Ruminations on Inconsistent Verdicts (3/21/16)

Today, I offer an interesting "order" in a tax conspiracy trial ruminating on the broad issue of inconsistent jury verdicts.  The case is United States v. Soderling, 2016 U.S. Dist. LEXIS 35456 (ND Cal. 2016), here.  Because the order is relatively short, I just cut and paste the contents of the order (caption and signature line omitted):

ORDER RE GOVERNMENT'S REQUEST TO AUTHORIZE THE JURY TO REACH INCONSISTENT VERDICTS 

The government charged Jay and Jessica Soderling, husband and wife, with a conspiracy to defraud the United States by preventing the Internal Revenue Service from collecting taxes. At trial, the government contended only that Jay and Jessica conspired with one another; it did not contend that there were unindicted co-conspirators. Shortly after the jury began deliberating, it asked a good question: "Can we find one defendant guilty of conspiracy and not the other?" The government urged the Court to answer "yes." The defendants urged the Court to answer "no." Although the jury has now reached a verdict, the Court files this order to give lawyers and trial judges food for thought in the event this instructional issue comes up again. 

By urging a "yes" answer, the government was asking the Court to tell the jury it could reach inconsistent verdicts. As set forth in the Ninth Circuit Model Criminal Jury Instruction on conspiracy to defraud the United States, a jury may only convict a defendant if the government proves three elements, the first of which is that there was an agreement between two or more people. In this case, the government contended that Jay and Jessica reached the agreement to defraud the United States; the government did not contend that anyone else was party to the agreement. Therefore, the only way the jury could have convicted Jay was to conclude he reached an agreement with Jessica. And the only way the jury could have convicted Jessica was to conclude she reached an agreement with Jay. A verdict acquitting one defendant would necessarily have been inconsistent with a verdict convicting the other defendant. 

In support of its request for an instruction authorizing the jury to reach inconsistent verdicts, the government cited United States v. Powell, 469 U.S. 57 (1984). In Powell, the jury convicted the defendant of several compound felonies, while acquitting her of the predicate felonies.n1 Ms. Powell argued that because these verdicts were necessarily inconsistent, her convictions had to be reversed. The Supreme Court rejected that argument, holding that the inconsistency among the verdicts was not a reason to overturn the convictions. As the Court explained, inconsistent verdicts can be the result of jury nullification — a decision by the jury to show mercy with respect to a particular charge by acquitting the defendant notwithstanding her guilt (and notwithstanding the trial court's instructions). There was no reason, in the Supreme Court's view, to give the defendant an additional benefit from that possible exercise of mercy by reversing her convictions for the other counts charged. Powell, 469 U.S. at 68-69. So long as there was enough evidence to support the charges the jury convicted the defendant of, it did not matter that the jury acquitted her on the other charges. Id. at 67-68. Since Powell, several Ninth Circuit decisions have applied this rationale to cases where a jury acquits all but one alleged conspirator. See, e.g., United States v. Ching Tang Lo, 447 F.3d 1212, 1226 (9th Cir. 2006); United States v. Hughes Aircraft Co., 20 F.3d 974, 977-78 (9th Cir. 1994); United States v. Valles-Valencia, 823 F.2d 381, 381-82 (9th Cir. 1987) modifying 811 F.2d 1232 (9th Cir. 1987) ("As the Supreme Court noted in United States v. Powell, however, inconsistent verdicts can just as easily be the result of jury lenity as a determination of the facts. Thus, the acquittal of all conspirators but one does not necessarily indicate that the jury found no agreement to act." (citations omitted)).
   n1  Specifically, the Powell jury convicted the defendant of using the telephone to facilitate conspiracy to possess with the intent to distribute cocaine, and using the telephone to facilitate possession with the intent to distribute cocaine. But the jury acquitted the same defendant of conspiracy to possess with intent to distribute cocaine, and of possession with intent to distribute cocaine. Powell, 469 U.S. at 59-60. 

It's one thing to say a conviction won't be overturned on appeal merely because it is inconsistent with a co-defendant's acquittal. It's quite another thing to instruct a jury, during trial or during deliberations, that it is authorized to reach inconsistent verdicts in the first place. Inconsistent verdicts are something we tolerate, but that doesn't change the fact that they are legally erroneous. And trial courts are supposed to instruct juries to follow the law; they are not supposed to give juries the green light to disregard the law and reach legally erroneous verdicts. United States v. Christensen, 801 F.3d 970, 1011-12 (9th Cir. 2015) (noting, in the course of affirming a court's dismissal of a juror for cause during deliberations, that "trial courts have the duty to forestall or prevent" jury nullification, id. at 1011 (quoting Merced v. McGrath, 426 F.3d 1079, 1080 (9th Cir. 2005)); accord United States v. Sepulveda, 15 F.3d 1161, 1189-90 (1st Cir. 1993) (approving trial court's statement to the jury, in response to a question about nullification, that it should follow the instructions the court had previously given and should convict if it finds the elements beyond a reasonable doubt). Indeed, the very first of the Ninth Circuit Model Criminal Jury Instructions states: "To the facts as you find them, you will apply the law as I give it to you, whether you agree with the law or not."

