On consideration of the petition for rehearing and suggestion for rehearing en banc filed in the above-entitled cause by plaintiff-appellee, a vote of the active members of the Court was requested, and a majority of the active members of the Court have voted to deny a rehearing en banc. All of the judges on the original panel have voted to deny the petition for rehearing. Accordingly,

IT IS ORDERED that the aforesaid petition for rehearing and suggestion for rehearing en banc be, and the same is hereby, DENIED.

EASTERBROOK, Circuit Judge, with whom POSNER, Chief Judge, and MANION and EVANS, Circuit Judges, join, dissenting from the denial of rehearing en banc.

The result of a 10-week trial should stand unless the verdict is unreliable. Yet two and a half years after this lengthy trial ended, the panel has set aside the verdicts for the sole reason that defense counsel were confused about the order in which members of the venire were to be seated, and therefore could not make the most advantageous use of their peremptory challenges. The panel did not conclude that this foul-up undermined confidence in the verdict; instead it refused to inquire whether the error was harmless. Quoting dicta in Swain v. Alabama, 380 U.S. 202, 219, 85 S. Ct. 824, 835, 13 L. Ed. 2d 759 (1965), that "[t]he denial or impairment of the right is reversible error without a showing of prejudice", the panel continued: "harmless-error analysis is inappropriate where a defendant's statutory right to peremptory challenge has been denied or impaired". United States v. Underwood, 122 F.3d 389, 392 (7th Cir.1997). This approach has the support of United States v. Taylor, 92 F.3d 1313, 1325 (2d Cir.1996); Kirk v. Raymark Industries, Inc., 61 F.3d 147, 158-62 (3d Cir.1995) (civil case); Knox v. Collins, 928 F.2d 657, 661 (5th Cir.1991) (collateral attack on state conviction); and United States v. Annigoni, 96 F.3d 1132 (9th Cir.1996) (en banc). But four other courts of appeals have held that an error that causes the defense to lose or waste peremptory challenges may be deemed harmless when, in the language of Fed.R.Crim.P. 52(a), the reduction in the number of available challenges does not affect "substantial rights". See McQueen v. Scroggy, 99 F.3d 1302, 1320-21 (6th Cir.1996) (collateral attack); United States v. Horsman, 114 F.3d 822, 825 (8th Cir.1997); Getter v. Wal-Mart Stores, Inc., 66 F.3d 1119, 1122 (10th Cir.1995) (civil case); United States v. Farmer, 923 F.2d 1557, 1566 & n. 20 (11th Cir.1991). As the ninth circuit did in Annigoni, we should resolve this issue en banc.

Although peremptory challenges are authorized in all federal jury trials, their use is increasingly contested now that Georgia v. McCollum, 505 U.S. 42, 112 S. Ct. 2348, 120 L. Ed. 2d 33 (1992), has held that prosecutors may object to challenges exercised by the defense. Assessing challenges on the fly yields many claims of error, and other things too can go awry, as they did in the jury selection preceding the Messinos' trial. Today's case starkly presents the question whether harmless-error analysis is necessary; no factual complications or extraneous legal questions impede decision. Our panel recognized that the subject is important, one judge wrote separately to express doubts about the doctrine of automatic reversal, and the conflict among the circuits demonstrates that the problem is common elsewhere. A significant, recurring legal issue is a prime candidate for hearing by the full court. That the panel has decided the issue incorrectly makes rehearing all the more appropriate.

Perfection is elusive. Appellate courts long ago ceased to be citadels of technicality and began to ask whether a particular error implies that the judgment is unreliable. To redo a lengthy trial--to redo it 3 years after the original trial, 5 years after the indictment, 17 years after the drug conspiracy began, as the panel has directed--is to ensure that the outcome will be less reliable than it was the first time even if there are no legal errors. Memories will have faded or become distorted, witnesses will have died, evidence will have vanished. New errors are bound to occur, perhaps more serious than those in the first trial. While the Messinos' second trial occupies the courtroom, other cases will fester in the queue, and litigants innocent of any responsibility for the delay will suffer losses as a result. As usual, the quest for the perfect is the enemy of the good.

