Daniel Taglia and Robert McDonnell were convicted by a jury of labor racketeering and sentenced to four and six years in prison, respectively.
Taglia was an acquaintance both of McDonnell, a lawyer, and of Robert Blessing, the nephew of an official , of a local union. Taglia and McDonnell conspired— so the jury could, and did, find — to bribe officials of the union to award a contract to McDonnell to furnish the union’s members with prepaid legal services financed by a deduction from wages. Blessing was to put Taglia and McDonnell into contact with his uncle. But Blessing was an FBI informant, whom the FBI instructed to play along with the scheme; and it also enlisted an honest union official to play the role, originally intended for the uncle, of corrupt union official. Taglia and McDonnell several times met or spoke by telephone with Blessing, with the official, or with both. A number of these conversations were tape recorded and played back to the jury. Conducted without code words or euphemisms, these conversations discussed the kickback scheme at length and in detail. It is impossible to place an innocent construction on them. Although Blessing testified at trial — was in fact the principal live witness for the prosecution — the evidence most damaging to the defendants was the tapes.
The appeals raise a large number of issues, only three of which require discussion; the others have no possible merit. First, both defendants complain about the district judge’s refusal to grant their motion for a new trial (Fed.R.Crim.P. 33) on the ground of newly discovered evidence. Blessing had testified in an unrelated trial, held the year before the trial in this case, that he had seen Taglia deliver a package of cocaine. He had described the package in detail. The testimony was false. He had not seen Taglia engage in a drug transaction. The defendants argued in their motion for a new trial that they should be allowed to use the evidence of this lie to impeach Blessing’s testimony against them. The district judge denied the motion on the ground that this impeachment would not have changed the outcome of the trial.
The government defends the judge’s ruling on the ground that newly discovered evidence that is merely impeaching is not a permissible ground for a new trial. There is language to this effect- in countless cases, illustrated by
Mesarosh v. United States,
Of course it will be the rare case in which impeaching evidence warrants a new trial, because ordinarily such evidence will cast doubt at most on the testimony of only one of the witnesses. The judicial language that seems to exclude impeaching testimony from the scope of Rule 33 thus illustrates the tendency to over generalize. It is easy to confuse a practice with a rule. The practice has been to deny new trials where the only newly discovered evidence was impeaching. But the practice should not be taken to imply a rule that even if the defendant proves that his conviction almost certainly rests on a lie, the district judge is helpless to grant a new trial. District judges do not in fact consider themselves *416 helpless in such circumstances, and they are right not to.
This, however, is not that rare case. Giving due regard to the limitations on our power to second guess a district judge’s decision not to grant a new trial,
United States v. Morales,
The second issue that merits discussion is whether Taglia’s motion for a severance should have been granted. McDonnell, his codefendant, was his former lawyer. So was Alan Ackerman, McDonnell’s counsel. Taglia feared that Ackerman and McDonnell, who with the court’s permission was acting as his own co-counsel — a procedure we have called impermissible,
United States v. Oakey,
The government argues that Taglia waived the issue whether his trial should have been severed from McDonnell’s by failing to renew the motion for severance. Had he renewed it, the government suggests, the judge might have placed limitations on any cross-examination of Taglia that McDonnell or Ackerman might want to conduct. The government adds that since Taglia would have denied any criminal conduct, it would hardly have been in his codefendant’s interest to try to shake his credibility by cross-examination.
We agree that there was waiver. Apparently the motion to sever was lost in the shuffle before trial. Those things happen. If a motion is not acted upon, a litigant had better renew it. He may not lull the judge into thinking that it has been abandoned and then, after he has lost, pull a rabbit out of his pocket in the form of the forgotten motion.
United States v. Brown,
The denial of motions to sever has become almost a reflex. This is not a surprise, for there is a presumption that participants in a conspiracy or other crimi
*417
nal scheme should be tried together, not only to economize on judicial and prosecu-torial resources but also to give the jury a fuller picture of the scheme.
