Retrial After a Hung Jury

Retrial After a Hung Jury

A jury that cannot reach a verdict produces the simplest of the endings that follow attachment of jeopardy, and the federal rules state the consequence directly. When the jury cannot agree on a verdict on one or more counts, the court may declare a mistrial on those counts, and the government may retry any defendant on any count on which the jury could not agree.¹ This is the first application of the idea set out in this chapter's introduction: the answer depends on why the first proceeding ended, and a deadlock is a failure of the process that leaves guilt undecided.² The four sections below take the rule apart piece by piece, so this section supplies the rule itself and the reason it is shaped as it is.

The reason starts with what a deadlocked jury has and has not done. Its members reached no decision, so the first proceeding produced neither an acquittal nor a conviction, and the government has not yet obtained the verdict it came to court for. In Richardson v. United States the defendant tried to block a second trial after a hung jury, and the Court held that the protection of the Clause applies only if there has been "some event, such as an acquittal, which terminates the original jeopardy."³ A mistrial declared because the jury was deadlocked is not such an event, so the second trial continues the first jeopardy rather than beginning a second one. The functional logic points the same way. A retrial after deadlock does not give the government a second try at a decision the first jury made; it supplies the decision the first jury could not reach.

That continuing-jeopardy reasoning looks to the end of the proceeding and asks whether anything terminated it. The manifest necessity line of cases, introduced in the last section, looks to the beginning: once a particular jury is sworn, the prosecutor may not try the defendant before another jury without showing a manifest necessity for the new proceeding.⁴ The two approaches reach the same result for a genuine deadlock. A jury that truly cannot decide leaves the court no way to obtain a verdict from that tribunal, and the mistrial ends nothing that the Clause protects.⁵

The fact that moved the answer is whether the jury produced a verdict at all. Acquittals are virtually always final and the government may not appeal one, which is why the second scenario closes the case, while the first leaves the original jeopardy running.⁶ Students lose this issue when they reason that the government has had its one chance and failed, because a hung jury is not an acquittal and the rule's own wording shows why: the mistrial is declared on counts the jury "could not agree" on, not on counts it decided.⁷ The rule also carries a built-in qualification, since retrial after a deadlock is permitted unless the court also enters an acquittal or its equivalent.⁸

The sections that follow take the pieces in order: why a deadlocked jury may be discharged at all, how a verdict on some counts limits what the government may retry, why a weak first trial does not by itself end the case, and why an informal report of the jury's votes is not a verdict. The first of these begins with the requirement of unanimity that makes a split jury a jury with nothing to report.

Deadlock as Manifest Necessity

The authority to retry after a hung jury begins with the way a federal verdict is made. The jury must return its verdict to a judge in open court, and the verdict must be unanimous.¹⁰ Because every juror's agreement is required for either outcome, one juror who cannot be persuaded can prevent a conviction, and one juror holding out the other way can prevent an acquittal. When the jurors split and stay split, they have therefore produced no verdict at all: no finding of guilt, no finding of innocence, and no default result in either direction. What the first trial produced is an impasse, and the question under the Double Jeopardy Clause is whether the government must accept that impasse as the end of its case. The answer has been no since 1824, and the reasoning starts with a single short opinion.

The foundation is United States v. Perez, in which the jury deadlocked, the court declared a mistrial without the defendant's consent, and Justice Story's opinion for the Supreme Court held that a second trial did not violate the Fifth Amendment.¹¹ The opinion reasoned that courts may discharge a jury from giving any verdict whenever, taking all the circumstances into consideration, "there is a manifest necessity for the act, or the ends of public justice would otherwise be defeated."¹² That formula is the manifest necessity standard, and it presupposes that the first jeopardy has already begun. In Crist v. Bretz the Court noted that Perez has long been understood to mean that jeopardy attached in the first trial and that a second trial was nonetheless allowed, because discharging the deadlocked jury was manifestly necessary.¹³ The standard therefore accepts that the defendant was once in jeopardy and identifies the circumstances in which a second trial may follow anyway.

