciary Policies and Procedures (1973). The material is periodically updated. Judges who have questions about whether their conduct would be consistent with the Code may write to the Codes of Conduct Committee for a written, confidential opinion. See Introduction, Code of Conduct. The Committee traditionally responds promptly. A judge may also seek informal advice from the Committee's circuit representative.

While some of the Code's Canons frequently generate questions about their application, others are straightforward and easily understood. Canon 3A(6) is an example of the latter. In forbidding federal judges to comment publicly "on the merits of a pending or impending action," Canon 3A(6) applies to cases pending before any court, state or federal, trial or appellate. See Jeffrey M. Shaman et al., Judicial Conduct and Ethics s 10.34, at 353 (3d ed. 2000). As "impending" indicates, the prohibition begins even before a case enters the court system, when there is reason to believe a case may be filed. Cf. E. Wayne Thode, Reporter's Notes to Code of Judicial Conduct 54 (1973). An action remains "pending" until "completion of the appellate process." Code of Conduct Canon 3A(6) cmt.; Comm. on Codes of Conduct, Adv. Op. No. 55 (1998).

The Microsoft case was "pending" during every one of the District Judge's meetings with reporters; the case is "pending" now; and even after our decision issues, it will remain pending for some time. The District Judge breached his ethical duty under Canon 3A(6) each time he spoke to a reporter about the merits of the case. Although the reporters interviewed him in private, his comments were public. Court was not in session and his discussion of the case took place outside the presence of the parties. He provided his views not to court personnel assisting him in the case, but to members of the public. And these were not just any members of the public. Because he was talking to reporters, the Judge knew his comments would eventually receive widespread dissemination.

It is clear that the District Judge was not discussing purely procedural matters, which are a permissible subject of public comment under one of the Canon's three narrowly drawn exceptions. He disclosed his views on the factual and legal matters at the heart of the case. His opinions about the credibility of witnesses, the validity of legal theories, the culpability of the defendant, the choice of remedy, and so forth all dealt with the merits of the action. It is no excuse that the Judge may have intended to "educate" the public about the case or to rebut "public misperceptions" purportedly caused by the parties. See Grimaldi, Wash. Post; Microsoft Judge Says He May Step down from Case on Appeal, Wall St. J., Oct. 30, 2000. If those were his intentions, he could have addressed the factual and legal issues as he saw them--and thought the public should see them--in his Findings of Fact, Conclusions of Law, Final Judgment, or in a written opinion. Or he could have held his tongue until all appeals were concluded.

Far from mitigating his conduct, the District Judge's insistence on secrecy--his embargo--made matters worse. Concealment of the interviews suggests knowledge of their impropriety. Concealment also prevented the parties from nipping his improprieties in the bud. Without any knowledge of the interviews, neither the plaintiffs nor the defendant had a chance to object or to seek the Judge's removal before he issued his Final Judgment.

Other federal judges have been disqualified for making limited public comments about cases pending before them. See In re Boston's Children First, 244 F.3d 164 (1st Cir. 2001); In re IBM Corp., 45 F.3d 641 (2d Cir. 1995); United States v. Cooley, 1 F.3d 985 (10th Cir. 1993). Given the extent of the Judge's transgressions in this case, we have little doubt that if the parties had discovered his secret liaisons with the press, he would have been disqualified, voluntarily or by court order. Cf. In re Barry, 946 F.2d 913 (D.C. Cir. 1991) (per curiam); id. at 915 (Edwards, J., dissenting).

In addition to violating the rule prohibiting public comment, the District Judge's reported conduct raises serious questions under Canon 3A(4). That Canon states that a "judge should accord to every person who is legally interested in a proceeding, or the person's lawyer, full right to be heard according to law, and, except as authorized by law, neither initiate nor consider ex parte communications on the merits, or procedures affecting the merits, of a pending or impending proceeding." Code of Conduct Canon 3A(4).