The panel approached this issue as if it had to make a decision about wise policy. Should it undertake harmless-error analysis, given the inscrutable effects of missed opportunities in jury selection, or should it insist on error-free implementation of the rules and avoid that task? As an original matter that would be an interesting question. Though I have my doubts about the panel's assessment of the competing goals--judicial review of peremptory challenges after Batson v. Kentucky, 476 U.S. 79, 106 S. Ct. 1712, 90 L. Ed. 2d 69 (1986), means that there are more errors, and a correspondingly high benefit of a harmless-error rule--the subject is not open to decision as an original matter. Rule 52(a) provides: "Any error, defect, irregularity or variance which does not affect substantial rights shall be disregarded." It does not except errors affecting peremptory challenges. Any error that does not affect substantial rights shall be disregarded. See also 28 U.S.C. § 2111 ("On the hearing of any appeal or writ of certiorari in any case, the court shall give judgment after an examination of the record without regard to errors which do not affect the substantial rights of the parties."); Fed.R.Civ.P. 61 ("The court at every stage of the proceeding must disregard any error or defect in the proceeding which does not affect the substantial rights of the parties.").

Occasionally courts declare that Rule 52(a) and § 2111 just don't apply to one or another right. Fifteen years ago we held that a prosecutorial request to the jury to draw an adverse inference from the defendant's exercise of his privilege against self-incrimination could not be harmless. But the Supreme Court disagreed, holding that Rule 52(a) applies, just as it says, to any error, even one of constitutional dimension. United States v. Hasting, 461 U.S. 499, 103 S. Ct. 1974, 76 L. Ed. 2d 96 (1983). Despite Hasting, another court of appeals soon held that misjoinder never could be harmless because it is "inherently prejudicial". Again the Court reversed, holding that Rule 52(a) must be applied across the board. United States v. Lane, 474 U.S. 438, 106 S. Ct. 725, 88 L. Ed. 2d 814 (1986). Soon a conflict among the circuits developed on the question whether a court of appeals could assert a supervisory power to reverse without a harmless-error analysis. Yet again the Supreme Court disapproved, writing that "federal courts have no more discretion to disregard the Rule's mandate than they do to disregard constitutional or statutory provisions." Bank of Nova Scotia v. United States, 487 U.S. 250, 255, 108 S. Ct. 2369, 2373-74, 101 L. Ed. 2d 228 (1988).

Our panel did not discuss or even cite Rule 52(a), § 2111, Hasting, Lane, or Bank of Nova Scotia. Nor did it acknowledge any of the four recent appellate decisions employing the harmless-error rule. The panel's treatment of authority is the norm among appellate courts that refuse to use a harmless-error approach when dealing with peremptory challenges. The following table illustrates:

                            Discussion or citation of...
    Case       Rule 52/61    § 2111    Hasting    Lane    Nova Scotia
   Taylor          N           N          N        N           N
    Kirk           N           N          N        N           N
    Knox           N           N          N        N           N
  Underwood        N           N          N        N           N
  Annigoni         Y           N          N        N           N

The solitary Y in this table means that the ninth circuit addressed Rule 52(a) in Annigoni. But in doing so it adopted a line of argument that the Supreme Court soon disapproved in Johnson v. United States, --- U.S. ----, 117 S. Ct. 1544, 137 L. Ed. 2d 718 (1997).

Annigoni distinguished "classic trial error," which it said was the domain of Rule 52(a), from "structural error," which it thought led to automatic reversal. 96 F.3d at 1143. After concluding that an improper rejection of a peremptory challenge is not "trial error" because it occurs before the presentation of evidence begins, 96 F.3d at 1144, and that its effect is hard to assess, the ninth circuit concluded that harmless-error analysis need not be undertaken. Now this approach has a lot of trouble with Lane and Bank of Nova Scotia, each of which involved hard-to-evaluate errors that preceded the trial (misjoinder in Lane, defective indictment in Bank of Nova Scotia). Let that pass. The deeper problem is that Rule 52(a) does not permit courts to distinguish "trial" from "structural" errors. "Any error, defect, irregularity or variance which does not affect substantial rights shall be disregarded." About the time it decided Annigoni, the ninth circuit held that Rule 52(b)--the plain-error companion to the harmless-error doctrine--likewise is limited to "trial" errors, and that "structural" errors always require reversal. United States v. Keys, 95 F.3d 874 (9th Cir.1996) (en banc). The Supreme Court replied in Johnson, --- U.S. at ----, 117 S. Ct. at 1548:

