A Fresh Look at Blackstone’s Solution
We can now say good-bye to same offense. The focus in this chapter is on trial procedure and outcome, rather than on the substantive question of when offenses are the same. The question is whether a second jeopardy is threatened or has occurred. This question is important, of course, only if the second jeopardy concerns the same offense; for purposes of this chapter and the next, we simply assume that the relevant offenses are the same offense. In most second-jeopardy cases, the charge at the second trial is identical to the first one, which satisfies all same-offense tests.
This chapter examines the Court’s approach to the twice-in-jeopardy problem and offers a case-by-case blameworthiness account as an alternative. For example, I argue that a hung jury based on insufficient evidence bars a new trial because the defendant’s lack of blameworthiness has been determined. The Court disagrees. In chapter 9 I offer a more global legislative-prerogative account of second jeopardies, focusing on the extent to which explicit legislative action can change traditional concepts of verdict and appeal.
Madison’s original language prohibited more than one trial for the same offense. This language presumably prohibits more than one exposure to verdict, and the reason for terminating the exposure is unimportant. But the Clause as adopted and ratified does not forbid more than one trial or exposure to verdict. It forbids more than one jeopardy. Until we know when one jeopardy is over, we do not know when the bar arises.
Blackstone encountered no difficulties here because the eighteenth-century rule about terminations of jeopardy was elegant: the pleas in bar arose only after a prior acquittal or conviction.1 This elegant rule worked in Blackstone’s day because juries were required to deliberate until they reached a verdict in capital cases and could be discharged in noncapital cases only if both parties consented.2 Blackstone’s rule is manifested in several state constitutions.3 Our more permissive modern procedure about jury deliberation demands a rule about early terminations. But the rule could still be Blackstone’s, which is that early terminations are just not jeopardy outcomes. What is wrong with returning to Blackstone’s formulation?
The answer is: (almost) nothing. The task of this chapter is to justify, and explain, that answer. In chapter 3 I argued that the conventional wisdom was wrong about Justice Story’s famous mistrial opinion in United States v. Perez.4 Story did not state a rule for reviewing trial court mistrials. Instead, Story stated that defendants cannot prevail on a double jeopardy claim unless they can show an acquittal or conviction for the same offense.
Verdicts end jeopardy because verdicts are the blameworthiness outcome the process is designed to produce. At that point, a rational system would have no interest in redetermining blameworthiness. In individual cases, the state’s interest in redetermining blameworthiness might be rational, even compelling, but the system has a whole would collapse under its own weight if some principle did not generally constrain redeterminations. That principle, embodied in the Code of Hammurabi, is a broadly stated res judicata rule.
As a general matter, there is no reason to believe that fact-finder #2 will render a more accurate judgment about blameworthiness than fact-finder #1 and thus no reason to permit #2 to proceed, at least when the question is a systemic one. In effect, the double jeopardy rule merely embodies in the Constitution that rational system of res judicata, presumably to prevent individual instances of behavior inconsistent with the rational systemic model.
But a second determination of blameworthiness is impossible without a first determination. Two easy examples exist of this principle, both of which have been part of the common law since 1200: conviction and acquittal.
Verdicts are the end of jeopardy, as they were in the king’s courts in the thirteenth century and five hundred years later when Blackstone wrote his Commentaries, but only if they remain undisturbed. If a defendant elects to attack a conviction and if the conviction is vacated, no verdict remains to serve as a Blackstonian plea in bar of another trial. The question of an acquittal remaining undisturbed is more complex than it first seems, and the next section begins with that issue.
Defendants do not, of course, appeal acquittals, and the federal government and most states do not permit a prosecution appeal. A largely ignored issue is whether double jeopardy prevents a prosecution appeal. The Court held, near the beginning of the twentieth century and over Holmes’s dissent, that the Double Jeopardy Clause bars a federal appeal of an acquittal.5 In a state case in 1986, the Court rather casually stated that double jeopardy also barred a state appeal of an acquittal if a reversal would “lead to proceedings that violate the Double Jeopardy Clause.”6
James Strazzella questions whether it is double jeopardy or the procedural law that prohibits the appeal of an acquittal.7 As he correctly notes, the acquitted defendant is not exposed to a second jeopardy by virtue of an appeal. A double jeopardy issue arises only if the acquittal is vacated and the defendant is forced to trial a second time. One way to conceptualize the issue is that double jeopardy bars a second trial in the face of an acquittal, and the procedural law bars a futile appeal.
The distinction is important in two kinds of cases. In the first kind, about which Professor Strazzella wrote, a few states permit the prosecution to appeal, typically to obtain “advisory rulings limited to future cases.”8 It is important for that category of cases to understand that the Court has never held that the Double Jeopardy Clause bars prosecution appeal when the only effect would be an advisory opinion. In the second category of cases, a legislature might permit a prosecutor to appeal trial court rulings that led to an acquittal and, if the trial judge were overruled, to retry the acquitted defendant with a correct set of legal rulings. Would the Double Jeopardy Clause permit this system? The next chapter considers this fundamental question.
But in describing the current state of the law, it is safe to say that an acquittal bars a second trial for the same offense, and the only real issue is to determine which outcomes are acquittals. Two kinds of questions arise. First, perhaps an acquittal obtained by fraud is not really an acquittal. Second, perhaps an acquittal is not really an acquittal if the trial judge acts arbitrarily or erroneously and deprives the state of its chance to obtain a conviction.
The most obvious way in which an acquittal may not be an acquittal is if the defendant bribes the judge to return a not-guilty verdict as trier of fact.9 Ann Bowen Poulin argues that courts should ignore the acquittal and permit a second trial “if the prosecution establishes, beyond a reasonable doubt, that the acquittal was a product of a corrupt [judicial] process.”10 In these cases, the defendant has relinquished “any claim to a protected interest in the finality of the fact finder’s verdict or in the termination of the stress and anxiety of the trial.”11 Professor Poulin argues that a similar exception should not be recognized when a defendant corrupts members of the jury; here, the corruption of the process is less profound because, presumably, some or most of the jury are not corrupted.
David Rudstein recognizes that an acquittal obtained by fraud is a corruption of the process but argues that it happens so infrequently that “the game hardly seems worth the candle.”12 The “candle,” as Professor Rudstein sees it, is the potential for harassment of acquitted defendants and the loss of finality of acquittals. The state could come forward years later, allege fraud, and force an acquitted defendant, at a minimum, to defend the charge of fraud.
As I noted in chapter 1, it seems unlikely that the legislature would want to give finality to an acquittal in the limited set of circumstances carved out by Poulin: when the state can prove beyond a reasonable doubt that the defendant corrupted the judge. Moreover, I can draw support from Blackstone here, as Rudstein acknowledges. In one statement of autrefois acquit, Blackstone limits it to cases “when a man is once fairly found not guilty upon any indictment.”13 Though these cases arise too infrequently to make much difference, the better rule for a legislative-intent account is Poulin’s.
In Sanabria v. United States,14 the trial judge excluded evidence on the erroneous belief that the indictment did not cover the proferred evidence. An alternative to excluding evidence because it was improperly alleged in the indictment would have been to dismiss the indictment because it improperly alleged one of the elements. Had the judge taken that option, the dismissal would not have been the end of jeopardy.15 But Sanabria held that an acquittal was an acquittal and could not be appealed.16
Westen and Drubel criticize this distinction, arguing that the two errors are “functionally identical.”17 That is the problem with legal-realist policy analysis; a clever advocate can make almost any two things appear functionally the same. In this case, Westen and Drubel focus on the procedural similarity: both cases involve defects in the charging document and consequent dismissal at the defendant’s behest. But the only function that should determine the end of jeopardy is substantive: if the judge merely dismissed the indictment for a procedural flaw, he would not have resolved the elements of the offense in the defendant’s favor. Once we free ourselves from interest balancing and policy analysis, it is sheer sophistry to equate a dismissal where the trier of fact has found the case unproven18 with a case in which the judge has found an element improperly alleged in the indictment.
Sanabria can be fully defended, however, only if we can explain why acquittals, but not convictions, are an absolute bar to further proceedings. Westen suggests that acquittals are an absolute bar because acquitted defendants enjoy a unique type of “relief and desire for repose” and because double jeopardy protects jury nullification.19 Another explanation is that acquitted defendants do not appeal their verdicts. On this Black-stonian view, convictions and acquittals are identical double jeopardy events when rendered. They acquire a different procedural status when a defendant chooses to appeal a conviction. If the conviction is vacated, the state is permitted to begin again, unless something about the reversal determines the blameworthiness issue in the defendant’s favor.
In effect, Blackstone’s verdict rule works as follows: when a defendant offers a double jeopardy defense to a trial, the court determines whether there is an extant verdict for the same offense (an acquittal, an undisturbed conviction, or a conviction overturned in a way that makes it substantively an acquittal). Looking at the end of the appeal process to see whether a verdict exists to bar another jeopardy provides a better explanation of the disparate treatment of acquittals and convictions than Westen’s views about nullification and repose. As to repose, it is not clear why double jeopardy should protect particular expectations about future proceedings. As to nullification, it is not clear why double jeopardy, and not the Sixth Amendment right to a jury trial, is the protector of nullification.
What is clear is that Blackstone would have insisted on an extant verdict to bar a second jeopardy. That still seems the best explanation for why convictions are today treated differently from acquittals. Some trial outcomes that are not formal verdicts can be “acquittal equivalents,” but I defer the discussion of these outcomes until we have disposed of formal verdicts. Acquittals pose no problem, at least outside the corrupted-judge context. Whether rendered by judge or jury, whether based on a fair interpretation of the evidence or a raw exercise in jury nullification, an un-corrupted acquittal is the end of any “life or limb” proceedings by the state against the defendant for the same offense. (As the world saw, with some misgivings, in the O. J. Simpson civil case, and as chapter 4 made plain, an acquittal in a criminal case does not bar a civil suit.) Of the two kinds of formal verdicts, only convictions present issues yet to be resolved.
Criminal procedure was far simpler in Blackstone’s day. If a defendant made a double jeopardy plea in bar, a judge merely had to determine whether a prior trial had ended in an acquittal, a conviction with a judgment, or a conviction without a judgment. If so, the defendant could successfully plead autrefois acquit, autrefois attaint, or autrefois convict.20 If no verdict existed, the plea was no good.