What did the reporters convey to the District Judge during their secret sessions? By one account, the Judge spent a total of ten hours giving taped interviews to one reporter. Auletta, World War 3.0, at 14 n.*. We do not know whether he spent even more time in untaped conversations with the same reporter, nor do we know how much time he spent with others. But we think it safe to assume that these interviews were not monologues. Interviews often become conversations. When reporters pose questions or make assertions, they may be furnishing information, information that may reflect their personal views of the case. The published accounts indicate this happened on at least one occasion. Ken Auletta reported, for example, that he told the Judge "that Microsoft employees professed shock that he thought they had violated the law and behaved unethically," at which time the Judge became "agitated" by "Microsoft's 'obstinacy'." Id. at 369. It is clear that Auletta had views of the case. As he wrote in a Washington Post editorial, "[a]nyone who sat in [the District Judge's] courtroom during the trial had seen ample evidence of Microsoft's sometimes thuggish tactics." Ken Auletta, Maligning the Microsoft Judge, Wash. Post, Mar. 7, 2001, at A23.

The District Judge's repeated violations of Canons 3A(6) and 3A(4) also violated Canon 2, which provides that "a judge should avoid impropriety and the appearance of impropriety in all activities." Code of Conduct Canon 2; see also In re Charge of Judicial Misconduct, 47 F.3d 399, 400 (10th Cir. Jud. Council 1995) ("The allegations of extra-judicial comments cause the Council substantial concern under both Canon 3A(6) and Canon 2 of the Judicial Code of Conduct."). Canon 2A requires federal judges to "respect and comply with the law" and to "act at all times in a manner that promotes public confidence in the integrity and impartiality of the judiciary." Code of Conduct Canon 2A. The Code of Conduct is the law with respect to the ethical obligations of federal judges, and it is clear the District Judge violated it on multiple occasions in this case. The rampant disregard for the judiciary's ethical obligations that the public witnessed in this case undoubtedly jeopardizes "public confidence in the integrity" of the District Court proceedings.

Another point needs to be stressed. Rulings in this case have potentially huge financial consequences for one of the nation's largest publicly-traded companies and its investors. The District Judge's secret interviews during the trial provided a select few with inside information about the case, information that enabled them and anyone they shared it with to anticipate rulings before the Judge announced them to the world. Although he "embargoed" his comments, the Judge had no way of policing the reporters. For all he knew there may have been trading on the basis of the information he secretly conveyed. The public cannot be expected to maintain confidence in the integrity and impartiality of the federal judiciary in the face of such conduct.

C. Appearance of Partiality

The Code of Conduct contains no enforcement mechanism. See Thode, Reporter's Notes to Code of Judicial Conduct 43. The Canons, including the one that requires a judge to disqualify himself in certain circumstances, see Code of Conduct Canon 3C, are self-enforcing. There are, however, remedies extrinsic to the Code. One is an internal disciplinary proceeding, begun with the filing of a complaint with the clerk of the court of appeals pursuant to 28 U.S.C. s 372(c). Another is disqualification of the offending judge under either 28 U.S.C. s 144, which requires the filing of an affidavit while the case is in the District Court, or 28 U.S.C. s 455, which does not. Microsoft urges the District Judge's disqualification under s 455(a): a judge "shall disqualify himself in any proceeding in which his impartiality might reasonably be questioned." 28 U.S.C. s 455(a). The standard for disqualification under s 455(a) is an objective one. The question is whether a reasonable and informed observer would question the judge's impartiality. See In re Barry, 946 F.2d at 914; see also In re Aguinda, 241 F.3d 194, 201 (2d Cir. 2001); Richard E. Flamm, Judicial Disqualification s 24.2.1 (1996).

"The very purpose of s 455(a) is to promote confidence in the judiciary by avoiding even the appearance of impropriety whenever possible." Liljeberg v. Health Servs. Acquisition Corp., 486 U.S. 847, 865 (1988). As such, violations of the Code of Conduct may give rise to a violation of s 455(a) if doubt is cast on the integrity of the judicial process. It has been argued that any "public comment by a judge concerning the facts, applicable law, or merits of a case that is sub judice in his court or any comment concerning the parties or their attorneys would raise grave doubts about the judge's objectivity and his willingness to reserve judgment until the close of the proceeding." William G. Ross, Extrajudicial Speech: Charting the Boundaries of Propriety, 2 Geo. J. Legal Ethics 589, 598 (1989). Some courts of appeals have taken a hard line on public comments, finding violations of s 455(a) for judicial commentary on pending cases that seems mild in comparison to what we are confronting in this case. See Boston's Children First, 244 F.3d 164 (granting writ of mandamus ordering district judge to recuse herself under s 455(a) because of public comments on class certification and standing in a pending case); In re IBM Corp., 45 F.3d 641 (granting writ of mandamus ordering district judge to recuse himself based in part on the appearance of partiality caused by his giving newspaper interviews); Cooley, 1 F.3d 985 (vacating convictions and disqualifying district judge for appearance of partiality because he appeared on television program Nightline and stated that abortion protestors in a case before him were breaking the law and that his injunction would be obeyed).

