UNITED STATES of America, Plaintiff-Appellant, v. Donnell YOUNG, Defendant-Appellee, 424 F.3d 499

OPINION RALPH B. GUY, JR., Circuit Judge. The government appeals from the order granting defendant Donnell Young’s motion to empanel separate juries to hear the guilt and penalty phases of his trial on capital and noncapital offenses. The district court found it had authority under 18 U.S.C. § 3593(b)(2)(C), on a showing of “good cause,” to decide to empanel a non-death-qualified jury to determine the question of guilt and then, if convicted of a capital offense, to empanel a second death-qualified jury to decide whether to impose the death penalty.

The government maintains that this dual-jury procedure violates the Federal Death Penalty Act, specifically, 18 U.S.C. § 3593(b), and that this court has jurisdiction to review the district court’s pretrial order either under the collateral order doctrine, or by treating the appeal as a writ of mandamus. After review of the law and the arguments presented on appeal, we VACATE the district court’s order and REMAND for further proceedings consistent with this opinion. I. This case began with a single-count, one-defendant indictment filed in March 1998.

A Second Superceding Indictment, filed in November 1998, expanded the drug charges in the original indictment to include defendant Donnell Young and nineteen other defendants (many of whom would ultimately enter guilty pleas). Shortly after being charged in this case, Young was transferred from a jail in Oklahoma City, Oklahoma, where he had been held for 18 months in connection with the murder of Woody Pilcher. Pilcher, an alleged coconspirator, was killed in Oklahoma City on August 2, 1997.

Capital murder charges were added against Young in a Fourth Superceding Indictment filed in September 1999, and repeated in a Fifth Superceding Indictment filed in September 2002. The Fifth Superceding Indictment charged 45 counts against five defendants: Jamal Shakir, Pacia Shakir, Donnell Young, Derrick Eatmon, and Eben Payne.

Jamal Shakir was charged with engaging in a continuing criminal enterprise, referred to as the “Shakir Enterprise,” which the district court described as “a faction of the Los Angeles-based street gang, the ‘Rollin’ 90s Crips,’ [that] allegedly transported and distributed controlled substances; used and possessed firearms; laundered money; and engaged in violent acts including assault, abduction, robbery, and murder.

The charges span a seven-year period from 1993 to September 1999 and allege wrongdoing in various communities, including Los Angeles, California, Oklahoma City, Oklahoma, Memphis, Tennessee, and Nashville, Tennessee.” United States v. Young, 376 F.Supp.2d 787, 789 (M.D.Tenn.2005). Young, who is from Los Angeles, claims not to have set foot in Tennessee before being taken into federal custody.

Young awaits trial on several drug and weapons offenses and three death-eligible counts related to the murder of Pilcher: specifically, (1) killing Pilcher in furtherance of a continuing criminal enterprise and drug conspiracy; (2) killing Pilcher to obstruct justice; and (3) causing the death of Pilcher by the use and carrying of a firearm during and in relation to a crime of violence or a drug trafficking crime (counts 37-39). In October 2002, the government gave notice of its intention to seek the death penalty against Young.

Severance was granted such that Young, Eatmon, and Payne would be tried jointly with each other, but separately from Jamal and Pa-cía Shakir. Because Payne was later found mentally incompetent to stand trial, Young (charged with capital and noncapi-tal offenses) currently awaits trial with Eatmon (charged with only noncapital crimes).

In a “Motion for Bifurcated Juries,” Young requested that the district' court empanel two juries — the first selected for the guilt phase without the process of “death qualification” and a second death-qualified jury selected for the sentencing phase in the event he is convicted of a capital offense. A hearing was held January 18, 2005, and the motion was granted in an opinion and order entered on April 22, 2005. Without articulating the rationale underlying its

interpretation of the statute, the district court concluded that it had discretion under 18 U.S.C. § 3593(b)(2)(

C) to decide before trial that “good cause” existed to empanel separate juries to decide the questions of guilt and punishment. In the alternative, the district court also found that the unitary jury requirement could be “waived” by the defendant’s motion to empanel separate juries. Although both of these conclusions had already been rejected by the Fifth Circuit in United States v. Williams, 400 F.3d 277, 281-83 (5th Cir.), cert. denied, — U.S. -, 125 S.Ct. 1611, 161 L.Ed.2d 289 (2005), the district court relied instead on the decision to the contrary in United States v.