If one ignores the modern procedural pertubations (the conditional guilty plea, appeal, motion for a new trial following conviction, and dismissal after a conviction), convictions produce the same result today. When Nathaniel Brown was faced with an auto theft charge based on the same taking for which he had already served thirty days in jail, he raised a plea of former conviction. And, as we have seen, the Supreme Court held that double jeopardy barred Ohio from prosecuting Brown for the same offense.
This part of double jeopardy has not changed for eight hundred years (even longer if we include Greek and Roman law) because it is the core of double jeopardy. Infused with notions of res judicata, the idea that one verdict on an issue ends the defendant’s jeopardy is both fair and procedurally necessary. As noted earlier, any system that determines outcomes must have a limit on the number of times the same question can be considered. And “one” is the only fair limit when “life or limb” is at risk.
Thus, the only double jeopardy issue that arises when a defendant relies on an undisturbed conviction (or acquittal, for that matter) is whether the new proceeding is for the same offense as the one that produced the verdict. But, once a conviction is disturbed, new issues arise.
In Blackstone’s day, a convicted defendant could seek to have the judgment “falsified, reversed, or avoided” by showing that the original judgment was rendered in some way that made it voidable. The example Blackstone gives is a judgment entered by those who lacked the authority to do so.21 A convicted defendant could also petition the king’s courts from the inferior courts by a writ of error that showed “notorious mistakes in the judgment or other parts of the record,” such as a misdescription of the accused, the sheriff, or the court.22 Leaving aside the technical (and outmoded) judgment of outlawry, the effect of reversing or falsifying the judgment of conviction was that “all former proceedings are absolutely set aside, and the party stands as if he had never been at all accused.”23 In Blackstone’s day, as in my account, the accused “still remains liable to another prosecution for the same offence; for the first being erroneous, he never was in jeopardy thereby.”24
This follows logically from Blackstone’s notion that a verdict is a plea in bar. Remove the verdict, and you remove the plea in bar. While this is not the modern English rule,25 it is the modern rule in the United States. In United States v. Tateo26 the Court reaffirmed an earlier precedent and held that a reversal clears the way for a new trial. The Tateo Court acknowledged the contrary modern English rule but rejected it largely on policy grounds. The Court wrote: “It would be a high price indeed for society to pay were every accused granted immunity from punishment because of any defect sufficient to constitute reversible error in the proceedings leading to conviction.”27
A more Blackstone-like explanation can be found in the case on which Tateo relied. This 1896 case is also useful in drawing a distinction between convictions and acquittals as verdict bars. Two defendants in Ball v. United States28 had been convicted in an earlier trial, and their convictions had been set aside on appeal because of a defect in the indictment. At the new trial, the trial judge overruled the plea of former jeopardy. The defendants were convicted and sentenced to death. Their claim in the Supreme Court was that a verdict ended jeopardy, whether or not it is overturned on appeal for error.
In deciding that question the Ball Court made a distinction that has rarely been noted. Westen and Drubel urge a different rationale, one drawn from the Court’s later explanation of Ball: “Given the difficulty of affording a defendant a completely error-free trial the first time around, the State is ‘excused’ for committing the error and allowed to assert its interest in law enforcement over the defendant’s interest in finality.”29 If one uses this rationale, one gets the right answer in Ball on the conviction issue—the indictment error can be excused and the government permitted to assert its interest in law enforcement. The government can disregard the outcome of the flawed proceeding and begin again. The problem with this characterization is that it cannot explain the acquittal issue also raised by Ball.
There was a third defendant in Ball, Miliard Filmore Ball,30 who was found not guilty by the jury in the initial trial. Since he was charged and tried under the same indictment as the other two defendants, the excusalof-error rationale means that the government could disregard the outcome of his flawed proceeding—the acquittal—and begin again with respect to him, too. In Ball, we do not have to argue about whether one error is more excusable than the other: it was precisely the same error in the same indictment that infected both the convictions and the acquittal. If the error justified overriding the convicted defendants’ finality interest in the first outcome, it also, without an additional parameter, overrides Miliard Filmore Ball’s interest in the first outcome.
But the Court held in Ball, after a rather lengthy and somewhat defensive analysis distinguishing English authorities, that Miliard Filmore Ball could not be reprosecuted. Thus, another parameter is needed. Westen and Drubel read the Ball opinion as suggesting that the most significant interest in the double jeopardy hierarchy of (legal realist) values is a defendant’s interest in jury nullification. On this reading of Ball, the additional parameter is that acquittals by a jury are categorically different from any other outcome.
Asserting a jury nullification parameter, as a brute fact, does harmonize the results in Ball, but it is historically and conceptually unsatisfactory. The historical flaw is that neither Blackstone nor any of the other English authorities ever favored the finality of acquittals over the finality of convictions.31 Moreover, support for this distinction in the Ball opinion results from a modernist misreading. Although the Ball Court stressed the jury finding of factual innocence, its purpose was to distinguish the English rule that a void indictment never gave rise to a jeopardy bar. The Ball Court asserted that the English doctrine of void indictments applied only to outcomes not based on factual innocence.
The distinction Ball made between the convictions and the acquittal is not based on acquittals being categorically different, or even on some abstract view of Miliard Filmore Ball’s factual innocence; the distinction is based on what Miliard Filmore Ball did following the outcome—nothing. The Court noted several times that what rendered the convicted defendants subject to reprosecution was that they had chosen to erase the conviction. For example, “[I]t is quite clear that a defendant who procures a judgment against him upon an indictment to be set aside may be tried anew.”32 In an even more instructive passage, the Court distinguished three situations: [1] the English void verdict rule; [2] the convicted defendant; and [3] the acquitted defendant. The point to distinguishing [1] is to reject the English rule that an acquittal on a void indictment cannot be used as a jeopardy bar:
[1] But, although the indictment was fatally defective, yet, if the court had jurisdiction of the cause and of the party, its judgment is not void, but only voidable by writ of error, and until so avoided cannot be collaterally impeached. [2] If the judgment is upon a verdict of guilty, and unreversed, it stands good, and warrants the punishment of the defendant accordingly, and he could not be discharged by writ of habeas corpus. [3] If the judgment is upon an acquittal, the defendant, indeed, will not seek to have it reversed, and the government cannot.33
A clearer statement of the relevant principle is difficult to imagine. It has nothing to do with jury nullification and everything to do with the defendant’s election to stand on or to reverse the initial outcome. The defendants in Ball were treated differently simply because Miliard Filmore Ball took no action to upset his acquittal, while the other defendants sought and obtained reversals of their verdicts. I have previously used “verdict election” to describe this principle,34 which is consistent with the Court’s explanation of Ball.35
Not only is the Ball Court’s description clear; the concept of verdict election is self-derivable from the way the criminal process treats verdicts. Defendants X and Y appeal their convictions and obtain a reversal. No verdict now exists that manifests the first jeopardy. No barrier exists to a reprosecution. Blackstone’s common law produces the same result. But defendant Miliard Filmore Ball does not appeal his verdict. An extant verdict manifests the first jeopardy; it matters not whether that verdict is an acquittal or a conviction. An extant verdict is an insurmountable barrier to a new trial. On this view, no double jeopardy issue arises in a mistrial case where no determination of guilt or innocence can be found in the outcome. Using blameworthiness to explain the second jeopardy aspect permits a return to Blackstone, to Story’s Commentaries, and to Story’s reading of Story’s opinion in Perez.36
By contrast, Westen and Drubel’s legal realism requires that their hierarchy of double jeopardy values be accepted and then balanced in a complex process that is more an ad hoc rationalization of the Court’s doctrine than a helpful analytical tool.37 Westen and Drubel confuse matters further when they seek to harmonize the mistrial and the appeal cases by use of their jury nullification value. They argue that the same rationale explains both categories of cases—whether the prosecution was manipulating error in the first trial to avoid an acquittal. Manipulation is one explanation of the mistrial doctrine, as we see later,38 but it has nothing to do with appeals. It does not explain why Miliard Filmore Ball was treated differently from his codefendants. More important, it is a question that need not be answered. If one takes the Blackstonian view that mistrials have nothing to do with cases involving verdicts, one is not forced on the Procrustean bed of trying to harmonize two very different lines of cases.
Once again, a careful look at what the earlier courts and commentators thought about double jeopardy provides a more satisfactory explanation of doctrine than a set of vague policies and values. Defendants who successfully appeal their convictions can be reprosecuted because no verdict exists as barrier. That rule does not apply to defendants who do not appeal their verdicts, which includes many convicted defendants as well as acquitted defendants. The double jeopardy world was still simple when Ball was decided in 1896, and it made a lot more sense than the Westen and Drubel world of 1978.
There are two double jeopardy questions involved with guilty pleas. First, how does double jeopardy doctrine operate as a framework within which prosecutors and defense counsel bargain for repose and finality? Stated differently, how do the various double jeopardy rules and exceptions—for example, dual sovereignty—affect the process by which guilty pleas are negotiated? This intensely pragmatic and important question is beyond the scope of the present book, but Daniel Richman’s examination is both thorough and thoughtful.39
The second question is theoretical: are the consequences of a guilty plea conviction different from those of a conviction after a trial? The answer is no and yes. For the typical guilty plea conviction, the procedural process by which it is obtained is irrelevant. If the defendant does not succeed in disturbing the conviction, it stands as a bar to another proceeding to the same offense.
But what happens when the guilty plea follows an earlier verdict for the same offense? If the defendant is permitted to appeal his guilty plea, the underlying conviction must be vacated. But the rule for appealing guilty plea convictions is very strict—about the only errors that can be raised on appeal are errors in the voluntariness of the guilty plea or in the procedure that led up to the plea.40 So the issue is whether a defendant can appeal on the ground that the guilty plea conviction violated the Double Jeopardy Clause.
The Court permits this kind of appeal.41 Double jeopardy violations are one of a very small subset of errors unrelated to the plea itself that can be raised on appeal.42 The reason for this exception, according to the Court, is that the double jeopardy violation deprives the state of the authority to bring a second proceeding. If the state lacked authority even to file charges, the entire process is tainted, and no reason exists to give preclusive effect to the guilty plea. This comports with Blackstone’s view of autrefois attaint (a common law plea now merged into the doctrine of former conviction): all proceedings after an accused had been attainted were superfluous. Similarly, with an extant verdict: all proceedings based on the same offense are superfluous and may be disregarded.43
A second procedural wrinkle involved with guilty-plea convictions is that defendants can promise to do certain things in the future in exchange for a guilty plea to a lesser offense. If the defendant fails to live up to the contractual agreement, what are the double jeopardy consequences? Again, if one examines this issue in the abstract, without worrying about fairness to a particular defendant, the answer seems pretty clear. Assuming that guilty pleas can be made conditional—which is not a double jeopardy issue—and assuming that the defendant failed to do what he promised, the contract would give the state the authority to vacate the underlying conviction. Once the underlying conviction has been vacated, the defendant is in the same position as if the conviction had been reversed on appeal.