And in this case, the law is that a conspiracy cannot exist unless it involves an agreement between two or more people. It would have been error for the jury to convict one defendant while acquitting the only potential co-conspirator. Therefore, it would have been error for the Court to authorize the jury, through the Court's response to the jury's question, to convict one defendant of conspiracy despite concluding that the government had not proven the only other possible co-conspirator's participation beyond a reasonable doubt.n2
   n2 Ninth Circuit Model Criminal Jury Instruction 1.13, the generic instruction for cases involving multiple defendants, provides in part: "The fact that you may find one of the defendants guilty or not guilty should not control your verdict as to any other defendant." This is no doubt true in most multi-defendant cases, but it is somewhat misleading in a case where there are only two alleged co-conspirators and no evidence of unindicted co-conspirators. 

And an error like this could, depending on the facts of the case, have grave consequences. Suppose, in a case where two defendants are charged with conspiracy and there are no other co-conspirators, the jury believes there is not enough evidence to conclude beyond a reasonable doubt that the defendants formed an agreement. This means both defendants are "not guilty" as a legal matter. But suppose the jury really wants to convict one of the defendants anyway. If a trial court suggests to a jury that it can do so, and the jury then does so, that would be a miscarriage of justice. Appellate courts might tolerate this possibility when reviewing inconsistent verdicts because they don't know (and won't inquire into) what the jury was thinking. But that doesn't mean trial courts can (or should) sanction the possibility when instructing the jury in the first place. 

Perhaps if the Court had given the government's proposed instruction and the jury had followed it, the error invited by the government would have gone uncorrected on appeal, in light of Powell. More likely, however, Powell would not have applied in this situation, because Powell was not a case in which the trial court invited the jury to err. Cf. Kleve v. Hill, 202 F.3d 1155, 1158 (9th Cir. 2000) (Kozinski, J., dissenting from denial of rehearing en banc) ("Powell holds that where the jury is properly instructed and nevertheless returns inconsistent verdicts, we will not question the validity of the guilty verdicts on some counts just because they are logically inconsistent with not guilty verdicts on other counts . . . . For Powell to apply, the jury must have been given a correct, consistent set of instructions."). 

In light of the foregoing, the Court declined the government's request to answer "yes" to the question whether the jury could "find one defendant guilty of conspiracy and not the other." Instead, the Court instructed the jury as follows: 

For a conspiracy to exist, two or more people must be members of the conspiracy. I refer you to Instruction 18, which discusses the three elements of conspiracy. The first element explains that to convict a defendant of Count 1, you must conclude beyond a reasonable doubt that the defendant formed an agreement with at least one other person to defraud the United States. If you find this element beyond a reasonable doubt, you should turn to the next two elements. 

In this case, the government has only contended that Mr. and Mrs. Soderling formed an agreement to defraud the United States. It has not contended that any other person agreed with one of the defendants to defraud the United States. 

Perhaps the government was worried that if the Court did not authorize the jury to reach inconsistent verdicts, both defendants would be acquitted of the conspiracy charge. As it turns out, the jury convicted both defendants. But had the Court accepted the government's invitation to instruct the jury that it could reach inconsistent verdicts, the government might not have gotten either of its convictions. The result may well have been one acquittal through jury nullification and one reversal on appeal.

Apart from the law, the strategy in each side's proposed response to the jury's question is interesting.  The Government obviously wanted at least one guilty verdict which is what the jury's question implied it might get.  On the other hand, by asking the judge to answer the question no, the defendants must have thought there was some reasonable prospect that the jury would resolve to a not guilty verdict for both.

The issue of inconsistent verdicts is a big issue.  I have not done the work or given sufficient thought to offer anything here that might be considered "state of the law" on the issue.  United States v. Powell, 469 U.S. 57 (1984), here, is essential reading.  I do offer certain excepts from two law review articles and then from a dissenting opinion in an en banc case from the Sixth Circuit.