United States v. Sanford applied the standard to a short set of facts. The defendants were indicted for poaching buffalo in Yellowstone National Park, and their first jury trial ended in a mistrial.¹⁴ While the government was preparing to retry them, the defendants moved to dismiss the indictment, the district court dismissed it, and the court of appeals affirmed on the theory that a second trial would put them twice in jeopardy. The Supreme Court agreed that jeopardy had attached in the first trial, but it pointed out that the trial had ended in a mistrial declared by the district court on its own motion, so the Perez test governed. Because the defendants did not dispute that the jurors were unable to reach a verdict, the case presented the very situation in Perez, and the Double Jeopardy Clause did not bar retrial.

The reason the defendant's interest in the first jury yields starts with unanimity. If a deadlock exhausted the government's opportunity, one holdout juror could in effect acquit any defendant, and the outcome would reflect no finding about guilt at all. The Court has described the result of such a rule as an insuperable obstacle to the administration of justice in cases that show no sign of the oppressive practices the Clause was written to stop, since a jury's failure to agree is an unforeseeable circumstance that makes completing the trial impossible.¹⁵ In Wade v. Hunter the Court therefore said that a defendant's "valued right to have his trial completed by a particular tribunal must in some instances be subordinated to the public's interest in fair trials designed to end in just judgments."¹⁶ A real deadlock is a choice made by no one. Neither the prosecutor nor the judge has engineered it, so the second trial does not become a device for the State to try again after a failure of its own making.

The government's side of the ledger supplies a second reason. In Arizona v. Washington the Court observed that courts have held without exception that the trial judge "may discharge a genuinely deadlocked jury and require the defendant to submit to a second trial," a rule that honors "society's interest in giving the prosecution one complete opportunity to convict those who have violated its laws."¹⁷ The word complete carries the point, since a trial that ends in deadlock has not delivered that opportunity: no jury ever decided the case. Richardson v. United States put the same idea in terms of entitlement, stating that the government, like the defendant, is "entitled to resolution of the case by verdict from the jury," so that jeopardy does not terminate when the jury is discharged because it cannot agree.¹⁸ Rule 31(b)(3) now states the federal result directly, permitting a mistrial on counts where the jury cannot agree and retrial of those counts.¹⁹

The same reasoning explains the limit that the word genuinely places on the exception. Deadlock is the clearest example of a case in which an impartial verdict cannot be reached, a circumstance that supplies manifest necessity.²⁰ A mistrial that fits no recognized ground is presumably improper, and a retrial after it is not permitted.²¹ The limit also removes a temptation. A prosecutor whose case is going poorly, or a judge with a crowded docket, would otherwise have a reason to describe the jury as hopelessly divided and start over before a fresh panel, which would turn a doctrine built for real impasse into a way out of a trial that is going badly. Requiring that the deadlock actually exist denies anyone that route. The standard governs mistrials that the defendant opposed, because a mistrial the defendant requested or consented to calls for a different analysis, taken up in a later section.²²

The fact that changed the result is the reason for the discharge. Elapsed time is not the dividing line: a short deliberation does not by itself make a deadlock false, and a long one does not by itself make it real, because manifest necessity turns on the circumstances of the particular case.²³ What the government needs is a jury that truly cannot agree. Students lose this issue when they assume that any mistrial after a failed deliberation permits retrial, because the deadlock must be genuine, and a discharge for the court's own convenience is no deadlock at all.²⁴

Partial Verdicts and Counts

A hung jury ends only what the jury left undecided, and the mechanism that makes this possible sits in the same federal rule that authorizes retrial. Rule 31(b) of the Federal Rules of Criminal Procedure lets a jury report agreement as it occurs instead of holding every result until the last question is answered.²⁶ With multiple defendants, the jury may return a verdict at any time during its deliberations as to any defendant about whom it has agreed.²⁷ With multiple counts, a jury that cannot agree on all counts as to a defendant may return a verdict on those counts on which it has agreed.²⁸ Each provision turns one agreement into a partial verdict, meaning a verdict that covers some defendants or some counts and leaves the others open. The unresolved remainder is then governed by the third provision, which limits what the government may do about it.

The reporting provisions and the retrial limit rest on different reasons, so each is worth stating separately. The reporting provisions answer a practical problem. A jury that could report only at the end of deliberations, and only as a package, would be discharged on a deadlock with its decided counts unannounced, so the jurors would have reached a result and the law would still record no verdict on it. Letting the jury report each agreement when it happens means that a decision actually made becomes a verdict, whatever becomes of the rest. The retrial limit answers a structural point about the mistrial exception itself. That exception exists because a deadlock produces no verdict, so it reaches only the places where a verdict is missing, and Rule 31(b)(3) says so in its terms by allowing the government to retry any defendant on any count on which the jury could not agree.²⁹ A count on which the jury did agree has a verdict, and the reason for allowing a second trial is absent for that count.