The Court reached this result in Ricketts v. Adamson.44 The agreement stated that in the event of breach, “the parties shall be returned to the positions they were in before this agreement.”45 Adamson pleaded guilty to second-degree murder in return for his promise to testify against others involved in the murder. When he refused to testify in a retrial of the others, the state court vacated the second-degree murder conviction and reinstated the original charges. He was subsequently convicted of first-degree murder and sentenced to death. The Supreme Court upheld the conviction and sentence.
One might draw a distinction between Adamson and the typical defendant who appeals his conviction and obtains a reversal. This distinction would make something of the typical defendant who exercises the appellate procedure to disturb his conviction, while Adamson did nothing. He did not seek legal action to disturb his conviction; indeed, he stood on the conviction as a jeopardy bar.46 But ultimately this distinction does not hold. Adamson did do something to upset his conviction. His refusal to execute the agreement was a legal action for which he could be held accountable.47 He chose a path, with legal consequences, just as the typical defendant who successfully appealed his conviction.
The final wrinkle in guilty-plea convictions intersects the acquittal doctrine. An acquittal is a manifestation of lack of blameworthiness, and a conviction can sometimes function as a partial acquittal. If, for example, a defendant is charged with first-degree murder and convicted of second-degree murder (the facts of Green v. United States),48 the fact-finder has implicitly acquitted the defendant of first-degree murder. Should the defendant obtain a reversal of the second-degree murder conviction, retrial is permitted for that charge but not for the higher charge of which he has been implicitly acquitted. So held in Green.
One might seek to expand the Green doctrine to guilty-plea convictions, but the analogy is inapt. When the prosecutor accepts a guilty plea to a lesser offense, there is no fact-finding about the blameworthiness of the defendant for the greater charge. Thus, as the Court has held, a guilty-plea conviction does not implicitly acquit of any other offense.49
When trial judges enter more than one conviction, defendants can raise a double jeopardy issue on appeal by claiming that two or more of these convictions are for the same offense. If the defendant is wrong, that’s the end of the story. If the defendant is right, appellate courts must decide which convictions to vacate. This issue turns out to be mostly about sentencing.
We should begin with what may be obvious but was not decided until 1985. More than one conviction for the same offense is a violation of double jeopardy without regard to the sentences imposed on the multiple convictions (or whether any sentence at all was imposed). The only possible remedy for a violation of the double jeopardy multiple-punishment doctrine is to vacate the offending convictions.50
The issue arose because courts adopted strategies to avoid vacating convictions (probably from a desire not to have to retry a defendant if one conviction was subsequently overturned). Courts sometimes left the offending convictions intact and ordered all but one sentence to be vacated. Alternatively, courts sometimes left the convictions intact and ordered the sentences to be served concurrently.51 But, in 1985, the Supreme Court held that the offending legal event was not the existence of multiple sentences but the existence of more than one conviction for the same offense.52
This is right on a Blackstone view because even a conviction without judgment permitted a defendant to claim the plea of former conviction in 1764. If one accepts the analogy between multiple convictions in one trial and multiple trials,53 then the existence of more than one conviction must offend double jeopardy in every context. Unfortunately, this recognition does not solve all the problems about how to remedy multiple convictions when the error is discovered on appeal.
While a full discussion of this highly technical issue is beyond the scope of the current work,54 the problem can be sketched here. Suppose the trial judge enters convictions for offense A, offense B, and offense C. The judge imposes five-year sentences for all convictions and orders the sentences for A and B to be served concurrently, with the sentence for C to be served consecutively. The net effect of this sentencing package is that the defendant faces a ten-year sentence, with a concurrent term of five years for the first part of the sentence. Now assume that the appellate court determines that the conviction for C violates the double jeopardy multiple-punishment doctrine. What is to be done?55
If the appellate court mechanically vacates the conviction for C and affirms the other two, the result is that the defendant faces two concurrent sentences of five years each, half as long as the sentence the trial judge meant to impose. As appellate courts lack the authority to resentence, the only remedy that permits the total sentence to be what the trial judge intended is to vacate all convictions and remand for resentencing. But assuming the defendant asked only for conviction C to be reversed, it is not clear that appellate courts possess the authority to reverse the others. And there are potential due process problems that are not discussed here.56
The Supreme Court has never spoken to this issue. It may depend more on notions of procedural finality in appellate review than on double jeopardy. Judges can avoid the problem by making sure that one conviction carries a sentence that reflects the judge’s view of the defendant’s blameworthiness. If this is not possible, because no one offense permits a sentence of that length, judges can order each sentence consecutive with each other (and the pairs concurrent with other pairs) to ensure that the sentencing package cannot be changed by an appellate reversal of one sentence. In the hypothetical, for example, the judge could have ordered A served consecutively with B, and B served consecutively with C, and A served consecutively with C, and all three pairs to be served concurrently.
The sentencing issues discussed so far arise only when more than one conviction is entered. Are there double jeopardy sentencing problems when only a single conviction is entered? Not in Blackstone’s day, but recent modern complexities in sentencing, particularly the federal sentencing guidelines, have opened up potential double jeopardy claims. So far, consistent with Blackstone, the Supreme Court has failed to find any double jeopardy problems associated with sentencing as long as there is only one conviction per same offense.
Some simple matters first. There are no double jeopardy problems in resentencing under a single conviction after an earlier conviction has been reversed on appeal and the defendant reconvicted. The sentencing judge must give the defendant credit for time already served but may otherwise impose any sentence permitted by law.57 Though there are potential due process problems here,58 there is no double jeopardy problem, the Court has told us, because the “slate has been wiped clean. The conviction has been set aside and the unexpired portion of the original sentence will never be served.”59 This manifests a Blackstonian view of second jeopardies.
So double jeopardy permits judges to resentence at will after a conviction that follows an appellate reversal. But what if the relevant sentencing guidelines require the prosecutor to disclose, and the judge to consider, uncharged conduct that constitutes a crime? If that uncharged conduct results in a sentence enhancement, can the prosecution later bring a prosecution based on the uncharged conduct? Yes, the Court held in Witte v. United States.60 Sentencing enhancements “do not punish a defendant for crimes of which he was not convicted, but rather increase his sentence because of the manner in which he committed the crime of conviction.”61
This perhaps abstract distinction follows from the Court’s century-long treatment of habitual-offender crimes. Habitual-offender crimes punish a defendant more severely if he has been previously convicted of other crimes. In a way, the defendant is being punished again for the past crimes, but the Court has consistently held that the legislature may view the past crimes as evidence suggesting that the defendant is more dangerous, more in need of reform, and thus a better candidate for a longer sentence. In the Court’s words, the increased penalty “is a stiffened penalty for the latest crime, which is considered to be an aggravated offense because a repetitive one.”62
The Court’s treatment of sentence enhancment fits nicely with the legislative-prerogative view of double jeopardy developed in this book. As long as the legislature is clear that it wants the additional penalty, which is clear in the habitual-offender laws, there is no double jeopardy reason to oppose the additional penalty. Indeed, a legislative-prerogative theory would also permit the government to use uncharged conduct for which the defendant had been tried and acquitted as long as the legislature made clear that it authorized this outcome. There is nothing magic about an acquittal that overrides the legislature’s ability to configure a procedural system that uses conduct to determine blameworthiness, as long as the prosecutor has to prove the conduct in the new proceeding. This, of course, follows from my Blackstonian view that only verdicts have jeopardy status. The Supreme Court agreed, in United States v. Watts,63 holding that the conduct underlying an acquittal may be used to enhance a sentence under the federal sentencing guidelines.
To be sure, there are difficult sentencing issues involving due process and construction of sentencing guidelines, but the only Model 3 double jeopardy sentencing requirement is that a trial judge cannot impose more than one conviction for the same offense. Because Congress has denominated enhancement as something other than a verdict, it does not matter that the enhancement is based on conduct for which the defendant was convicted or acquitted. On this issue, the Court has closely paralleled the Blackstone blameworthiness account of double jeopardy. Other nuances appear when we consider the concept of “acquittal equivalence” in the next three sections.
There are two categories of outcomes that lack the status of a formal verdict: mistrial and dismissal. The procedural distinction, roughly, is that a mistrial keeps the original indictment or other charging instrument in place and thus contemplates a retrial. A dismissal, as its name suggests, dismisses the indictment or charging document and is indifferent as to retrial. For a few years, the Court seemed to intimate that a substantive jeopardy distinction might exist between these outcomes. But the only distinction the Court has made so far is that dismissals, but not mistrials, can be viewed as equivalent to an acquittal.64 A jeopardy bar arises, for example, if a judge dismisses an indictment because the state has failed to prove that the defendant was sane at the time the crime was committed.
In chapter 2 I argued that jeopardy policy has nothing to do with preventing harassment and everything to do with preventing unauthorized judgments, which include convictions and dismissals of defendants who are entitled to acquittals.65 Ireland’s Case 66 shows why this “acquittal equivalence” must be accepted as a jeopardy policy, whatever else may also be included. Five defendants were charged with conspiracy to murder the king. The year was 1678, and the previous century had seen the bitter Reformation, the residue of which lingers to this day. Only a few years prior to Ireland’s Case, the authorities had thwarted the famous Gunpowder Plot to blow up Parliament that is still memorialized as Guy Fawkes Day. The prosecutor in Ireland’s Case sought to show that the plot against King James was motivated by the same English “papists” who, “being moved and seduced by the instigation of the devil,”67 had sought to “introduce the popish religion, and to destroy the established Protestant religion in England.”68
At the close of the evidence, the judge summarized the case for the jury. He noted the requirement that the crime be proved by two witnesses and stated that, with respect to two of the defendants, Whitebread and Fenwick, the requirement had not been met. The prosecution had proved the existence of a plot to kill the king, according to the judge, but not against these defendants. The judge concluded, “[S]o that though the testimony be so full, as to satisfy a private conscience, yet we must go according to law, too.” Thus, “it is a great evidence that is against them [Whitebread and Fenwick]; but it not being sufficient in point of law, we discharge you of them; it is not a legal proof to convict them by, whatsoever it may be to satisfy your consciences. Therefore remove Mr. Fenwick and Mr. Whitebread from the Bar.” The judge ordered Fenwick and Whitebread returned to prison “until more proof may come in.”69 Both were later retried, convicted, and executed,70 their plea of former jeopardy rejected because no formal verdict existed.71
It is unlikely that our system, with independent judges, would ever produce a travesty of justice on the order of Ireland’s Case. Nonetheless, the possibility of prosecutorial manipulation of the process argues for a mistrial doctrine that discourages prosecutors from using early trial termination as acquittal avoidance. Drawing on Ireland’s Case, an early termination should not prevent an analysis of the evidence to see whether an acquittal would have occurred.