While s 455(a) is concerned with actual and apparent impropriety, the statute requires disqualification only when a judge's "impartiality might reasonably be questioned." 28 U.S.C. s 455(a). Although this court has condemned public judicial comments on pending cases, we have not gone so far as to hold that every violation of Canon 3A(6) or every impropriety under the Code of Conduct inevitably destroys the appearance of impartiality and thus violates s 455(a). See In re Barry, 946 F.2d at 914; see also Boston's Children First, 244 F.3d at 168; United States v. Fortier, 242 F.3d 1224, 1229 (10th Cir. 2001).

In this case, however, we believe the line has been crossed. The public comments were not only improper, but also would lead a reasonable, informed observer to question the District Judge's impartiality. Public confidence in the integrity and impartiality of the judiciary is seriously jeopardized when judges secretly share their thoughts about the merits of pending cases with the press. Judges who covet publicity, or convey the appearance that they do, lead any objective observer to wonder whether their judgments are being influenced by the prospect of favorable coverage in the media. Discreet and limited public comments may not compromise a judge's apparent impartiality, but we have little doubt that the District Judge's conduct had that effect. Appearance may be all there is, but that is enough to invoke the Canons and s 455(a).

Judge Learned Hand spoke of "this America of ours where the passion for publicity is a disease, and where swarms of foolish, tawdry moths dash with rapture into its consuming fire...." Learned Hand, The Spirit of Liberty 132-33 (2d ed. 1953). Judges are obligated to resist this passion. Indulging it compromises what Edmund Burke justly regarded as the "cold neutrality of an impartial judge." Cold or not, federal judges must maintain the appearance of impartiality. What was true two centuries ago is true today: "Deference to the judgments and rulings of courts depends upon public confidence in the integrity and independence of judges." Code of Conduct Canon 1 cmt. Public confidence in judicial impartiality cannot survive if judges, in disregard of their ethical obligations, pander to the press.

We recognize that it would be extraordinary to disqualify a judge for bias or appearance of partiality when his remarks arguably reflected what he learned, or what he thought he learned, during the proceedings. See Liteky v. United States, 510 U.S. 540, 554-55 (1994); United States v. Barry, 961 F.2d 260, 263 (D.C. Cir. 1992). But this "extrajudicial source" rule has no bearing on the case before us. The problem here is not just what the District Judge said, but to whom he said it and when. His crude characterizations of Microsoft, his frequent denigrations of Bill Gates, his mule trainer analogy as a reason for his remedy--all of these remarks and others might not have given rise to a violation of the Canons or of s 455(a) had he uttered them from the bench. See Liteky, 510 U.S. at 555-56; Code of Conduct Canon 3A(6) (exception to prohibition on public comments for "statements made in the course of the judge's official duties"). But then Microsoft would have had an opportunity to object, perhaps even to persuade, and the Judge would have made a record for review on appeal. It is an altogether different matter when the statements are made outside the courtroom, in private meetings unknown to the parties, in anticipation that ultimately the Judge's remarks would be reported. Rather than manifesting neutrality and impartiality, the reports of the interviews with the District Judge convey the impression of a judge posturing for posterity, trying to please the reporters with colorful analogies and observations bound to wind up in the stories they write. Members of the public may reasonably question whether the District Judge's desire for press coverage influenced his judgments, indeed whether a publicity-seeking judge might consciously or subconsciously seek the publicity-maximizing outcome. We believe, therefore, that the District Judge's interviews with reporters created an appearance that he was not acting impartially, as the Code of Conduct and s 455(a) require.

D. Remedies for Judicial Misconduct and Appearance of Partiality

1. Disqualification

Disqualification is mandatory for conduct that calls a judge's impartiality into question. See 28 U.S.C. s 455(a); In re School Asbestos Litig., 977 F.2d 764, 783 (3d Cir. 1992). Section 455 does not prescribe the scope of disqualification. Rather, Congress "delegated to the judiciary the task of fashioning the remedies that will best serve the purpose" of the disqualification statute. Liljeberg, 486 U.S. at 862.