The limit also removes a temptation that would otherwise attach to every multi-count case. If a deadlock on one count reopened the whole indictment, a prosecutor who had lost a different count would have a reason to welcome a hung count elsewhere, since it would carry the lost count back into court for a second attempt. Confining retrial to the counts without a verdict takes that reason away, because a verdict already returned cannot be revived by a failure somewhere else in the case. The cleanest way to see the result is to run it on a small set of counts.

The same logic operates defendant by defendant when several defendants are tried together.

In both scenes the fact that fixed the result was whether a verdict existed for that particular defendant on that particular count. The unit of retrial is therefore the defendant-count pair, and a guilty verdict on a decided count remains in place alongside the retrial of a hung one, just as an acquittal does.³² Students lose this issue when they treat the mistrial as the end of the whole trial, because a mistrial on one count is not a mistrial on every count, and the retrial that follows is only as wide as the deadlock. A related question is whether a hung count in a weak prosecution should end the case on that count, and the next section takes up that argument.

Hung Jury Versus Insufficient Evidence

A defendant whose first jury deadlocked has a tempting next argument: if the government's evidence was thin enough that the jurors could not agree, the government should not get a second chance to fill the gaps. The argument borrows its force from Burks v. United States, which holds that when a conviction is reversed on appeal because the evidence was legally insufficient to support the verdict, the Double Jeopardy Clause bars retrial, since that reversal is equivalent to an acquittal.³⁴ The logic of the Burks bar is easy to state. A ruling that the evidence was legally insufficient means the prosecution had its full opportunity to prove guilt and failed, so a second trial would let the government repair a failure a court has already declared. The full treatment of Burks comes in the section on retrial after reversal. Here it serves as the contrast point, because the question is whether a hung jury produces the same kind of ruling.

In Richardson v. United States the defendant made exactly this argument. He contended that if the government failed to introduce sufficient evidence at his first trial to establish guilt beyond a reasonable doubt, he could not be tried again after a mistrial because of a hung jury, and he relied on Burks.³⁵ The Supreme Court rejected the claim. Burks involved a defendant who had obtained an unreversed appellate ruling that the evidence was insufficient, and it did not address a mistrial declared because the jury could not agree. The law governing hung juries has its own sources and logic, and retrial after a deadlock remains permissible regardless of how strong the first trial's evidence was.³⁶

The reason for the different treatment lies in what each situation tells the court. A Burks ruling is a resolution in the defendant's favor: someone with authority has decided that the prosecution's proof cannot support a conviction. A hung jury is the absence of any resolution. The jurors failed to agree, and that failure shows only that twelve people did not reach unanimity. It does not show that the evidence could not support a conviction. Richardson put the point in terms of when jeopardy ends, explaining that the protection of the Clause "applies only if there has been some event, such as an acquittal, which terminates the original jeopardy."³⁷ A discharge for deadlock is not such an event, so the second trial continues the first jeopardy and does not begin a new one.

Two practical consequences reinforce the rule. If thin evidence could turn a deadlock into a bar, courts would have to weigh the strength of the government's case after every hung jury, and a defendant would walk away with a judgment that no factfinder ever reached. The Court has also said that the government, like the defendant, is entitled to resolution of the case by verdict from the jury, and a deadlock leaves that entitlement unmet on both sides.³⁸ The weakness of the first trial is therefore an argument about what a jury might eventually do, and the Clause is triggered only by a ruling about what the evidence legally could do.