Petitioner argues that she need not fall within the "limited" and "circumscribed" strictures of [United States v.] Olano[, 507 U.S. 725, 113 S. Ct. 1770, 123 L. Ed. 2d 508 (1993) ], because the error she complains of here is "structural," and so is outside Rule 52(b) altogether. But the seriousness of the error claimed does not remove consideration of it from the ambit of the Federal Rules of Criminal Procedure. None of the cases discussing "structural error," upon which petitioner relies, were direct appeals from judgments of conviction in the federal system. Several came from state courts which had considered the claimed error under their own rules. See Gideon v. Wainwright, 372 U.S. 335, 83 S. Ct. 792, 9 L. Ed. 2d 799 (1963); Arizona v. Fulminante, 499 U.S. 279, 111 S. Ct. 1246, 113 L. Ed. 2d 302 (1991); Sullivan v. Louisiana, 508 U.S. 275, 113 S. Ct. 2078, 124 L. Ed. 2d 182 (1993). Others came here by way of federal habeas challenges to state convictions. See Vasquez v. Hillery, 474 U.S. 254, 106 S. Ct. 617, 88 L. Ed. 2d 598 (1986); McKaskle v. Wiggins, 465 U.S. 168, 104 S. Ct. 944, 79 L. Ed. 2d 122 (1984). None of them were subject to the provisions of Rule 52.

But it is that Rule which by its terms governs direct appeals from judgments of conviction in the federal system, and therefore governs this case. We cautioned against any unwarranted expansion of Rule 52(b) in United States v. Young, 470 U.S. 1, 105 S. Ct. 1038, 84 L. Ed. 2d 1 (1985), because it "would skew the Rule's 'careful balancing of our need to encourage all trial participants to seek a fair and accurate trial the first time around against our insistence that obvious injustice be promptly redressed,' " id., at 15, 105 S.Ct. at 1046 (quoting United States v. Frady, 456 U.S. 152, 163, 102 S. Ct. 1584, 1592, 71 L. Ed. 2d 816 (1982)). Even less appropriate than an unwarranted expansion of the Rule would be the creation out of whole cloth of an exception to it, an exception which we have no authority to make. See Carlisle v. United States, 517 U.S. 416, ----, 116 S. Ct. 1460, 1466, 134 L. Ed. 2d 613 (1996).

In other words: Rule 52 governs every issue in all federal criminal prosecutions; a distinction between "trial" and "structural" error is appropriate only in state cases, to which Rule 52 does not apply; judges have "no authority" to make exceptions to Rule 52. The Court followed up by vacating United States v. Keys, --- U.S. ----, 117 S. Ct. 1816, 137 L. Ed. 2d 1025 (1997). Although certiorari was not sought in Annigoni, its foundation has been washed away. Yet our panel followed Annigoni without discussing Johnson.

Before taking up Rule 52, Annigoni asserted: "Given the obvious importance of the right of peremptory challenge, a criminal defendant has long had a powerful remedy for the impairment or denial of this right: automatic reversal of the conviction." 96 F.3d at 1141. For this proposition it cited Harrison v. United States, 163 U.S. 140, 142, 16 S. Ct. 961, 961, 41 L. Ed. 104 (1896); Gulf, Colorado & Santa Fe Ry. v. Shane, 157 U.S. 348, 351, 15 S. Ct. 641, 642, 39 L. Ed. 727 (1895); and Lewis v. United States, 146 U.S. 370, 376, 13 S. Ct. 136, 138, 36 L. Ed. 1011 (1892). These cases have one essential feature in common: they were decided more than a century ago, during the Bad Old Days when appellate courts reversed for any error. It is exactly this inflexibility that § 2111 and Rule 52(a) jettison. All doubt was stilled in Lane:

The Court of Appeals held that misjoinder "is inherently prejudicial." ... The Circuits are divided on the question whether misjoinder requires automatic reversal, or whether the harmless-error rule governs. Most Circuits that have adopted the per se approach have relied on McElroy v. United States, 164 U.S. 76[, 17 S. Ct. 31, 41 L. Ed. 355] (1896), where this Court applied the joinder statute then in force and reversed convictions of jointly tried defendants after rejecting the Government's argument that there was no showing of prejudice. Id., at 81[, 17 S.Ct. at 33].

McElroy, however, was decided long before the adoption of Federal Rules of Criminal Procedure 8 and 52, and prior to the enactment of the harmless-error statute, 28 U.S.C. § 2111, which provides that on appeal we are to ignore "errors or defects which do not affect the substantial rights of the parties." Under Rule 52(a), we are similarly instructed that any error "which does not affect substantial rights shall be disregarded."