Green (Green I), 343 F.Supp.2d 23, 30-32 (D.Mass.2004), a decision that was subsequently reversed by the First Circuit in United States v. Green (Green II), 407 F.3d 434 (1st Cir.2005). The district court went on to explain at length its reasons for finding that “good cause” existed in this case; including “case management” issues, concerns about the impact of death qualification on the racial composition of the jury, and social science evidence suggesting death-qualified jurors may be more prone to convict and may decide sentencing issues before the penalty phase. Young, 376 F.Supp.2d at 791-800.

Specifically, the district court found “case management” concerns established “good cause” to adopt a dual-jury procedure. First, judicial time and resources would be saved in this and related cases (which include two other capital defendants) by proceeding without a death-qualified jury because the court could avoid both extensive voir dire (or at least most of it) and the need to resolve the pending disputes concerning the aggravating factors asserted by the government before trial.

Time was found to be especially important in this case because charges had already been pending for more than seven years, and a number of codefendants who entered guilty pleas remain in custody but will not be sentenced until after they have testified. Besides, a sentencing hearing will be unnecessary if Young is acquitted on the capital offenses.

Finally, the court found there was a substantial difference in the evidence that would likely be presented in each phase; that the burden on the cooperating witnesses was lessened because multiple trials were already contemplated; and that steps, such as stipulations or summaries of evidence, may be used to reduce duplication of evidence. In addition, the district court found it highly likely that prospective black jurors would be disproportionately excluded for cause due to the higher rate of opposition to the death penalty among blacks.

Coupling this with the small percentage of blacks in the local jury pool, the district court found further good cause to empanel a guilt-phase jury without death qualification. The court felt this was especially significant here because Young is an African-American being prosecuted far away from his home in Los Angeles. As for the propensities of death-qualified jurors, the district court was troubled by research showing that 49.2% of capital-qualified jurors make the sentencing decision before the penalty phase.

The district court relied on other social science evidence indicating that death-qualified jurors may be less likely to believe in the presumption of innocence or may otherwise be more prone to convict in the guilt phase of the trial. Finally, the district court rejected the government’s contention that a defendant might lose the benefit of any “residual doubt” that a unitary jury might carry over from the guilt phase.

In the absence of research concerning the “residual doubt” effect, the district court concluded that the government’s interest in a unitary jury procedure was outweighed by other facts that create good cause to “bifurcate the jury.” The government appealed, and hearing on the matter was expedited at the defendant’s request. II. A. Jurisdiction Defendant concurs in the government’s assertion of appellate jurisdiction and, although jurisdiction may not be conferred by the consent of the parties, we agree that we may review the government’s challenge to this pretrial order.

First, under the collateral order doctrine, an interlocutory order may be immediately reviewable as a “final decision” under 28 U.S.C. § 1291 when the order (1) conclusively determines (2) an important legal issue completely separate from the merits of the action, which is (3) effectively unreviewable on appeal from a final judgment. Cohen v. Beneficial Indus. Loan Corp., 337 U.S. 541, 546, 69 S.Ct. 1221, 93 L.Ed. 1528 (1949); see also United States v. Caggiano, 660 F.2d 184-90 (6th Cir.1981) (allowing government appeal from order disqualifying United States Attorney’s office from participating in criminal case).

This exception to the final judgment rule must be strictly construed in criminal cases. Midland Asphalt Corp. v. United States, 489 U.S. 794, 799, 109 S.Ct. 1494, 103 L.Ed.2d 879 (1989); see also United States v. Yeager, 303 F.3d 661, 666 (6th Cir.2002) (no jurisdiction over defendant’s appeal from dismissal of charges without prejudice). In this case, the district court conclusively determined that a non-death-qualified jury will decide the question of guilt and that a second death-qualified jury will only be empaneled in the event Young is convicted of a capital offense.