The result in Ireland’s Case is wrong, on this view, not because the trial ended prior to verdict but because there is no reason to distinguish between defendants who are acquitted and those who would have been acquitted had the jury been permitted to return a verdict.72 Deciding which defendants belong in the latter category is obviously difficult. If we hypothetically had foreknowledge of the evidence that would be admitted, we could construct a universe of defendants against whom the state will present a legally insufficient case. A double jeopardy account sensitive to substance rather than procedure would not permit this class to be subdivided into those who are acquitted and those whose trial is ended by the prosecutor’s motion.
We do not have the ability to predict the future, of course. But that does not mean we should ignore the problem of acquittal avoidance. Some amendment to Blackstone’s pleas in bar is thus necessary. The nature of the amendment is implicit in the concept of blameworthiness. The entire double jeopardy principle underlying this book can be restated as a prohibition of treating singular blameworthiness as anything other than singular blameworthiness. In the first seven chapters, this principle allowed me to simplify the same-offense doctrine and to unify it with collateral estoppel. In the second jeopardy context, this principle means that no possible double jeopardy issue can arise unless the first trial has produced a finding as to blameworthiness.
We can relax Blackstone’s reliance on formal verdicts by recognizing that lack of blameworthiness can be manifested in outcomes other than formal verdicts. Other endings can be “acquittal equivalents,” by which I mean an ending that resolves the blameworthiness issue in favor of the defendant. An example is a dismissal granted on the ground of insufficient evidence.
At this point, the argument is vulnerable to the legal-realist claim that something as vague as acquittal equivalence opens the way, at best, to interpretational difficulties and, at worst, to the infinite regress to the bargain basement of legal realism where each case requires its own rule. It is a fair criticism, but some soft-edged principle cannot be avoided if we wish to give courts the authority to prevent or rectify Ireland’s Case. Acquittal equivalence requires courts to look at the record in each nonverdict jeopardy case, but that is what courts do when defendants appeal convictions on the ground of insufficient evidence. Moreover, it is possible to limit acquittal equivalence to keep it from consuming the entire mistrial doctrine.
Verdicts end jeopardy. But verdicts can occur in ways other than a formal entry of judgment. Acquittal equivalents occur in three procedural settings: at the end of trial, on appeal, and (rarely) in mistrials. The issues at the end of trial and on appeal are conceptually linked and can usefully be considered as paradigm cases of acquittal equivalence.
An outcome that resolves the facts of the case in the defendant’s favor is an acquittal, whether or not it is called by that name. Green v. United States73 is the classic example. Green was found guilty of second-degree murder following a trial on a charge of first-degree murder. The Court held that the effect of the formal verdict was to acquit Green of first-degree murder,74 a rule that Bishop had recognized early in the twentieth century.75 As odd as it may sound, even a conviction can be an acquittal. The Court showed how seriously it took the analogy to acquittal when it later held that the mere prosecution of the “acquitted” offense was a violation of double jeopardy, even though the defendant was convicted of the same lesser-included offense as in the first trial (so, in one sense, defendant was not harmed by the second trial).76 This is precisely how a formal acquittal works, and the Court held that an implicit acquittal works the same way.
The Court refused to extend Green to routine sentencing. In United States v. DiFrancesco,77 the government appealed a sentence on the ground that it was “clearly erroneous”—too lenient—under the relevant dangerous-special-offender sentencing provisions, which permitted government appeal. The defendant claimed that the government was forbidden to appeal his favorable sentence because it was tantamount to an implicit acquittal of a harsher sentence. The Supreme Court rejected this claim, noting a “distinction between acquittals and sentences” that “does not require that a sentence be given a degree of finality that prevents its later increase.”78
The Court was right to reject this analogy. Double jeopardy blameworthiness is an either/or proposition. Either Green is guilty of first-degree murder or he is not. If not, either he is guilty of second-degree murder or not. Judgments that a defendant is not guilty of X crime is a judgment that the defendant is blameless for X crime.
By contrast, the “acquittal” of longer sentences does not entail a finding that the defendant is blameless. It manifests, instead, a judicial notion about proportionality of the penalty to the demonstrated blameworthiness. But, as we saw in chapter 1, double jeopardy has nothing to say about proportionality, deferring that question to the Cruel and Unusual Punishment Clause or to the Due Process Clause.79
The Court made an exception to DiFrancesco for death penalty sentencing. A finding against the death penalty by the judge or jury “acquits” the defendant of the death penalty, and the state cannot seek that penalty if a new sentencing hearing is later required.80 Again, the Court’s result follows by analogy to verdict blameworthiness. The two-tier procedure in death cases requires a separate sentencing hearing at which new evidence is admitted under relaxed rules of evidence, and the only issue is whether to apply the death penalty. The process of hearing new evidence and making an either-or decision on the death penalty very much resembles the either-or decision on guilt. Indeed, if one views the death penalty as different from all others, the analogy is striking. A defendant is either “guilty” in the sense of deserving this unique penalty or he is not.
The Court muddied the jeopardy waters for a time by introducing a concern with finality or harassment rather than acquittal equivalence in dismissal cases. United States v. Jenkins held that a dismissal barred further proceedings “devoted to the resolution of factual issues going to the elements of the offense charged.”81 Even if these factual issues were not resolved in the defendant’s favor in the first trial, that they remain unresolved was enough to convert the dismissal into a jeopardy bar.
The unanimous opinion, written by Justice Rehnquist, relied on Justice Black’s oft-quoted language from Green about the double jeopardy limit on the power of the state “to make repeated attempts to convict an individual for an alleged offense, thereby subjecting him to embarrassment, expense and ordeal and compelling him to live in a continuing state of anxiety and insecurity, as well as enhancing the possibility that even though innocent he may be found guilty.”82 But, as already noted, Justice Black’s eloquent language is not analytically helpful. Yes, it is more costly for defendants to undergo two trials, but the costs of a second trial are unacceptable only when the first trial has produced an outcome that we recognize as the end of jeopardy for a single offense. Drawing on Michael Moore, I attempted in chapter 2 to dispel the confusion that some magical quality exists about second prosecutions that changes the same-offense analysis.
Similarly, the threat of a second prosecution does not magically transform dismissals into verdicts. The costs to the defendant of two proceedings are severe regardless of the reason for having a second proceeding, but these costs are cognizable in double jeopardy policy only when there is a reason to treat the first outcome as the end of jeopardy. Is a dismissal the end of one jeopardy just because it would require a new proceeding if reversed on appeal? Jenkins held yes, but the outcome is in no way predetermined by Justice Black’s quote.
Three years later, in Scott v. United States, the Court admitted that “we pressed too far in Jenkins the concept of the ‘defendant’s valued right to have his trial completed by a particular tribunal.’”83 In an opinion by Justice Rehnquist, the author of Jenkins, the Court overruled Jenkins by a vote of 5-4, candidly admitting that “vastly increased exposure to the various facets of the Double Jeopardy Clause” had led it to conclude that Jenkins was “wrongly decided.”84 It may well be the only time a Justice has authored an opinion overruling a case in which he wrote the Court’s opinion.
Rehnquist acknowledged in Scott that Black’s quote is not helpful: “These historical purposes are necessarily general in nature, and their application has come to abound in often subtle distinctions which cannot by any means all be traced to the original three common-law pleas.”85 Rehnquist is half right in this observation. Yes, Black’s quote is more oratory than analytic. But most of the subtle distinctions that have developed in defining second jeopardy can be traced to Blackstone’s common-law pleas. The ones that cannot be traced to Blackstone should be abandoned. Specifically, Scott, but not Jenkins, can be traced to the plea of autrefois acquit.
Indeed, Scott noted that Justice Black’s quote is fully honored if the government is forbidden to reprosecute when Blackstone’s common-law pleas would forbid a retrial.86 The Court followed this insight by recognizing that the label “acquittal” does not necessarily answer the question of whether a dismissal operates as an acquittal. Scott held that a dismissal requested by the defendant was a bar to retrial only when “the ruling of the judge, whatever its label, actually represents a resolution [in the defendant’s favor], correct or not, of some or all of the factual elements of the offense charged.”87 This is what I mean by “acquittal equivalence.” Thus, with more agonizing than necessary, Scott held exactly what Blackstone would have predicted: an acquittal bars a new trial, and a midtrial dismissal that is not an acquittal does not bar a new trial or an appeal.
Nonacquittal dismissals simply have no double jeopardy effect. Thus, United States v. Wilson88 held that a nonacquittal dismissal after conviction does not bar an appeal. The Court’s rationale was that no new trial would be required. Either the trial judge will be upheld or the jury verdict of guilt will be reinstated. But an unresolved issue is posed in those (surely rare) cases when the judge grants an acquittal notwithstanding the jury’s verdict based on the judge’s view that the evidence was insufficient.89 The problem here is not that we do not have a verdict in form or substance, which is the problem in most dismissal and mistrial cases; rather, it is that we have too many verdicts, both formally and substantively. The jury concluded that the defendant was guilty. The judge, hearing the same evidence, concluded that he was not guilty. Is there any reason to prefer the judge’s conclusion to that of the jury?
The Wilson rationale seems robust enough to permit an appeal here.90 Double jeopardy protection operates against a second verdict for the same offense. As James Strazzella recognizes, the bar of an appeal following a jury acquittal is merely an ancillary rule designed to protect the defendant against having to defend a meaningless appeal.91 When the trial judge overrules a jury conviction, the appeal is not meaningless because the jury verdict can be reinstated and no new trial need occur. Wilson should permit an appeal here, as at least one state court has held.92 The decision on blameworthiness was, after all, a split decision.
Dismissals that determine the lack of legislatively decreed blameworthiness, usually on the ground of insufficient evidence, are acquittal equivalents that should be treated exactly like acquittals. Dismissals after conviction permit appeal by the prosecution, because a new trial is never necessary. Other forms of dismissals can be appealed by the state, and, if the appeal is successful, a new trial can be held.