474 U.S. at 444-45, 106 S. Ct. at 729 (footnotes omitted). McElroy is of a piece with Harrison, Shane, and Lewis. Their approach is defunct--terminated for federal cases by statute and rule, for state cases by Chapman v. California, 386 U.S. 18, 87 S. Ct. 824, 17 L. Ed. 2d 705 (1967), and its successors, such as Brecht v. Abrahamson, 507 U.S. 619, 113 S. Ct. 1710, 123 L. Ed. 2d 353 (1993). To overthrow the automatic-reversal rule of one exemplar is to overthrow that approach altogether. Yet Annigoni did not cite or discuss Lane. Neither did our panel.

Swain, the solitary decision of the Supreme Court on which the panel relied, offers poor support for automatic reversal. The brief passage that the panel quoted was dictum based on Harrison, Shane, and Lewis. Swain did not contend that he had been deprived of a peremptory challenge. He attacked the prosecutor's challenges and lost on the merits. Two years later, Chapman changed the landscape for cases coming from state courts--as § 2111 and Rule 52 had done for cases in the federal courts. Johnson holds that the Court's approach to the review of state judgments (of which Swain was one) is not instructive in federal prosecutions. Then there is the fact that Batson overruled Swain. If the doctrine of automatic reversal ever had a foundation, it was that peremptory challenges are, well, peremptory, to be exercised in the parties' unfettered discretion, and that harmless-error analysis is a form of judicial inquiry into the utility of a challenge. "For it is, as Blackstone says, an arbitrary and capricious right; and it must be exercised with full freedom, or it fails of its full purpose." Lewis, 146 U.S. at 378, 13 S.Ct. at 139, quoted in Swain, 380 U.S. at 219, 85 S.Ct. at 835. After Batson and McCollum, however, peremptory challenges are subject to judicial control, and there is no more support for automatic reversal than there is when the judge resolves any of the thousand other issues that influence the conduct of the trial.

I have lavished attention on Annigoni because it is the most detailed of the panel's precursors. The other three recent appellate decisions announcing support for automatic reversal add nothing of value. As the table shows, none of them discusses § 2111, Rule 52(a) (or Fed.R.Civ.P. 61), or any of the important decisions of the Supreme Court. Taylor tossed off a single unreasoned sentence; like Swain, the passage in Taylor is dictum, for the court held that there had been no error. Kirk invokes Harrison, Shane, and Lewis in addition to Swain's dictum and does not offer an independent justification of its conclusion. Knox is a collateral attack on a state conviction, to which Rule 52 does not apply. It says that defendants have a constitutional right to peremptory challenges, and that harmless-error analysis is "accordingly" inapplicable. 928 F.2d at 661. Both the premise and the conclusion of Knox are untenable. The constitutional right is to an impartial jury, not to remove members of the venire who the judge rightly concludes could render an unbiased verdict. Ross v. Oklahoma, 487 U.S. 81, 88, 108 S. Ct. 2273, 2278, 101 L. Ed. 2d 80 (1988) (the right to peremptory challenges is "not of constitutional dimension"); accord, McCollum, supra; cf. United States v. Boyd, 86 F.3d 719 (7th Cir.1996). Anyway, Hasting and Chapman show that the constitutional footing of a right does not preclude a harmless-error inquiry. Other cases, such as United States v. Vargas, 606 F.2d 341, 346 (1st Cir.1979), and United States v. Rucker, 557 F.2d 1046, 1048-49 (4th Cir.1977), that announced an automatic-reversal rule before Hasting, Lane, Bank of Nova Scotia, Batson, McCollum, and Ross, do not require any additional comment.