The decision resolves the important legal issue of whether the Federal Death Penalty Act gives the district court discretion to adopt a dual-jury procedure on a pretrial finding of good cause; an issue that is completely separate from the merits of the underlying action. The issue is likely to recur as capital defendants, including others with charges pending before this district judge, are likely to argue for a dual jury process with a nondeath-qualified jury to hear the guilt phase of the trial. Finally, the government’s challenge is effectively unre-viewable after final judgment.

The government will be unable to appeal whether the defendant is found not guilty, is convicted but not sentenced to death, or is found guilty and sentenced to death by separate juries. See Sanabria v. United States, 437 U.S. 54, 69, 98 S.Ct. 2170, 57 L.Ed.2d 48 (1978); Fong Foo v. United States, 369 U.S. 141, 143, 82 S.Ct. 671, 7 L.Ed.2d 629 (1962). Construing the collateral order exception narrowly, we find the requirements are met in this case. Accord Williams, 400 F.3d at 280; but see Green II, 407 F.3d at 438 (declining to rely on collateral order exception, but granting mandamus relief).

Alternatively, there is also jurisdiction to review the appeal as a petition for writ of mandamus. Mandamus is a drastic remedy that should be invoked only in extraordinary cases where there is a clear and indisputable right to the relief sought. In re Parker, 49 F.3d 204, 206 (6th Cir.1995). This court has adopted a five-factor test for determining whether extraordinary circumstances warrant mandamus relief. Id. at 207. These factors are cumulative and must be balanced, and the absence of any factor is not controlling. Id.

The five factors are: 1) whether the party seeking the writ has no other adequate means, such as direct appeal, to attain the relief needed; 2) whether the petitioner will be damaged or prejudiced in a way not correctable on appeal; 3) whether the district court’s order is clearly erroneous as a matter of law; 4) whether the district court’s order is an oft-repeated error, or manifests a persistent disregard of the federal rules; [and] 5) whether the district court’s order raises new and important problems, or issues of law of first impression. 49 F.3d at 207.

Here, the government will be unable to appeal from this order after judgment is entered, and prejudice from having to present the case to separate juries in the guilt and penalty phases would not be correctable on appeal. There is also every reason to expect that this issue will recur in other capital cases, particularly given that this is the third case to present the issue in the courts of appeals. The district court’s

interpretation of the “good cause” exception in § 3593(b)(2)(

C) involves an important question of law and is a matter of first impression in this circuit. Lastly, for the reasons set forth below, we find the district court’s

interpretation of § 3593(

b) is erroneous as a matter of law. Accordingly, appellate jurisdiction exists over this matter whether under the collateral order exception or as a request for mandamus relief. Accord Williams, 400 F.3d at 280-81; Green II, 407 F.3d at 439-40. B. “Good Cause” under § 3593(b)(2)(

C) The Federal Death Penalty Act requires a separate sentencing hearing when the government seeks to impose the death penalty and specifies before whom the bifurcated penalty phase may be conducted. The question before us is the proper

interpretation of § 3593(b), which states as follows: (

b) Hearing before a court or jury. — If the attorney for the government has filed a notice as required under subsection (

a) and the defendant is found guilty of or pleads guilty to an offense described in

section 3591, the judge who presided at the trial or before whom the guilty plea was entered, or another judge if that judge is unavailable, shall conduct a separate sentencing hearing to determine the punishment to be imposed. The hearing shall be conducted— (1) before the jury that determined the defendant’s guilt; (2) before a jury impaneled for the purpose of the hearing if— (

A) the defendant was convicted upon a plea of guilty; (

B) the defendant was convicted after a trial before the court sitting without a jury; (

C) the jury that determined the defendant’s guilt was discharged for good cause; or (