Defendants can appeal convictions, of course. Sometimes defense appeals will produce an acquittal equivalent.
We saw earlier that defense appeals that produce a reversal leave no extant formal verdict to operate as a jeopardy bar. But could not an appellate reversal leave an acquittal equivalent? Scott recognized that a resolution of facts in the defendant’s favor is an acquittal equivalent. The same concept should apply in appeals. The defendant in Burks v. United States93 won his appeal on the grounds that the conviction was not supported by sufficient evidence. The Court held that this was equivalent to an acquittal, thus barring a new trial.
Ball v. United States,94 which permitted a new trial after reversal, did not have to be overruled. As Burks noted, the Ball reversal was based on a flawed indictment, not on insufficient evidence. Verdict election makes sense in the former situation but not the latter. The latter outcome means that the trial jury should never have been permitted to deliberate. Had the trial judge done his job, or had the jury done its job, Burks would have received an acquittal at the trial level.
Burks is merely an application of the basic principle that jeopardy ends when a defendant is found substantively not guilty. Findings of not guilty at the trial court level—either by judge or by jury—end the matter forever,95 but findings of legally insufficient evidence by appellate courts are a little more complicated. In Forman v. United States,96 a panel of the Court of Appeals reversed the defendant’s conviction with instructions to enter a judgment of acquittal on the ground that, as the case had been tried, no other verdict was possible. Upon rehearing, however, the Court of Appeals changed its mind, noting a theory that would have permitted a conviction, and ordered a new trial. The defendant appealed the second judgment, arguing that the first Court of Appeals judgment operated as an acquittal.
The Supreme Court affirmed the second judgment. Though this case predated Burks, the issue should come out the same way today. Once a defendant puts into play the question of whether his conviction is supported by legally sufficient evidence, the issue is not resolved until the review is complete. Evidence is not legally insufficient until there is a final, unreviewable order holding that it is legally insufficient.
The post-Burks Court was willing to engage in quite fine distinctions about what constitutes an insufficient-evidence reversal. The Court ultimately drew a distinction between reversals when evidence is insufficient as a matter of law and reversals when evidence is against the weight of the evidence.97 Though Justice White may be right that the effect of this fine distinction is to undermine Burks,98 the Court’s theory is right: no jeopardy bar should arise unless the reversal effectively finds the defendant not guilty.
The Court has also held that the Burks determination must be based on the evidence the trial court actually admitted, even if some of it should have been suppressed and thus not available to the jury.99 The defendant argued that he was entitled to a review of sufficiency based only on the evidence that should have been presented. The Court disagreed. Again, the theory is right. The reason to treat reversal for insufficient evidence as an acquittal equivalent is that no rational jury could have convicted on that evidence. The reviewing court thus must put itself in the position of the jury. Moreover, as a practical matter, if the trial court had refused to admit certain state evidence, the prosecution might have been able to find other, similar evidence that was admissible.100
Burks, like Scott, applies only to outcomes that resolve blameworthiness. An example of a case that might or might not be an acquittal equivalent is the one in which the prosecutor knowingly withheld exculpatory evidence. The Pennsylvania Supreme Court tackled this issue in Commonwealth v. Smith,101 and Anne Bowen Poulin has written an excellent article on the double jeopardy implications of this violation.102
Both Smith and Poulin focus on the misconduct of the prosecutor and view the Double Jeopardy Clause as harmed by the covert undermining of the fairness of the first trial. On this view, the second trial is a double jeopardy harm of the harassment variety that we saw in chapter 2. As my double jeopardy construct has no place for harms that result from generalized due process unfairness, the focus is on whether the first trial would have resolved the defendant’s blameworthiness in his favor. This requires a counterfactual determination: had the exculpatory evidence been admitted, would the defendant have been acquitted?
Details of the counterfactual inquiry remain to be developed; admittedly, this test is no easier to satisfy than the amorphous “harassment” inquiry required by traditional double jeopardy policy. But counterfactual analysis at least asks the right question—did the first trial constitute a verdict equivalent? Because the state is at fault for the failure to disclose in Smith, the mechanism for sorting out the counterfactual claims should be reasonably easy for the defendant to satisfy in this category. With that in mind, I would apply a version of the harmless error rule: considering the exculpatory evidence together with the evidence that was admitted, would the state have had overwhelming evidence of guilt? If so, the defendant has suffered no cognizable double jeopardy harm. If not, the first outcome should be treated as an acquittal.
One advantage of this approach is that it should temper the anger and surprise that greet decisions like Smith.103 Rather than viewing double jeopardy as a reward to defendants who have been treated unfairly or as a punishment for prosecutors who have misbehaved, a bar of a second trial should be less controversial if the outcome is seen as equivalent to an acquittal. Viewing an acquittal as a likely outcome of the first trial had it been fair makes a double jeopardy bar logically appropriate. This is, of course, the rationale of Burks, made more difficult here by the counter-factual inquiry.
The argument to this point is that the Court’s second jeopardy doctrine is informed by acquittal equivalence—the Green implicit acquittal doctrine, the Scott dismissal rule, and the Burks insufficient-evidence exception to the Ball general rule that a successful appeal permits a new trial. Acquittal equivalance can be extended to some mistrials, like Ireland’s Case. But can it be extended to more typical mistrials?
Judicial manipulation as in Ireland’s Case has not appeared in the reported cases in this country. Moreover, whatever the extent of prosecutorial manipulation, defendants almost never win jeopardy claims based on mistrials. Of sixteen mistrial cases in the Supreme Court, the defendant gained a majority vote in only one, Downum v. United States, 104 and that by a 5-4 margin. The question is whether the acquittal equivalence concept can do any work in this category of cases.
Bishop in 1923 drew a distinction between the double jeopardy effect of discharging a jury on the ground that the jurors were later found to be disqualified and discharging juries for other reasons.105 The former category permitted retrial, but not the latter. It seems likely from Bishop’s analysis that he viewed the typical jury discharge as equivalent to an acquittal.
Whenever, after the jury is sworn, it is found that the evidence is not sufficient to convict; or that a material witness for the prosecution is absent; or that such witness is unacquainted with the nature of an oath, and so requires instruction before testifying; or that the witness is suddenly taken too ill to proceed,—no second trial can be had.106
The first example (evidence not sufficient) is clearly an acquittal equivalent, and the others could be if Bishop envisioned the trial continuing without a key witness for the prosecution. This reading of Bishop is supported by an earlier reference to the effect of a prosecutor discontinuing an indictment during trial. Bishop states, without explanation, that “the legal effect is an acquittal.”107 If my reading of Bishop is right, the idea of acquittal equivalence in mistrial cases is an old one.
When the prosecutor requests a mistrial, two concerns arise: the unfairness of forcing the defendant to gear up for another trial and the Ireland’s Case unfairness of permitting the state to avoid an acquittal. The first kind of unfairness can be obviated by getting the defendant’s consent to the mistrial; it is less clear that the Ireland’s Case concern is completely addressed by the defendant’s consent to the mistrial. Focusing on the first kind of unfairness for the moment, in chapter 2 I identified the general unfairness of facing a second trial as harassment and concluded that it is a soft-edged policy that does not give courts much guidance in resolving individual cases.
Implicitly accepting the soft-edged characterization, the Court has adopted a legal realist approach to mistrials requested by the prosecution, as we saw in detail in chapter 3. The Court purports to balance the relevant interests, the fulcrum being the “manifest necessity” phrase taken from Justice Story. This balance involves finality, the weakest of the values in the Westen and Drubel heirarchy. On this reading of the Court’s doctrine, a no-verdict outcome forced on the defendant violates the weak finality interest and becomes a weak presumption that no new trial can begin. But because this is a weak interest, and a weak presumption, the state can rebut the presumption by showing “manifest necessity,” whatever that means.
A fuller quotation from Story’s opinion observes that the trial court may discharge the jury without a verdict upon finding “a manifest necessity for the act, or the ends of public justice would otherwise be defeated.”108 One analysis has demonstrated that this statement is susceptible to three interpretations: (1) either manifest necessity or the ends of public justice may justify a mistrial, (2) manifest necessity justifies a mistrial and the ends of public justice justify permitting a new trial, and (3) manifest necessity must exist and the ends of public justice must be served before a mistrial can be declared.109 The “ends of public justice” is as vague as “manifest necessity,” thus magnifying the difficulty of using this standard.
The vagueness of the concepts, and the inevitable tendency of judges to permit the state to have one full shot at defendants, has produced a jurisprudence that recognizes almost any governmental excuse as “manifest necessity.”110 The mistrial cases involving the right to a verdict typically explain the facts, quote the “manifest necessity” language, and then either conclude that manifest necessity was present, as a ritualistic incantation, or defer to the judgment of the trial court. It is difficult to quarrel with Stephen Schulhofer’s conclusion that the manifest necessity standard is a “thoroughly deceptive misnomer, perhaps not rivaled even by the Holy Roman Empire.”111
The defendant won Downum v. United States,112 however, and this case may provide a clue about what really counts in the mistrial analysis. The prosecution moved for mistrial before any witnesses testified, on the ground that the key prosecution witness on two counts was unavailable. The judge granted this motion over defense objection, and the Supreme Court held that double jeopardy prohibited a retrial. One way to explain Downum is to say that prosecutors are simply stuck with their mistakes. If a prosecution-requested mistrial resulted from prosecutorial error, an unfairness that rises to the level of a jeopardy bar has occurred. That reading does not, however, survive Illinois v. Somerville,113 where the prosecutor negligently drafted the indictment but prevailed against a jeopardy challenge to the prosecution-requested mistrial.
Justice Rehnquist, who wrote the majority opinion in Somerville, offered an explanation for Downum that did not focus on the prosecutor’s mistakes. The key for Rehnquist was whether the mistrial “operated as a post-jeopardy continuance to allow the prosecution an opportunity to strengthen its case.”114 If, as in Downum, the prosecutor will likely strengthen its case following the mistrial, courts should not find manifest necessity. If no reason exists to think that the state’s case will look different at the next trial, as in Somerville,115 then no jeopardy bar arises.
Rehnquist is on the right track. Focusing on the prosecutor’s chance to strengthen the state’s case raises the second kind of unfairness, the kind that my blameworthiness theory recognizes: the possibility that defendants will be denied acquittals to which they were entitled. This chapter amplifies Rehnquist’s standard by seeking to make it more precise and by linking it with my blameworthiness theory. Before doing that, I examine the other category of mistrials: those requested by defendants.