At a minimum, s 455(a) requires prospective disqualification of the offending judge, that is, disqualification from the judge's hearing any further proceedings in the case. See United States v. Microsoft Corp., 56 F.3d 1448, 1463-65 (D.C. Cir. 1995) (per curiam) ("Microsoft I"). Microsoft urges retroactive disqualification of the District Judge, which would entail disqualification antedated to an earlier part of the proceedings and vacatur of all subsequent acts. Cf. In re School Asbestos Litig., 977 F.2d at 786 (discussing remedy options).

"There need not be a draconian remedy for every violation of s 455(a)." Liljeberg, 486 U.S. at 862. Liljeberg held that a district judge could be disqualified under s 455(a) after entering final judgment in a case, even though the judge was not (but should have been) aware of the grounds for disqualification before final judgment. The Court identified three factors relevant to the question whether vacatur is appropriate: "in determining whether a judgment should be vacated for a violation of s 455(a), it is appropriate to consider the risk of injustice to the parties in the particular case, the risk that the denial of relief will produce injustice in other cases, and the risk of undermining the public's confidence in the judicial process." Id. at 864. Although the Court was discussing s 455(a) in a slightly different context (the judgment there had become final after appeal and the movant sought to have it vacated under Rule 60(b)), we believe the test it propounded applies as well to cases such as this in which the full extent of the disqualifying circumstances came to light only while the appeal was pending. See In re School Asbestos Litig., 977 F.2d at 785.

Our application of Liljeberg leads us to conclude that the appropriate remedy for the violations of s 455(a) is disqualification of the District Judge retroactive only to the date he entered the order breaking up Microsoft. We therefore will vacate that order in its entirety and remand this case to a different District Judge, but will not set aside the existing Findings of Fact or Conclusions of Law (except insofar as specific findings are clearly erroneous or legal conclusions are incorrect).

This partially retroactive disqualification minimizes the risk of injustice to the parties and the damage to public confidence in the judicial process. Although the violations of the Code of Conduct and s 455(a) were serious, full retroactive disqualification is unnecessary. It would unduly penalize plaintiffs, who were innocent and unaware of the misconduct, and would have only slight marginal deterrent effect.

Most important, full retroactive disqualification is unnecessary to protect Microsoft's right to an impartial adjudication. The District Judge's conduct destroyed the appearance of impartiality. Microsoft neither alleged nor demonstrated that it rose to the level of actual bias or prejudice. There is no reason to presume that everything the District Judge did is suspect. See In re Allied-Signal Inc., 891 F.2d 974, 975-76 (1st Cir. 1989); cf. Liberty Lobby, Inc. v. Dow Jones & Co., 838 F.2d 1287, 1301-02 (D.C. Cir. 1988). Although Microsoft challenged very few of the findings as clearly erroneous, we have carefully reviewed the entire record and discern no basis to suppose that actual bias infected his factual findings.

The most serious judicial misconduct occurred near or during the remedial stage. It is therefore commensurate that our remedy focus on that stage of the case. The District Judge's impatience with what he viewed as intransigence on the part of the company; his refusal to allow an evidentiary hearing; his analogizing Microsoft to Japan at the end of World War II; his story about the mule--all of these out-of-court remarks and others, plus the Judge's evident efforts to please the press, would give a reasonable, informed observer cause to question his impartiality in ordering the company split in two.

To repeat, we disqualify the District Judge retroactive only to the imposition of the remedy, and thus vacate the remedy order for the reasons given in Section V and because of the appearance of partiality created by the District Judge's misconduct.

2. Review of Findings of Fact and Conclusions of Law

Given the limited scope of our disqualification of the District Judge, we have let stand for review his Findings of Fact and Conclusions of Law. The severity of the District Judge's misconduct and the appearance of partiality it created have led us to consider whether we can and should subject his factfindings to greater scrutiny. For a number of reasons we have rejected any such approach.

The Federal Rules require that district court findings of fact not be set aside unless they are clearly erroneous. See Fed. R. Civ. P. 52(a). Ordinarily, there is no basis for doubting that the District Court's factual findings are entitled to the substantial deference the clearly erroneous standard entails. But of course this is no ordinary case. Deference to a district court's factfindings presumes impartiality on the lower court's part. When impartiality is called into question, how much deference is due?