Eric L. Muller, The Hobgoblin of Little Minds? Our Foolish Law of Inconsistent Verdicts, 111 Harv. L. Rev. 771 (1998) (footnotes omitted [SSRN version here]:

What should a court do when a single criminal jury reaches logically irreconcilable verdicts - for example, acquitting a defendant of possessing drugs with an intent to distribute, but convicting her of using a phone to commit that offense? The Supreme Court has unanimously decreed that a court should leave the verdicts undisturbed. To reverse the inconsistent conviction would not only require guesswork about what produced the inconsistency, but would also be unfair to the government, which cannot appeal the inconsistent acquittal. In this Article, Professor Muller demonstrates that this do-nothing approach to inconsistent verdicts masks a deep pro-government bias. He argues that the Court has both underestimated the risks that inconsistent verdicts pose to criminal defendants and overstated the impact of such verdicts on the government. He also observes that the Court has imagined only all-or-nothing solutions to the problem of inconsistent verdicts: either reverse all inconsistent convictions, or affirm them all. Professor Muller suggests three intermediate solutions that would allow courts to distinguish among inconsistent verdicts and disturb only those that are likely to have resulted from pro-government jury error. 

* * * * 

1. Multiple-Count Inconsistencies. - Juries sometimes return inconsistent verdicts on multiple counts against a single defendant. Claire's case, the third of the hypotheticals with which this Article began, is an example of multiple-count inconsistency - an example that closely tracks the facts of United States v. Powell, 23Link to the text of the note the leading Supreme Court case on the topic. Powell involved a compound federal crime, the crime of using a telephone to commit a federal drug felony. 24Link to the text of the note The crime is "compound" because it is a crime piggybacking on top of another crime: to be guilty of the telephone facilitation charge, the defendant must also be guilty of committing the underlying predicate drug felony. In Powell, the inconsistency arose because the jury acquitted the defendant of the underlying felony (conspiring to possess cocaine with the intent to distribute it), but convicted her of using a telephone to commit that very felony. 25Link to the text of the note The jury is instructed, in a case like Powell, that the underlying felony is an element of the telephone facilitation crime that must be proven beyond a reasonable doubt. 26Link to the text of the note When the jury in Powell delivered its inconsistent verdicts, it declared the underlying felony simultaneously proven and not proven, suggesting that the jury must have disregarded the court's instructions. 

Multiple-count inconsistencies are possible whenever the government tries a compound crime and its predicate offense together.  Compound offenses have proliferated in recent decades; Congress has created compound offenses in the areas of racketeering, 28Link to the text of the note money laundering, and narcotics trafficking. 30Link to the text of the note As the number of compound crimes has increased, the chances for multiple-count inconsistencies have increased as well. 

2. Multiple-Defendant Inconsistencies. - The second common type of inconsistent verdict is a single jury's acquittal of all but one of multiple defendants charged with jointly committing a crime that requires at least two participants. The common-law crime of riot is a good, if ancient, example. Under English law, that crime required three participants. It was legally impossible for fewer than three people to riot. Thus, if three defendants stood trial together for riot, and the jury convicted only one of them, this was a fatally inconsistent verdict, and the conviction could not stand. Other crimes that have been understood to require more than one participant are adultery, incest, fornication, bigamy, miscegenation, and dueling.  

Some of these are no longer crimes in this country; others are rarely, if ever, prosecuted. But conspiracy is still quite commonly prosecuted, and it requires at least two participants. The essence of conspiracy is an agreement between two or more people to commit some unlawful act. Thus, in order to sustain a defendant's conviction for conspiracy, the government must prove beyond a reasonable doubt that the defendant agreed with someone that an unlawful act would be committed.  

Conspiracy cases obviously present a significant risk of inconsistent verdicts when all of the alleged co-conspirators are tried together. In fact, the risk was great enough, and the occurrence of such verdicts frequent enough, that courts developed a rule for resolving challenges to inconsistent conspiracy convictions. That rule - now followed in only a few jurisdictions - is known as the "rule of consistency." Based on the idea that "it is impossible in the nature of things for a man to conspire with himself," the rule sees the failure of a conspiracy charge against one of two alleged co-conspirators as a failure against both. It therefore "requires the reversal of a defendant's conspiracy conviction if all his coconspirators are acquitted of the same conspiracy charges" in a single trial.  

To date, almost every federal court of appeals has either abandoned or severely criticized the rule. Unfortunately, they have not replaced it with any rational alternative. Before examining the courts' current response to inconsistent verdicts in greater detail, I must first explain what prompts juries to reach inconsistent verdicts.