The Court applies a different mistrial standard when the defendant moves for a mistrial. At first glance, it might seem that a defense motion for a mistrial should have no jeopardy implications at all. After all, the defendant is getting what she wants; why should she be allowed to complain later? But as the Court has recognized, if a defendant-requested mistrial could never be grounds for barring retrial, the state might seek to insulate the prosecutor’s conduct from the scrutiny that attends a prosecution-initiated mistrial by forcing the defense to make the motion. Moreover, if the defendant is entitled to an acquittal, perhaps her ill-advised motion for a mistrial should be ignored.
Oregon v. Kennedy116 is the seminal case on defense-requested mistrials. Defense counsel attempted to establish bias on the part of a prosecution witness by questioning him about a criminal complaint that he had filed against the defendant. The prosecutor then sought to rehabilitate the witness by eliciting the reasons why he had filed the complaint. The trial court sustained a series of objections to this line of questioning, rulings that the Oregon Court of Appeals later characterized as “probably wrong.”117
Faced with substantial impeachment of one of her witnesses and an inability to rehabilitate the witness because of probably erroneous evidentiary rulings, the prosecutor then asked if the witness had ever done business with the defendant. When the witness said no, the prosecutor asked, “Is that because he is a crook?”118 Not surprisingly, the judge granted the defense motion for a mistrial. Should this mistrial be the end of jeopardy?
The state appellate court “accepted the trial court’s finding that it was not the intent of the prosecutor to cause a mistrial” but held reprosecution barred in any event because the prosecutor was guilty of “overreaching.”119 The Supreme Court reversed, in an opinion by Justice Rehnquist that adopted as a standard whether the prosecutor “intended to ‘goad’ the defendant into moving for a mistrial.”120 But, as Justice Stevens pointed out in a separate opinion, it is “almost inconceivable” that the defendant could show “intent to provoke a mistrial instead of an intent simply to prejudice the defendant.”121
The Rehnquist standard (in fact if not in theory) permits prosecutors free rein to introduce prejudicial elements into a trial that is not going well. The worst that can happen is that the defendant obtains a mistrial, and the state can begin again. If the defendant does not obtain a mistrial, the jury will have seen or heard the prejudicial evidence. These options must seem appealing to a prosecutor whose case is not going well.
Unfortunately, the test proposed by Stevens (writing for three other members of the Court as well as himself) is also flawed. It is whether “egregious prosecutorial misconduct has rendered unmeaningful the defendant’s choice to continue or to abort the proceeding.”122 How would a reviewing court know when the defendant’s choice to continue the trial had become “unmeaningful”? This is pure legal realism that seeks a unique right outcome for each case. But how could this standard help anyone decide an actual case?123
The dilemma, then, is that the Rehnquist standard has relatively hard edges but is virtually impossible to meet, and the Stevens standard is possible to meet but has hopelessly soft edges. There must be a better way. Perhaps a focus on blameworthiness can provide a better test for mistrials than the Court’s twin tests—“manifest necessity” for prosecution-requested mistrials and “intent to goad” for defense-requested mistrials.
Acquittal equivalence in the mistrial context is obviously more speculative than a reversal of a conviction on grounds of legally insufficient evidence that follows a review of the record of a full trial. The record is incomplete in most mistrial cases, and the question must be answered as a counterfactual: would the defendant have been entitled to an acquittal, as a matter of law, had the trial judge denied the motion for a mistrial? Something close to this kind of counterfactual inquiry underlies Justice Rehnquist’s comment that the Downum mistrial “operated as a post-jeopardy continuance to allow the prosecution an opportunity to strengthen its case.”124 The prosecution’s concession that its key witness on two counts was not available created an inference that the defendant might have won at least those counts had the case gone forward.
Indeed, Downum quoted with approval from a Ninth Circuit case that used an analysis very much like counterfactual acquittal equivalence:
[T]he district attorney entered upon the trial of the case without sufficient evidence to convict.… There is no difference in principle between a discovery by the district attorney immediately after the jury was impaneled that his evidence was insufficient and a discovery after he had called some or all of his witnesses.125
It is not quite as simple as the Ninth Circuit makes it seem. Deciding that the evidence would have been insufficient requires conjecture. One way of making the test less speculative is to put the burden of proof on the defendant and to weigh the evidence at the point of the mistrial motion unless there is a basis to predict the rest of the state’s case. The defendant will lose if no plausible acquittal claim suggests itself. If a defendant makes a plausible counterfactual acquittal claim, however, one could indulge a bias in favor of finding a jeopardy bar. This bias can be defended as a way of discouraging prosecutors from begininng trial with weak cases and as a way of incorporating part of the finality interest that can be more directly addressed as a due process claim.126
Incorporating this bias into the counterfactual inquiry merely requires softening the standard. Rather than ask whether the defendant would have been entitled to an acquittal as a matter of law if the motion had been denied, we could ask whether there was a realistic possibility that the state’s case would have been legally insufficient had the trial gone forward. Even on this softer standard, defendants will not win very many of these cases, but they win very few under the current doctrine. At least on the counterfactual test, we know and can state clearly why defendants lose—they cannot demonstrate a realistic possibility of a legally insufficient case had the judge denied the mistrial motion.
Sometimes the counterfactual inquiry is quite simple. In Somerville, the basis for the mistrial was a fatal defect in the indictment. The Court permitted a retrial on the ground that the error did not suggest manipulation, and it made little sense to force the state to proceed with a trial that contained automatic reversible error. On the counterfactual account, defendants have the burden of demonstrating that an acquittal was a realistic possibility; as the record contained no evidence, Somerville cannot make a plausible claim to acquittal equivalence. (I assume here, as did the Court, that state law permitted a conviction despite the flawed indictment.)
The defendant in United States v. Jorn127 was prosecuted for assisting the preparation of fraudulent income tax returns. The government subpoenaed five taxpayer witnesses to testify against Jorn. The judge warned the first of these witnesses of his privilege against self-incrimination and then refused to allow him to testify, despite his expressed willingness, until he had consulted a lawyer.128 The judge ascertained from the prosecutor that the remaining witnesses were “similarly situated” and discharged the jury “so abruptly”129 that there was no opportunity for the prosecutor to “suggest a continuance” or the defendant “to object to the discharge of the jury.”130
A plurality of the Court held that the mistrial was a jeopardy bar, focusing on the defendant’s right to a particular jury. But Jorn is best explained as acquittal equivalence. The judge terminated the trial before evidence of guilt was introduced and in a way that suggested the government’s case was going to be insufficient had the trial continued. Without any of its taxpayer witnesses, the government’s case was unlikely to be sufficient. While defendants normally lose if the record is bare of evidence, because the burden is on the defendant to show a realistic likelihood of an acquittal, the judge’s actions in Jorn permit the defendant to carry that burden.
Justices Black and Brennan would have decided Jorn on the ground that the “action of the trial judge amounted to an acquittal.”131 The plurality rejected this resolution because the record did not show that the trial judge relied “on facts relating to the general issue of the case.”132 Had the trial judge relied on the facts, Jorn would have been acquitted in the strong, Scott sense of a finding of evidence insufficiency. But a blameworthiness account suggests that counterfactual acquittal equivalence should also be treated as the end to jeopardy.
Jorn should not be tried again, in the view of Black and Brennan, because “the action of the trial judge amounted to an acquittal.” Precisely right. It was not an acquittal, as is a dismissal for insufficient evidence, but it amounted to an acquittal because Jorn was deprived of a realistic possibility of acquittal had the case gone forward. What the plurality’s “right to go to a particular tribunal”133 means, on this view, is the right to have the particular government’s case evaluated as if it had gone to the jury.
The counterfactual acquittal equivalence in Downum is not quite as obvious as in Jorn. The only “evidence” in the Downum record was the content of the prosecutor’s mistrial motion that cited a missing key witness on two counts. As to those two counts, the prosecutor’s assertion that he lacked a key witness should satisfy the counterfactual test for acquittal. The difficulty with Downum is that the Court barred retrial on all counts. Perhaps this is justifiable as a sort of penalty on the prosecution for dismissing the uninvolved counts along with the weak counts or as a blanket inference of insufficiency infecting all counts. But the jeopardy bar on the other counts is much more difficult to justify as a counterfactual acquittal equivalent.
The defendant in Brock v. North Carolina134 had a compelling acquittal equivalence claim but lost because the Court analyzed the claim under the Due Process Clause. Two of the state’s witnesses refused to testify on self-incrimination grounds. The prosecutor then told the court that the testimony of those witnesses “was necessary for the State to present its case fully before the jury” and moved for a mistrial, which the trial judge granted.135 Brock was decided prior to incorporation of the Double Jeopardy Clause into the Fourteenth Amendment, and the Court explicitly noted that it was not reaching the double jeopardy issue.136 Under the Due Process Clause, as we see in chapter 9, the test was: “Is that kind of double jeopardy to which the state has subjected [the defendant] a hardship so acute and shocking that our polity will not endure it?”137
The statement of the test implies that an outcome can be double jeopardy without simultaneously violating due process. Using the very high due process threshold, the Court found no violation. In dissent, Justice Douglas put the issue in acquittal equivalence terms, claiming that both double jeopardy and due process should prevent prosecutors from being able to “call a halt in the middle of a trial in order to await a more favorable time, or to find new evidence, or to make up the deficiencies in the testimony of its witnesses.”138
Whatever the right outcome under the Due Process Clause, Douglas’s view endorses my blameworthiness account of the Double Jeopardy Clause. The blameworthiness question is the counterfactual—would the case have proceeded favorably to the accused?—not whether the prosecutor was at fault for the lack of evidence. The state’s concession in Brock that the missing testimony was “necessary for the State to present its case fully” is more than enough to meet the counterfactual test. If Brock occurred today, with the Double Jeopardy Clause fully enforceable against the states, the defendant should win on the implicit acquittal equivalence rationale in Downum.
Acquittal equivalence in mistrial cases also illuminates the defense-requested mistrial category. As noted earlier, the Court’s standard in Oregon v. Kennedy was whether the prosecution intended to goad the defendant into moving for a mistrial, while the Stevens alternative standard was whether the defendant’s choice to continue the trial was rendered “un-meaningful.”