United States v. Velasquez,
If the judge would have had to grant Taglia’s motion to sever had he renewed it during the trial, then it might seem that the failure to renew the motion must have been harmless. Not so. Taglia may not have renewed the motion because as the trial developed he became content with or even came to prefer a joint trial. We cannot countenance a system in which a defendant first tries to see whether he can get an acquittal in a joint trial, and then when he is convicted renews his motion to sever so that he can have another crack at a jury. This is not a case like
United States v. Brown, supra,
Last we consider McDonnell’s challenge to the effectiveness of attorney Alan Ackerman’s representation of him. McDonnell makes a number of points, including the brevity of Ackerman’s closing statement, that neither singly nor in combination establish either that Ackerman’s representation of him fell below minimum professional standards or that he might have been acquitted had Ackerman done a better job. And he must show both things to upset his conviction on the ground of ineffective assistance of counsel.
Strickland v. Washington,
A brief comment on the procedural posture of the claim of ineffective assistance may provide some guidance for the future. A defendant wanting to make such a claim has three potential choices, depending on the precise nature and basis of the claim. He can ask the district judge for a new trial on the basis either of what the trial record discloses about his lawyer’s performance or of extrinsic evidence — evidence that was not part of the trial record, such as an affidavit from his lawyer or from some other criminal lawyer concerning the character or motivations of his lawyer’s trial strategy and tactics.
United States v. Pinkney,
Are these exclusive choices? The first two are not, provided that the defendant does not need to present extrinsic evidence in order to demonstrate his lawyer’s ineffectiveness. The defendant can ask the district judge for a new trial on the ground that the trial record shows a denial of effective assistance, or he can ask us on the same ground (or he can do both). But in either case he faces an uphill fight. When the only record on which a claim of ineffective assistance is based is the trial record, every indulgence will be given to the possibility that a seeming lapse or error *418 by defense counsel was in fact a tactical move, flawed only in hindsight. It is no surprise that such claims almost always fail.
The defendant has a better chance if he files a motion under section 2255 and asks for an evidentiary hearing at which he can call his former counsel as a witness and try to show that those apparent lapses were real lapses — there was no tactical reason for them. But he is unlikely to get to first base if he already has presented a claim of ineffective assistance of counsel to the court of appeals unsuccessfully. It is true that the doctrine of res judicata is inapplicable in postconviction proceedings, such as those under section 2255,
Sanders v. United States,
If, however, presenting a claim of ineffective assistance on direct appeal is Scylla, could not failing to present it on direct appeal be Charybdis?
Norris v. United States,
But if, in the postconviction proceeding, the defendant turns out to have no extrinsic evidence to present in support of his claim of ineffective assistance — if his position is that the claim is established by the trial record alone — then it is a fair question why he has waited, and if he has no answer then
Norris
will bar the claim.
Cartee v. Nix,
Dently v. Lane,
The principle that emerges from these cases is that the defendant must make his ineffective assistance claim at the earliest feasible opportunity, but that if he wants to support the claim with facts that require evidence to establish he will be well advised to wait till the postconvietion stage and will be safe in doing so. In this connection we point out that the Norris bar, like the law of the case doctrine, is not absolute; it can be waived for good cause.
For completeness we note that the defendant can seek an evidentiary hearing by way of a motion for a new trial as well. But the new-trial route will require him to scramble in order to gather within seven days sufficient evidence to persuade the trial judge either to grant him a hearing or to grant him an extension of time for presenting the motion. We add that neither under Rule 33 nor under section 2255 is a hearing mandatory in all cases; if there is no reason to suppose that a hearing would produce evidence justifying the grant of a new trial, there is no reason to hold a hearing.
United States v. Pinkney, supra,
McDonnell coupled his motion for a new trial with a request for an evidentiary hearing, and this was a proper procedure to follow, as we have said. The judge denied the request; given the paucity of the showing made by McDonnell that Ackerman rendered ineffective assistance to him and by doing so probably caused him to be convicted unjustly, we cannot say that the judge abused his discretion in doing so.
Affirmed.