D) after initial imposition of a sentence under this section, reconsideration of the sentence under this

section is necessary; or (3) before the court alone, upon the motion of the defendant and with the approval of the attorney for the government. A jury impaneled pursuant to paragraph (2) shall consist of 12 members, unless, at any time before the conclusion of the hearing, the parties stipulate, with the approval of the court, that it shall consist of a lesser number. This

section codifies the constitutional requirement that the guilt and penalty phases of a capital offense be bifurcated, and adopts, as a default rule, a requirement that a unitary jury must decide both guilt and punishment. This much is without dispute. We are called upon to interpret the exception in § 3593(b)(2)(C). The exception contains two prerequisites to the consideration of penalty-phase jury composition required by § 3593(b)(1) — (3). First, the government attorney must have filed the notice described in subsection (a). Second, the defendant must have either pleaded guilty or been found guilty. If the prerequisites are satisfied, § 3593(

b) requires “the judge who presided at the trial or before whom the guilty plea was entered, or another judge if that judge is unavailable,” to conduct a separate sentencing hearing before a penalty-phase jury described in § 3593(b)(1) — (3). The guilt prerequisite has not been satisfied in this case. The plain language clearly demands conclusion of the guilt phase prior to application of § 3593(b)(1) — (3). The court therefore should not have reached the composition of the penalty-phase jury under § 3593(b)(1) — (3).

Additionally, we find that use of the past-tense word “determined” in § 3593(b)(2)(C) “makes clear that the phrase ‘discharged for good cause’ refers to discharge for events arising after the guilt phase has been concluded.” Green II, 407 F.3d at 441 (emphasis in original); see generally Jones v.

United States, 527 U.S. 373, 418, 119 S.Ct. 2090, 144 L.Ed.2d 370 (1999) (“Discharge for ‘good cause’ under § 3593(b)(2)(C) ... is most reasonably read to cover guilt-phase (and, by extension, penalty-phase) juror disqualification due to, e.g., exposure to prejudicial extrinsic information or illness.”) (Ginsburg, J., dissenting); id. at 381, 119 S.Ct. 2090 (agreeing with dissent’s

interpretation of the statute) (majority opinion). The court in Williams similarly reasoned that: The “good cause” language pertains to discharging a jury that has already decided the defendant’s guilt. To reach this language, or the scenario contemplated by the statute, a motion would need to be made following a determination of guilt by the jury after trial.

For example, if the jury found the defendant guilty, and then, before the sentencing phase, certain members were disqualified because of their exposure to outside influences, a district court could entertain a motion to discharge the jury and could find “good cause” to grant such a motion. The provision does not allow a pretrial option for a bifurcated jury. 400 F.3d at 282; see also United States v. O’Driscoll, 250 F.Supp.2d 429, 430-31 (M.D.Pa.2001) (“Section 3593(b)(2)(

C) does not permit us at the beginning of the case to make a decision to empanel two juries. Instead some event must occur subsequent to a guilty verdict which would justify the discharge of the jury and the empaneling of a second jury to decide the penalty to be imposed.”). This

interpretation is reinforced by the structure of § 3593(

b) as a whole and by reading the exceptions in § 3593(b)(2) together. This

section provides first and in expressly mandatory language that the sentencing hearing shall be conducted “before the jury that determined the defendant’s guilt.” This unitary jury rule is followed by four conditional exceptions that allow for a new jury to be empaneled for the sentencing hearing “if’ one of four situations arises, including when “the jury that determined the defendant’s guilt was discharged for cause.” § 3593(b)(2)(C).

It also provides, as a third option found in § 3593(b)(3), that the defendant may essentially “waive” a jury in the penalty phase by moving, with the approval of the attorney for the government, for the penalty to be determined by the judge. The four exceptions enumerate the circumstances in which a district court is authorized to empanel a jury to hear the sentencing issue; each plainly involves a situation where there is no jury that can hear the penalty phase.