The question implies that there is some middle ground, but we believe there is none. As the rules are written, district court factfindings receive either full deference under the clearly erroneous standard or they must be vacated. There is no de novo appellate review of factfindings and no intermediate level between de novo and clear error, not even for findings the court of appeals may consider sub-par. See Amadeo v. Zant, 486 U.S. 214, 228 (1988) ("The District Court's lack of precision, however, is no excuse for the Court of Appeals to ignore the dictates of Rule 52(a) and engage in impermissible appellate factfinding."); Anderson v. City of Bessemer City, 470 U.S. 564, 571-75 (1985) (criticizing district court practice of adopting a party's proposed factfindings but overturning court of appeals' application of "close scrutiny" to such findings).

Rule 52(a) mandates clearly erroneous review of all district court factfindings: "Findings of fact, whether based on oral or documentary evidence, shall not be set aside unless clearly erroneous, and due regard shall be given to the opportunity of the trial court to judge of the credibility of the witnesses." Fed. R. Civ. P. 52(a). The rule "does not make exceptions or purport to exclude certain categories of factual findings from the obligation of a court of appeals to accept a district court's findings unless clearly erroneous." Pullman-Standard v. Swint, 456 U.S. 273, 287 (1982); see also Anderson, 470 U.S. at 574-75; Inwood Labs., Inc. v. Ives Labs., Inc., 456 U.S. 844, 855-58 (1982). The Supreme Court has emphasized on multiple occasions that "[i]n applying the clearly erroneous standard to the findings of a district court sitting without a jury, appellate courts must constantly have in mind that their function is not to decide factual issues de novo." Zenith Radio Corp. v. Hazeltine Research, Inc., 395 U.S. 100, 123 (1969); Anderson, 470 U.S. at 573 (quoting Zenith).

The mandatory nature of Rule 52(a) does not compel us to accept factfindings that result from the District Court's misapplication of governing law or that otherwise do not permit meaningful appellate review. See Pullman-Standard, 456 U.S. at 292; Inwood Labs., 456 U.S. at 855 n.15. Nor must we accept findings that are utterly deficient in other ways. In such a case, we vacate and remand for further factfinding. See 9 Moore's Federal Practice s 52.12[1] (Matthew Bender 3d ed. 2000); 9A Charles A. Wright & Arthur R. Miller, Federal Practice and Procedure s 2577, at 514-22 (2d ed. 1995); cf. Icicle Seafoods, Inc. v. Worthington, 475 U.S. 709, 714 (1986); Pullman-Standard, 456 U.S. at 291-92.

When there is fair room for argument that the District Court's factfindings should be vacated in toto, the court of appeals should be especially careful in determining that the findings are worthy of the deference Rule 52(a) prescribes. See, e.g., Thermo Electron Corp. v. Schiavone Constr. Co., 915 F.2d 770, 773 (1st Cir. 1990); cf. Bose Corp. v. Consumers Union of United States, Inc., 466 U.S. 485, 499 (1984). Thus, although Microsoft alleged only appearance of bias, not actual bias, we have reviewed the record with painstaking care and have discerned no evidence of actual bias. See S. Pac. Communications Co. v. AT & T, 740 F.2d 980, 984 (D.C. Cir. 1984); Cooley, 1 F.3d at 996 (disqualifying district judge for appearance of partiality but noting that "the record of the proceedings below ... discloses no bias").

In light of this conclusion, the District Judge's factual findings both warrant deference under the clear error standard of review and, though exceedingly sparing in citations to the record, permit meaningful appellate review. In reaching these conclusions, we have not ignored the District Judge's reported intention to craft his factfindings and Conclusions of Law to minimize the breadth of our review. The Judge reportedly told Ken Auletta that "[w]hat I want to do is confront the Court of Appeals with an established factual record which is a fait accompli." Auletta, World War 3.0, at 230. He explained: "part of the inspiration for doing that is that I take mild offense at their reversal of my preliminary injunction in the consent-decree case, where they went ahead and made up about ninety percent of the facts on their own." Id. Whether the District Judge takes offense, mild or severe, is beside the point. Appellate decisions command compliance, not agreement. We do not view the District Judge's remarks as anything other than his expression of disagreement with this court's decision, and his desire to provide extensive factual findings in this case, which he did.