* * * * 

VI. Conclusion 

The Supreme Court's response to the problem of inconsistent criminal jury verdicts is distressing. The Court correctly concedes that inconsistent verdicts reveal jury error, but sees only all-or-nothing remedies for that error: either reverse all inconsistent convictions or affirm them all. The Court has rightly seen the former remedy as unfair to the government, because it would strip the government of even those convictions in which it, rather than the defendant, was the victim of the jury's error. And so the Court has turned to the remedy of doing nothing - automatically affirming all inconsistent convictions.
Doing nothing will not do. The error that inconsistent verdicts reveal may harm the government; it may result from a genuine desire on the part of the jury to show leniency. Or it may not, and if it does not, the error may compromise the reasonable doubt standard, a fundamental protection of the criminal defendant, or it may reveal hostility, bias, or some other improper animus on the jury's part. At the very least, an inconsistent verdict is a warning signal that something may have gone awry in the jury room, to the defendant's detriment. We do not generally ignore warning signals: the flashing red light on the dashboard means that it is time to open the hood and check the engine. If we take warning signals seriously in life, we should also do so in the law. The reform proposals in this Article have the virtue of taking the jury's warning signal seriously. 

Of course, when it comes to jury verdicts, we cannot do the equivalent of throwing open the hood and looking at the engine, because we are deeply committed to the secrecy of jury deliberations. We believe that juries will deliberate better when they deliberate in confidence, so we blind ourselves to post-verdict evidence of jury error. That is probably a wise judgment, and in any event, it is an ancient one we are unlikely to revisit. But there is one narrow window into the jury's deliberations that we allow ourselves to peer through. That window is the jury's verdict. When we also blind ourselves to the error that appears through that one window, we achieve a very foolish consistency.

Irinia Kotchach Bleustein, Anna E. Bodi and Anna Majestro, Federal Criminal Conspiracy, 52 Am. Crim. L. Rev. 1089, 1113-1114 (2015) [no link; some footnotes omitted].

B. Joinder and Severance 

The prosecutor must decide whether to try co-conspirators jointly or separately. n146 Although joint trials create the danger that the fact-finder will not be able to distinguish the more guilty parties from other less culpable co-conspirators, n147 conspiracy charges usually provide a proper basis for joinder. n147 Severance will be granted only if a defendant can make "a strong showing of prejudice." n149
   n146 Traditionally, under the "rule of consistency," if conspirators were tried together, one conspirator could not be convicted under § 371 if the other conspirators were acquitted. See United States v. Velasquez, 885 F.2d 1076, 1090-91 (3d Cir. 1989) (holding when evidence was insufficient to convict co-conspirator of conspiracy, the evidence was also insufficient to convict defendant). See generally Eric L. Muller, The Hobgoblin of Little Minds? Our Foolish Law of Inconsistent Verdicts, 111 HARV. L. REV. 771, 781-82 (1998) (describing traditional rule). Today, this rule--which never required the acquittal of the last-tried defendant if the defendants were tried separately, has been rejected by all but the Seventh Circuit. See, e.g., Robertson v. Klem, 580 F.3d 159, 161 n.1 (3d Cir. 2009) ("[T]he fact that all alleged coconspirators are acquitted does not undermine a defendant's otherwise valid conviction for a conspiracy offense."); United States v. Nichols, 374 F.3d 959, 971 n.9 (10th Cir. 2004) (recognizing for the first time that the rule of consistency, as applied to co-conspirators, is no longer good law), vacated on other grounds, 543 U.S. 1113 (2005); United States v. Crayton, 357 F.3d 560, 564-65 (6th Cir. 2004) (noting that the "rule of consistency" is no longer considered good law). But see United States v. Mancari, 875 F.2d 103, 104 (7th Cir. 1989) (acquitting defendant of conspiracy due to the government's concessions regarding the "rule of consistency" and a lack of evidence that defendant had other co-conspirators during the time frame alleged in the indictment).
   n147 See Kotteakos v. United States, 328 U.S. 750, 774 (1946) (explaining danger of transference of guilt is "so great that no one really can say prejudice to substantial right has not taken place"); United States v. McVeigh, 169 F.R.D. 362, 370-71 (D. Colo. 1996) (recognizing joint trial of defendants presented unacceptable risk of prejudice to both and granting defendants' motions for severance).
   n148 [footnote omitted]
   n149 [footnote omitted]

Majority and Dissenting opinions in Getsy v. Mitchell, 495 F.3d 295, 319-322 (6th Cir. 2007) (en banc), here.  Getsy involved the issue of whether two different sentences -- one a death sentence and one a life sentence -- imposed in two separate trials were unconstitutionally disproportionate.  I quote first from the majority opinion, focusing only on the inconsistent verdict portion of the opinion (pp. 304-309):

The primary issue raised by Getsy, and the only issue certified for appeal by the district court, is whether Getsy's sentence was unconstitutionally arbitrary or disproportionate in relation to that received by Santine. Getsy's argument ultimately rests on the fact that Santine, the mastermind who directed codefendants Getsy, Hudach, and McNulty to kill Charles Serafino, did not receive the death penalty. Santine's indictment mirrored Getsy's. In a separate trial that took place after Getsy's, Santine was convicted of aggravated murder and aggravated burglary, but was acquitted of all the capital specifications charged and thus ineligible for the death penalty. Getsy claims that this disparity renders his death sentence arbitrary and disproportionate.