The “common denominator” is that “they all appear to represent situations in which it is either impossible or impracticable to apply the default rule of a unitary jury.” Green II, 407 F.3d at 441. That is, subpart (

A) and (

B) pertain to situations where there was no guilt-phase jury at all because there was either a guilty plea or a bench trial. Likewise, subpart (

D) applies when reconsideration of an imposed sentence is necessary, which arises almost certainly following direct or collateral review well after the jury’s discharge. Id. As the First Circuit so succinctly held in Green II, 407 F.3d at 441-42: Structurally, this leads to the conclusion that subparagraph (

C) should be read as referring to those situations in which empaneling a fresh penalty phase jury is unavoidable because of some exigency associated with, or arising after, the determination of the defendant’s guilt. This is the most natural (and therefore, the favored) reading of the statute. So construed, subsection (

b) presents a coherent, unified theme: a single, properly constituted jury will hear both phases of a federal capital trial unless circumstances definitively rule out that option. We find that the plain meaning of § 3593(b)(2)(

C) does not permit a pretrial determination that “good cause” exists to adopt a dual-jury procedure in federal capital cases. Without offering an alternative statutory

interpretation, defendant argues that the “good cause” exception must be viewed in the context of the district court’s broad discretion to control the proceedings and tempered by the heightened need for reliability in capital cases.

Defendant argues that by allowing for a separate penalty-phase jury under § 3593(b)(2)(C), Congress recognized that there may be circumstances when the guilt-phase jury should be discharged in furtherance of “higher interests.” While this much is apparent from the inclusion of the “good cause” exception, the question here is whether Congress left room for a district court to decide before trial that “good cause” exists to empanel separate juries to decide guilt and punishment.

In other words, the issue is whether the district court may make a pretrial determination that the case should not be tried before a unitary jury. The heart of defendant’s argument is that, in the absence of clear and unmistakable congressional intent to restrict the district court’s powers, the court should construe § 3593(

b) in a way that preserves the district court’s inherent authority to manage trials and select juries as it sees fit. Defendant analogizes this case to the inherent power of the federal courts to sanction a party for bad-faith conduct, which the Supreme Court held was not displaced by provisions authorizing the imposition of sanctions under 28 U.S.C. § 1927 and Fed.R.Civ.P. 11. Chambers v. NASCO, Inc., 501 U.S. 32, 111 S.Ct. 2123, 115 L.Ed.2d 27 (1991) (inherent power to control conduct may be invoked even though a procedural rule covers the same conduct).

Tellingly, the Court in Chambers concluded that the mechanisms for imposing sanctions under § 1927 and Rule 11 were not substitutes for the inherent power of the court because “that power is both broader and narrower than other means of imposing sanctions.” Id. at 46, 111 S.Ct. 2123. That is, Congress did not manifest an intention to supplant the district court’s power to impose sanctions to control the conduct of litigants. See also Link v.

Wabash R.R., 370 U.S. 626, 630, 82 S.Ct. 1386, 8 L.Ed.2d 734 (1962) (inherent power to dismiss sua sponte for failure to prosecute not clearly abrogated by language in Rule 41(

b) that appeared to require a motion from a party). It is also clear that the district court’s inherent powers may not be exercised in a way that conflicts with federal statutes or rules. See Bank of Nova Scotia v. United States, 487 U.S. 250, 254, 108 S.Ct. 2369, 101 L.Ed.2d 228 (1988) (holding supervisory power may not circumvent mandatory language of harmless-error inquiry required by Fed.R.Crim.P. 52(a)).

The Supreme Court has rejected reliance on the district court’s “inherent authority” as a basis for granting a motion for judgment of acquittal filed only one day after the time limit imposed by Fed.R.Crim.P. 29(c), explaining that the case law did “not establish any ‘inherent power’ to act in contravention of applicable Rules” and that the Court was “not at liberty to ignore the mandate of Rule 29 in order to obtain ‘optimal’ policy results.” Carlisle v. United States, 517 U.S. 416, 428 & 430, 116 S.Ct. 1460, 134 L.Ed.2d 613 (1996).

A statutory procedural rule framed “in terms of the mandatory ‘shall’ ... normally create[] an obligation impervious to judicial discretion.” Lexecon Inc. v. Milberg Weiss Bershad Hynes & Lerach, 523 U.S. 26, 35, 118 S.Ct. 956, 140 L.Ed.2d 62 (1998).