On a blameworthiness account of double jeopardy, there is only one question in cases that end short of verdict: would the defendant have had a realistic chance at an acquittal if the judge had denied the mistrial motion? There is no reason to treat the identity of the party who moves for a mistrial as anything other than an evidentiary fact in judging the defendant’s chance of an acquittal. We do not need to worry about “overreaching” or “goading” or “egregious prosecutorial misconduct” that “has rendered unmeaningful the defendant’s choice to continue or to abort the proceeding.” We need only ask the blameworthiness question: can this defendant show a realistic possibility of an acquittal had the judge denied a mistrial?
If the mistrial occurs during the state’s case, as in Kennedy, an acquittal equivalence test would be almost impossible for the defendant to satisfy unless the record manifested a weak case.139 Assuming a case of typical strength, the transcript would disclose only the prosecution witnesses and their cross-examination. That would be not be enough, in a typical case, to allow the defendant to meet the burden of showing that he would have had a realistic chance of an acquittal at the point of mistrial. When the mistrial is granted well into the defense case, the reviewing court is better able to judge whether there was a realistic possiblity that the state’s case would be legally insufficient. When this judgment favors the defendant, the mistrial should bar a second trial even though granted on the defendant’s motion.
The policy justification for an acquittal equivalence mistrial rule inheres in the actual harm done (or not done) to the defendant. If the defendant had a realistic chance at an acquittal, the prosecutor has violated the defendant’s double jeopardy rights in a very real sense. If not, the only harm that is done, and it is not a cognizable double jeopardy harm, is that the defendant has to endure a second trial to remedy the flaw in the first trial. What we would like to know before deciding whether a defendant should face a second trial is whether real double jeopardy harm was done, so why not ask it directly and have the jeopardy question turn exclusively on the answer?
A thought experiment demonstrates the wisdom of a mistrial doctrine based solely on acquittal equivalence.140 Combine the facts of Downum and Somerville: the prosecutor wants a mistrial because her key witness cannot be found, but she coincidentally discovers a fatal defect in the indictment (that cannot be cured by amendment). She candidly discloses both grounds for mistrial to the judge. Because Somerville held that this indictment defect is manifest necessity that justifies a mistrial, the prosecutor’s acquittal-avoidance motive is presumably irrelevant. But this is just the kind of mistrial that double jeopardy should count as a bar. On an acquittal equivalence account, a second trial is barred because the focus is on the chance for acquittal, not on the state’s grounds for a mistrial or the likelihood that a conviction would be automatically reversed on appeal.
The next category is the no-verdict termination imposed by the judge after the jury fails to reach a verdict, commonly called a “hung” jury. To begin with the obvious, the Court has held that a jeopardy bar arises when the jury cannot return a verdict and the judge grants the defendant’s motion for acquittal.141 This, of course, is an acquittal, rather than a no-verdict outcome.
It should not be surprising that a judge might grant a motion for acquittal following a hung jury. Although hung juries can be caused by other factors, the most likely cause is that some of the jury find the evidence insufficient to convict. The overlap between hung juries and acquittals appears in Hotema v. United States,142 discussed in chapter 7 for its collateral estoppel implications. Hotema was charged with three murders; he was tried for two murders in a single trial and acquitted on the ground of insanity. The jury hearing the last murder case could not agree on a verdict, and the government reprosecuted. Hotema was convicted when a new jury heard the government present its case for the third time. That one jury acquitted strongly suggests the hung jury resulted from juror doubts about Hotema’s sanity.
A defendant in Hotema’s situation can make a motion for acquittal following the jury’s failure to reach a verdict. The difficult issue is whether the Double Jeopardy Clause requires appellate courts to entertain an appeal when the trial judge overrules the motion for acquittal following a hung jury.143 It might seem obvious that defendants should have this right. After all, it is no more difficult to evaluate the evidence underlying a hung jury than that underlying a conviction, and appellate courts regularly perform the latter analysis.
The Court had, for a century and a half, routinely permitted retrial following discharge of a hung jury,144 but, in 1985, for the first time, it addressed a claim specifically based on the failure of the trial judge to recognize evidence insufficiency. The argument is powerful: if the jury and trial judge have made a mistake, if the defendant is legally entitled to an acquittal, why should he not be given the benefit of a jeopardy bar? Moreover, as the defendant in Richardson v. United States145 recognized, a hung jury resulting from legally insufficient evidence is no different from a conviction based on insufficient evidence. When the Court held in Burks that a conviction based on insufficient evidence was an acquittal equivalent,146 it opened the door to the argument that a hung jury based on insufficient evidence should be treated the same way. Moreover, the venerable Perez case is not precedent to reject Richardson’s argument; there was no motion for an acquittal following the hung jury in Perez. One can distinguish a defendant who offers a hung jury as a jeopardy bar and one, like Richardson, who offers the failure of the trial judge to grant a motion for acquittal as a potential jeopardy bar.
Probably only a few hung juries result from evidence that an appellate court would find legally insufficient. A jury can be hung (in most states) by only one holdout juror, who may have an idiosyncratic view of the evidence. The due process standard for legally sufficient evidence is a low one—whether any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt.147 There are many cases in which the evidence is weak but legally sufficient under this test; only egregiously defective cases warrant reversal.
Any case that fails to meet the Court’s evidence sufficiency standard is woefully weak. The jury should have acquitted, and the trial judge should have granted the postdismissal motion for acquittal. The appellate court can remedy this double failure of the process in the same way the Court now requires appellate courts to remedy convictions based on insufficient evidence. As the Court implicitly recognized in Richardson, an insufficient-evidence hung jury is substantively the same as a formal acquittal. Why not treat it the same way?148
The defendant in Richardson was simply asking the Court to elaborate the mistrial rule in the same way that the Court in Burks elaborated the appellate reversal rule. Richardson was not arguing that all mistrials constitute the end to jeopardy; he was not even arguing that all hung jury mistrials constitute ended jeopardies; he asked only that appellate courts judge whether a hung jury mistrial is based on legally insufficient evidence and, in that small category of cases, declare that acquittal equivalence bars a new trial.
The Court did not seize the opportunity to make the mistrial rule symmetrical with the appellate reversal rule. Instead, Justice Rehnquist wrote for the Court that the mistrial rule “has its own sources and logic”149 and held that a retrial following a hung-jury mistrial is always permissible, relying on Story’s opinion in Perez. Perhaps unconvinced that even the internal logic of the mistrial rule demanded this result, Rehnquist also noted that “Justice Holmes’ aphorism that ‘a page of history is worth a volume of logic’ sensibly applies here.”150 In effect, the Court held that precedent, and not logic, justified its refusal to recognize the Burks symmetry.
Justice Brennan noted in dissent that, under the majority’s rule, “a defendant who is constitutionally entitled to an acquittal but who fails to receive one—because he happens to be tried before an irrational or lawless factfinder or because his jury cannot agree on a verdict—is worse off than a defendant tried before a factfinder who demands constitutionally sufficient evidence.”151 True enough, but the majority’s rule is even more bizzare than Brennan’s argument implies. Assuming legally insufficient evidence, the defendant who receives a hung jury is not only worse off than the defendant who is acquitted; he is also (in theory) worse off than the defendant who is convicted.152 The latter defendant can appeal and receive an appellate reversal that is a bar to a new trial; the hung jury defendant has to suffer a new trial, during which the government has a chance to hone its case.
The implications of the majority’s position are startling. Consider a situation in which the evidence is demonstrably insufficient, and eleven members of the jury vote to acquit but one juror votes to convict. If the eleven cannot convince the one to change his mind, the defendant would be better off (again, in theory) if the eleven switched their votes to conviction. Then the defendant could test the sufficiency of the evidence without giving the government another chance to present its case.
Perhaps the Court was concerned about the additional burden on appellate courts if it recognized Richardson’s argument. But that is inadequate reason to ignore the logic of Burks. Either Burks or Richardson must be wrong, and, on a blameworthiness account, it is Richardson. Only two members of the Court dissented from the decision in Richardson to treat a hung-jury mistrial less favorably than a conviction based on insufficient evidence.153 This suggests how conceptually barren is the Court’s double jeopardy universe.
If a future Court is willing to rethink this problem, it need only recognize the distinction drawn earlier between the defendant who offers a hung-jury mistrial as a bar to retrial and a defendant who seeks review of the trial court’s denial of the motion for acquittal. As to the former defendant, no procedure exists to decide whether the hung jury resulted from legally insufficient evidence, and one could defer to the historical argument that a hung jury never constitutes a jeopardy bar. As to the latter defendant, however, the existence of a procedure for appealing the trial court’s denial of the motion for acquittal suggests that the proper course is to treat this situation like any other appeal and reach the merits of the claim.
Oddly, eight members of the Court in Richardson agreed that the federal statute governing appeal permitted “review of the sufficiency of the evidence … as a necessary component” of a double jeopardy claim,154 only to hold that a hung-jury claim could never be meritorious. After holding that Richardson had a “colorable double jeopardy claim appealable under 28 U.S.C. § 1291,”155 the Court dropped an Alice-in-Wonder-land foonote, which says in its entirety:
It follows logically from our holding today that claims of double jeopardy such as petitioner’s are no longer “colorable” double jeopardy claims which may be appealed before final judgment. A colorable claim, of course, presupposes that there is some possible validity to a claim. Since no set of facts will support the assertion of a claim of double jeopardy like petitioner’s in the future, there is no possibility that a defendant’s double jeopardy rights will be violated by a new trial, and there is little need to interpose the delay of appellate review before a second trial can begin.156
Thus, Richardson had a colorable claim, appealable under 28 § U.S.C. 1291, but, once the Court reached the merits of his appeal, there could never be a colorable claim in the future, and thus no future defendant would have a colorable claim. It may be, as the Court asserts, that the hung-jury doctrine “has its own sources and logic,”157 but it is very bad logic indeed if it supports the Court’s view that defendants have (had?) a colorable claim that is not now (never was?) a colorable claim.
When a defendant seeks appellate review of the order denying the motion for acquittal, Perez is technically inapplicable, and Richardson holds that this claim is procedurally appealable. All that is necessary to restore coherence to this corner of double jeopardy is to permit the hung-jury appeal to be heard substantively, as a type of Burks claim.
Though Richardson based its procedural appealablity ruling on a federal statute, the requirement of procedural review is also embedded in the Double Jeopardy Clause. The Court noted in Abney v. United States that “if a criminal defendant is to avoid exposure to double jeopardy and thereby enjoy the full protection of the Clause, his double jeopardy challenge to the indictment must be reviewable before that subsequent exposure occurs.”158 If the conceptual incoherence of Richardson can be rectified, the solution recommended here can be imposed on states that permit motions for acquittal following a hung jury. If a state does not permit such motions, there would be no procedural predicate upon which a substantive blameworthiness argument could work. But as long as a procedural predicate exists, the Double Jeopardy Clause should require appellate review of orders that overrule motions for acquittal, whether they follow convictions or mistrials.