The contrast case shows how little needs to change for the answer to flip. In United States v. Martin Linen Supply Co. the jury hung and a mistrial was declared, but the trial court then granted the defense's timely motion under Federal Rule of Criminal Procedure 29(c) for a judgment of acquittal, and the Supreme Court treated that disposition as an acquittal that the government could not appeal.³⁹ The jury behaved the same way in Richardson and in Martin Linen. What differed was the presence of a judicial ruling, because an acquittal is a ruling that resolves some or all of the factual elements of the offense in the defendant's favor, and it counts even if the ruling was wrong.⁴⁰ That is why the hung-jury rule carries its own qualification: retrial is permitted unless the court also enters an acquittal or its equivalent.⁴¹

The question to ask on an exam stem is who, if anyone, ruled that the evidence was legally insufficient. An argument that the first trial looked weak does not answer it, and a hung jury is not a finding that the evidence was insufficient. If the stem reports only a deadlock and a mistrial, retrial is allowed whatever the strength of the first case. If it reports that a judge granted a judgment of acquittal, the analysis shifts to the acquittal rule.

Deadlock Reports and Polling

A defendant whose first jury hung may turn to a different argument, one built on what the jurors said while they were still deliberating. If the foreperson reported that the jury had voted against guilt on a charge, the argument runs, the government should not get to try that charge again. The answer depends on what ends jeopardy on a charge, and that event is a verdict. A verdict is the jury's final decision, unanimous and announced formally in court, and it is the only act in this setting that resolves a count.⁴⁵ A report of how the jurors currently stand describes an unfinished deliberation and resolves nothing.

The reason a tally falls short is that deliberation is a process of persuasion, and the law keeps each juror's position open until the verdict is final. A juror who votes against guilt on an early ballot may be convinced otherwise by the discussion that follows, or may convince others, and that exchange is the purpose of sending twelve people into a room together. If a mid-deliberation report operated as an acquittal, every such report would freeze positions the jurors were still free to revise. A judge who asked the jury to keep working would then risk reopening a count the law already regarded as closed. Treating the report as a snapshot keeps the deliberation from being cut short by its own progress reports, and it leaves the verdict as the single act that settles the question.

In Blueford v. Arkansas the defendant was charged with capital murder, an offense that carried the lesser offenses of first-degree murder, manslaughter, and negligent homicide, and the jury was instructed to consider each lesser offense only after finding him not guilty of the greater ones.⁴⁶ The jury first reported that it was unanimous against guilt on capital murder and first-degree murder and deadlocked on manslaughter, and, consistent with the instruction, it had not considered negligent homicide. The judge then gave an Allen instruction, which urges jurors to try to reach a verdict and sends them back to continue deliberating, and a short time later the jury reported that it was hopelessly deadlocked.⁴⁷ The Supreme Court held that the Double Jeopardy Clause did not bar retrial on capital murder or first-degree murder, reasoning that because the jury kept deliberating after reporting those votes, they were not final verdicts and remained subject to reconsideration.⁴⁸ The foreperson's report was accurate about where the jurors stood at that moment, but jurors who return to the jury room are free to change their minds, which is why the Court has treated foreperson statements reporting deadlock on charges as no bar to retrial.⁴⁹

Blueford also marks a boundary around an idea that a later section develops, namely that a jury's handling of greater and lesser charges can itself carry consequences for retrial. Whatever that idea reaches, it does not reach a report of votes made while deliberations are still under way.⁵⁰ The question here is only whether the jury's decision had become final, and a vote relayed to the judge before the jury went back to work had not.

The Federal Rules supply the procedure for testing whether an announced verdict is final. Under Rule 31(d), after a verdict is returned but before the jury is discharged, the court must poll the jurors individually if a party asks and may do so on its own.⁵¹ The poll answers a question the foreperson's announcement cannot, because a foreperson speaks for the group, while a poll asks each juror separately whether the announced result is that juror's own verdict. Because a verdict must be unanimous, a single juror who does not assent shows that the announcement overstated the agreement. The rule then gives the court two choices: it may direct the jury to deliberate further, or it may declare a mistrial and discharge the jury.⁵² In either case no final verdict ever existed, and a mistrial that follows rests on the same absence of a verdict that supports retrial after any genuine deadlock.

The polling hypothetical and Blueford turn on the same fact, which is whether the jury's decision was final when the statement about it was made. In Blueford the jurors went back to deliberate and ended deadlocked, so the earlier votes were never a decision. In the hypothetical, the poll either confirms unanimity and makes the verdict final or exposes a dissent and shows that no verdict existed. Students lose this issue when they treat a unanimous vote as a verdict, because a report of votes is not a verdict, and a jury that keeps deliberating after the report shows that the votes were still open.