Even Fed.R.Crim.P. 57(b), which defendant points to as a codification of the trial court’s discretion, expressly provides that “[a] judge may regulate practice in any manner consistent with federal law, these rules, and the local rules of the district.” As the court in Green II explained: “The bottom line is this: where Congress has provided a specific panoply of rules that must be followed, the district court’s discretionary powers simply do not come into play.

Because this is such an instance, there is no reason to distort the plain meaning of the statutory text in an effort to preserve those powers.” 407 F.3d at 443. In response to the government’s argument that, under the district court’s reasoning, the exception threatens to swallow the rule, defendant contends that the district court’s findings rest on extraordinary case-specific circumstances that would recur in only the rarest of cases.

In particular, defendant emphasizes that the case management concerns included the outstanding disputes over the aggravating factors alleged by the government, the fact that charges had been pending for more than seven years, and the impact on cooperating codefendants who are still waiting to be sentenced.

Despite the reliance on these case-specific facts, the district court also relied on findings that a dual jury procedure could avoid much of the voir dire that would be necessary to death qualification and that no sentencing hearing would be required unless the defendant were convicted on one or more capital offenses. Both are true in all capital cases being tried to a unitary jury.

Moreover, the district court’s further concerns that a death-qualified guilt phase jury might be less diverse, more prone to convict, and more likely to prematurely decide punishment, are widely applicable in other federal death penalty cases. Defendant’s effort to limit the implications of the district court’s pretrial finding of “good cause” under § 3593(b)(2)(

C) is unavailing. C. Waiver Finally, the government challenges the district court’s alternative finding that the defendant could “waive” the unitary jury requirement. Defendant has abandoned the waiver argument on appeal, accepting that the district court’s authority to empanel separate juries under § 3593(

b) would not be dependent on the positions of the parties. We find the district court’s decision may not be sustained on the basis of waiver. As the First Circuit in Green II persuasively reasoned:

Section 3593 sets forth a set of general rules that govern all parties and the court itself. In those instances in which the statute does create rights that accrue to only one side or the other, the statute is explicit. For example,

section 3593(

c) specifies that the government “shall open the argument” and the defendant “shall [then] be permitted to reply.” So too

section 3593(b)(3), which makes specific reference to the ability of the parties jointly to waive certain of the rules (e.g., the right to a sentencing jury). The statute does not offer any such option with respect to the unitary jury rule of subsection (b)(1). The intentional inclusion of a waiver mechanism in one part of the statute persuasively indicates that the exclusion of such a waiver provision in another part of the same statute was intentional....

A federal capital defendant can no more waive the default rule of a unitary jury than the government can waive the rule that requires the court to instruct the jury to disregard the defendant’s race. If more were needed — and we doubt that it is — we would find it surpassingly difficult to believe that a statute that requires government approval of the defendant’s motion to dispense with a sentencing jury would sub silentio permit a defendant to obtain two juries as of right.

With respect to the deployment of a unitary jury in a federal capital case, “Congress intended to give no options, only commands.” Williams, 400 F.3d at 282. Green II, 407 F.3d at 443-44. Because we find the that § 3593(b)(2)(

C) may not be invoked pretrial on a finding of “good cause” and that the defendant may not “waive” the unitary jury requirement, the district court’s order is VACATED and the case is REMANDED for further proceedings consistent with this opinion. . The government responds that much of the guilt-phase evidence would have to be repeated to a separate penalty-phase jury and objects to the suggestion that it may be forced to accept the admission of summaries or stipulations during the penalty phase. See Old Chief v.

United States, 519 U.S. 172, 187-89, 117 S.Ct. 644, 136 L.Ed.2d 574 (1997) (holding government may not be forced to accept the defendant's offers to stipulate to evidence). . The advisory committee’s notes to Rule 57(

b) explain that the federal rules do not try to prescribe uniform practice as to matters of detail such as "the mode of impaneling a jury, the manner and order of interposing challenges to jurors, [or] the manner of selecting a foreman of a trial jury.” This case does not involve a detail that was left to the individual courts to regulate.