If jeopardy does not end until verdict—formal verdict or acquittal equivalence—it does not matter when it begins. In Blackstone’s day, the beginning of jeopardy was the same as the end of jeopardy. No procedural mechanism existed to terminate a trial prior to verdict, and it would have made no sense to speak of a beginning or ending of jeopardy.
Today, our criminal procedure permits preverdict terminations out of a concern for defendants. In Blackstone’s day, the rules permitted more certainty and less fairness on this issue. Defendants knew that they would get a verdict and that no new trial could follow as long as the verdict stood. The difficulty, of course, is that defendants could not get a mistrial if they wanted one, regardless of how many errors had infected the trial.
In the Court’s doctrine, a gap exists between the beginning of jeopardy and the end of jeopardy. The beginning of jeopardy is important in the Court’s doctrine because it signals when a mistrial must be tested for manifest necessity and a dismissal must be tested for acquittal equivalence. Prior to the beginning of jeopardy, a termination has no jeopardy implications. Suppose a judge grants a mistrial because the jury is hung; the government reschedules a trial, and the judge grants a pretrial motion for acquittal. The judge’s action is explicitly based on the ground that he has already heard the government’s case and is now persuaded that the government has insufficient evidence to convict. As we have seen, a judicial finding of insufficient evidence counts as an acquittal and thus is a bar to a new trial for that offense.
But not on the Court’s view of the “attachment” of jeopardy. The second dismissal is not a verdict, regardless of its grounds, because it does not end a jeopardy. It does not end a jeopardy because no jeopardy on the second attempt to convict had begun. What has not begun cannot end. While this seems unduly formalistic, it was the holding in United States v. Sanford.159
Two years later, in Crist v. Bretz, the Court held that the Double Jeopardy Clause requires jeopardy to attach when the jury is sworn, striking down Montana’s provision that jeopardy attached when the first witness was sworn.160 The Court stated, in dicta, that the Montana rule was appropriate for a trial without a jury but constitutionally defective for trials with juries. Since the difference between the time the jury is sworn and the time the first witness is sworn is a short period that includes only opening arguments, one is left to wonder what is so crucial about this period that it necessitated invalidating Montana’s rule that both kinds of jeopardy begin when the first witness is sworn.
No coherent policy reason has been offered for the distinction, as Justice Powell pointed out in his dissent in Crist.161 The Court quoted the platitude that the Double Jeopardy Clause protects “the defendant’s valued right to have his trial completed by a particular tribunal”162 and further articulated the genuine (but necessarily general) concern that the Double Jeopardy Clause is designed to minimize “harassing exposure to the harrowing experience of a criminal trial.”163 These platitudes do not, however, require the attachment of jeopardy at any particular point. As Justice Blackmun remarked, these interests could just as easily “support a conclusion that jeopardy attaches at the very beginning of the jury selection process.”164
The Court noted several possible lines for attachment—for example, jury selection and the prima facie case. The government recommended a “sliding scale” that would have abolished a fixed line. On this view, “the further the trial has proceeded the more justification is required for a mid-trial termination.”165 Stephen Schulhofer, in his usual thoughtful style, defended a variety of jeopardy lines in a way roughly consistent with the government’s position (and one year earlier).166 On a legal-realist view that exalts individual solutions for individual cases, no other approach seems defensible. Westen and Drubel are true to their legal realism when they note that the choice of a single line from among many is arbitrary “if by arbitrary one means that the decision is essentially ‘an exercise in line drawing.’”167 Though the Court has been heavily influenced by legal realism in double jeopardy analysis, it rejected the individual-case approach to the attachment of jeopardy. We are thus left with a “bright line” that has no convincing rationale beyond the perceived need to have a point of demarcation.
A blameworthiness account rejects the Court’s bright-line attachment doctrine, not in favor of a sliding scale as favored by the legal realists but in favor of returning to Blackstone’s notion that jeopardy begins and ends at verdict. It does not matter whether a judge resolves the facts in the defendant’s favor before or after the process has crossed some imaginary line. Once the facts are resolved in the defendant’s favor, by dismissal or as a counterfactual mistrial inquiry, jeopardy both begins and ends. Blackstone turns out to be right again.
The Court’s solution is, at best, arbitrary. At worst, it is nonsensical. The outcome of the pretrial dismissal case, Sanford, depended completely on the fortuity of whether the judge decided the motion for acquittal before or after the trial began. The Court was wrong to reject Sanford’s jeopardy claim. To be sure, virtually all pretrial dismissals do not represent a judgment on the merits of the government’s case,168 but when they do, as in Sanford, the acquittal equivalence of the outcome should be recognized, not submerged beneath a formalistic doctrine about the attachment of jeopardy.
In this chapter I have argued that the no-verdict cases are best understood as requiring the defendant to demonstrate that the outcome is some form of acquittal equivalent. Acquittal equivalents come in two forms: dismissals that actually resolve the factual elements in the defendant’s favor and the counterfactual determination that, on the basis of the record, the defendant can show a realistic possibility that the state’s case would have been legally insufficient had it gone forward.
As noted throughout this chapter, several Supreme Court opinions contain dicta suggesting a broader purpose behind the double jeopardy mistrial rule—to protect a defendant’s “valued right” to a particular jury.169 These statements never explain the origin of this “valued right” beyond a generalized fear of repetitive trials. The origin of this right is a finality interest, which exists even in cases where an early outcome would not qualify as an acquittal equivalent.
Although precisely predicting how acquittal equivalent claims will be decided is impossible, we can say precisely what interest of the defendant is at stake: defendants should not be deprived of their right to go to verdict when they have a realistic chance of an acquittal. That defendant has already demonstrated lack of blameworthiness, at least in a counterfactual sense, and should receive the shield of an acquittal.
The finality interest, on the other hand, is much more amorphous. Why does a defendant have an interest in a particular jury? However the interest is defined, it winds up being a fairness interest. Even the standard for determining when that interest is violated, derived from Perez, sounds in fairness: a defendant has a right to a particular jury unless the prosecution can show a “manifest necessity for the [mistrial], or the ends of public justice would otherwise be defeated.”170
I once argued that the finality interest in going to verdict sounds in due process, not double jeopardy.171 If mistrial claims premised on finality are qualitatively different from acquittal equivalent claims, as my argument presupposed, one might ask why the Supreme Court began analyzing these finality claims under the Double Jeopardy Clause. A careful look at the cases demonstrates that the Court drifted into the practice, first pointing out that no claim could be made at all (Perez), then maintaining that no claim could be made but, if one could, this would not be one,172 and finally discussing the merits of the claim as if they mattered.173 After that, it was a short trip to recognizing the claim.174 The reason the Court fell down this slippery slope is that the two categories of mistrial claims look procedurally alike. The only difference is the substance of the claim, and sloppy analysis could see this difference as one of degree rather than kind.
But a difference in kind does exist. Acquittal equivalence mistrial claims are about verdicts, and other mistrial claims are about fairness. If fairness is the issue, courts must balance the defendant’s finality interest against the need for the mistrial in light of the “ends of public justice.” What will this balance look like? What kinds of cases justify mistrials outside of acquittal equivalence? Nothing in double jeopardy theory is helpful; there is no verdict, and no analogy to a verdict.
The focus on fairness suggests locating this part of the mistrial doctrine in the Due Process Clause, which traditionally uses a balancing test to decide whether a right has been denied on the facts of a particular case.175 The relocation project is merely a change in labels. Mistrial finality claims have always been analyzed differently from other double jeopardy questions, reflecting implicit judicial recognition that finality claims are conceptually different from other double jeopardy claims. Double jeopardy analysis in other areas depends on a binary analytical structure: same offense, different offense; same sovereign, different sovereign; civil remedy, criminal punishment; one conviction, more than one conviction.176 The Court’s use of interest balancing to resolve mistrial finality cases suggests that they are generically different from the rest of double jeopardy. As James Shellenberger and James Strazzella put it: “The Supreme Court has maintained a strict double jeopardy bar in those circumstances in which the common law rules imposed a finality on judgments when it came to successive prosecutions. However, where the Court has moved beyond these common law situations, its results appear to represent a majority assessment of a basically pragmatic balancing of interests.”177 My acquittal equivalance theory is an explicit recognition of the difference between verdict finality and pragmatic balancing of interests.
To the extent that a coherent doctrine can be distilled from (or imposed upon) these mistrial finality cases, it is difficult to improve on Stephen Schulhofer’s two-tier standard.178 Late mistrials would receive strict scrutiny because of the defendant’s greater interest in completing the trial and the greater potential harm in a new trial. The later the mistrial, the stricter the scrutiny.
Mistrials declared early in the trial are less harmful. Schulhofer concludes that early mistrials should be evaluated under a more flexible “sound judicial adminstration” standard. Under this standard, “the defendant is entitled to some protection against the burden of reprosecution, but the nature of the burden ordinarily will not be sufficient to justify awkward or expensive alternatives to mistrial, and the trial judge normally will be in a position to render a fair and trustworthy judgment on the need to abort the proceedings.”179 The test is whether the trial court’s judgment is “responsible” and whether appellate review of that judgment should be deferential.180 This standard is essentially what the Court later adopted in Arizona v. Washington,181 which is appropriate for the finality subset of mistrial claims but not, as I have argued in this chapter, for the acquittal equivalence claims.
By subtracting the finality mistrial claims from the second jeopardy doctrine, while adding the Ireland class of acquittal equivalents, we wind up with a second jeopardy doctrine based on blameworthiness. Defendants who have not elected to void a criminal judgment have a good double jeopardy claim when the first trial has ended in a formal verdict, in an outcome that is functionally the same as an acquittal, or in an outcome that suggests an acquittal would have followed had the trial not terminated early. Any other bar to a second trial must be found in the Due Process Clause or in a due-process style double jeopardy analysis that is consciously distinct from the acquittal equivalence analysis.
That this approach can be traced directly to Blackstone is not reason enough to accept it. But I am not alone in recognizing the virtue of returning to something like Blackstone’s pleas in bar. Jay Sigler has argued that returning to the English final-judgment rule, with added procedural limitations to make it difficult to withdraw a case, is fairer both to the accused and to the prosecution. Requiring a final judgment would “insur[e] a full hearing for both the state and the accused.”182
I agree.