One final possibility: perhaps § 2111 and Rule 52(a) apply, but an error concerning a peremptory challenge always affects a "substantial right". A right is "substantial" when it is one of the pillars of a fair trial. Trial before an orangutan, or the grant of summary judgment against the accused in a criminal case, would deprive the defendant of a "substantial" right even if it were certain that a jury would convict. Sullivan v. Louisiana, 508 U.S. 275, 279, 113 S. Ct. 2078, 2081-82, 124 L. Ed. 2d 182 (1993); cf. California v. Roy, --- U.S. ----, 117 S. Ct. 337, 136 L. Ed. 2d 266 (1996). For the same reason, a biased tribunal always deprives the accused of a substantial right. Bracy v. Gramley, --- U.S. ----, 117 S. Ct. 1793, 138 L. Ed. 2d 97 (1997); Cartalino v. Washington, 122 F.3d 8 (7th Cir.1997). See also Gomez v. United States, 490 U.S. 858, 876, 109 S. Ct. 2237, 2248, 104 L. Ed. 2d 923 (1989) (decision by an unauthorized tribunal deprives the defendant of a substantial right). Deprivation of counsel likewise so undermines the ability to distinguish the guilty from the innocent that it always leads to reversal. See United States v. Cronic, 466 U.S. 648, 658-59, 104 S. Ct. 2039, 2046-47, 80 L. Ed. 2d 657 (1984); Castellanos v. United States, 26 F.3d 717 (7th Cir.1994). After McCollum and Ross, however, it is impossible to group an error concerning peremptory challenges with the denial of counsel or trial before a bribed judge. If the jury that actually sits is impartial, the defendant has enjoyed the substantial right. A reduction in the number of peremptory challenges from 10, the number Fed.R.Crim.P. 24(b) now provides in felony cases, to 3, the number it makes available in misdemeanors, could not be called a "deprivation" of a "substantial" right; indeed, after McCollum even the abrogation of peremptory challenges could not be so labeled. Many thoughtful persons, including one Justice of the Supreme Court, see Batson, 476 U.S. at 107-08, 106 S.Ct. at 1728-29 (Marshall, J., concurring), have called for the abolition of peremptory challenges, which are unknown in most legal systems outside the United States. Peremptory challenges enable defendants to feel more comfortable with the jury that is to determine their fate, but increasing litigants' comfort level is only one goal among many, and reduced peace of mind is a bad reason to retry complex cases decided by impartial juries. What was designed as a buffer zone to enable litigants to avoid jurors whose impartiality is in question, but not enough to support a challenge for cause, has become an invitation to racial (and other) discrimination, see McCollum and Batson, and a mechanism used by both sides to skew the jury in one's favor by removing members of the venire whose very impartiality is seen by the litigant as an obstacle to be overcome. Because these efforts offset each other, and it is hard to predict jurors' votes, litigants rarely succeed in producing a more favorable panel. See Hans Zeisel & Shari Seidman Diamond, The Effect of Peremptory Challenges on Jury and Verdict: An Experiment in a Federal District Court, 30 Stan. L.Rev. 491 (1978); Symposium on the Selection and Function of the Modern Jury, 40 American L.Rev. (Win.1991). It is therefore inconceivable that if the defendant freely exercises 8, or 5, or 3, peremptory challenges, rather than 10, any "substantial" right has been lost. Cases such as United States v. Magana, 118 F.3d 1173, 1206 (7th Cir.1997), and United States v. Farmer, 924 F.2d 647, 653 (7th Cir.1991), which say that it is within the judge's discretion to provide or withhold additional peremptory challenges in multi-defendant cases, recognize that a small change in the number of peremptory challenges does not affect litigants' vital interests. Cf. United States v. Cambara, 902 F.2d 144, 147-48 (1st Cir.1990) (an erroneous deprivation of an extra peremptory challenge awarded under Rule 24(b) may be deemed harmless).

In any given case there remains the possibility that a blunder affects a right that is substantial in the sense of Kotteakos v. United States, 328 U.S. 750, 66 S. Ct. 1239, 90 L. Ed. 1557 (1946): that the error "had substantial and injurious effect or influence in determining the jury's verdict". 328 U.S. at 776, 66 S.Ct. at 1253. See also O'Neal v. McAninch, 513 U.S. 432, 115 S. Ct. 992, 130 L. Ed. 2d 947 (1995). In a trial like this, the possibility that lawyers' confusion in the exercise of peremptory challenges altered the outcome is too remote to be worth investigating. The confusion affected only two members of the venire--one of whom was objectionable to just one attorney (whose clients were acquitted), and the other of whom the defense deemed an acceptable alternate juror. Perhaps it is impossible to show that the loss or disallowance of a peremptory challenge affected the outcome of a trial--but inability to trace adverse effects to a mistake does not justify reversing a conviction; it shows instead that there is no warrant for disturbing the judgment. United States v. Morrison, 449 U.S. 361, 101 S. Ct. 665, 66 L. Ed. 2d 564 (1981).

Whether harmless-error analysis is essential is a subject well worth the time of the court en banc. The question has arisen in almost every circuit; it is bound to come up more frequently now that McCollum requires judges to scrutinize defendants' and prosecutors' challenges alike; and the Supreme Court seems to have had trouble finding an appropriate case in which to resolve the conflict among the circuits, so the issue may be in our own bailiwick for a while. My colleagues' willingness to leave our circuit in the wrong camp pending that resolution is regrettable.