* * * *

Moreover, we have long held that the common-law rule of consistency has no application to conflicting verdicts returned by different juries in separate trials. See United States v. Newton, 389 F.3d 631, 636 (6th Cir. 2004) (noting that the rule of consistency "was not applied if co-conspirators were separately tried"), vacated in part on other grounds, 546 U.S. 803, 126 S. Ct. 280, 163 L. Ed. 2d 35 (2005); United States v. Sachs, 801 F.2d 839, 845 (6th Cir. 1986) ("[I]f coconspirators are tried separately, the acquittal of all other coconspirators does not mandate acquittal as to the remaining conspirator. . . . In other words, it is not necessarily inconsistent for two juries to reach differing results."); see also Cortis v. Kenney, 995 F.2d 838, 840 (8th Cir. 1993) (same); United States v. Lewis, 230 U.S. App. D.C. 212, 716 F.2d 16, 22 (D.C. Cir. 1983) (same); United States v. Sangmeister, 685 F.2d 1124, 1126-27 (9th Cir. 1982) (same); United States v. Espinosa-Cerpa, 630 F.2d 328, 333 (5th Cir. 1980) (same). This well-established precedent squarely precludes the old common-law rule from applying under the circumstances of this case. 

Getsy simply had no constitutional guarantee that his jury would reach the same results as prior or future juries dealing with similar facts, irrespective of the offense with which he was charged. Criminal defendants are instead protected from irrational convictions by the due process requirement that a conviction must be supported by sufficient evidence. Powell, 469 U.S. at 67 ("[A] criminal defendant already is afforded protection against jury irrationality or error by the independent review of the sufficiency of the evidence undertaken by the trial and appellate courts. We do not believe that further safeguards against jury irrationality are necessary."); see also Espinosa-Cerpa, 630 F.2d at 332 n.5 (explaining the ancient origin of the English common-law rule of consistency and "its inappropriateness to a modern American criminal system in which all verdicts obviously are, and always have been, subject to independent review for evidentiary support"). Thus, the constitutionality of Getsy's murder-for-hire conviction turns not on any fortuity regarding when he was tried or with whom, nor on the caprice permissible in another jury's decision to acquit on similar facts, but rather on the sufficiency of the evidence presented at his own trial. 

Only where a court declares that the evidence is legally insufficient to support the conspiracy conviction of one defendant must the conviction of the sole coconspirator also be voided. Morrison v. California, 291 U.S. 82, 93, 54 S. Ct. 281, 78 L. Ed. 664 (1934) (reversing two defendants' joint conspiracy convictions where due process precluded the state's reliance on a legal presumption to establish an element of the conspiracy). We pause to emphasize that, contrary to the view of the dissent, "[a] court's determination that there is insufficient evidence to convict cannot be equated with a jury's determination that a defendant, for whatever reason, should be acquitted." Crayton, 357 F.3d at 566. Apparently recognizing this fundamental distinction, Getsy himself has never argued that Morrison applies to his case, even in the wake of the original-panel majority's unwarranted reliance on that decision. Nevertheless, the dissent presses on with this argument, overlooking the critical distinction between a determination made by a court as a matter of law -- with which Morrison dealt -- and a jury verdict. (All discussion in this opinion of the dissent or the dissenting opinion refers to the lead dissent authored by Judge Merritt.) 

Santine's case was allowed to go to a jury, and that jury ultimately acquitted him of the murder-for-hire specification. But the very fact that the issue was submitted to a jury indicates that the evidence against him was not so deficient that the trial court could decide the question as a matter of law. Furthermore, jury verdicts differ intrinsically from decisions made by a court. See Crayton, 357 F.3d at 566; see also Powell, 469 U.S. at 66 (noting, in the context of inconsistent verdicts in a single trial, that "[t]he fact that the inconsistency may be the result of lenity, coupled with the Government's inability to invoke review, suggests that inconsistent verdicts should not be reviewable"). Although the dissent points out that the Supreme Court has "never retracted or narrowed" the holding in Morrison, neither has the Court ever expanded it to require the reversal of one conspirator's conviction or sentence in light of a coconspirator's acquittal by a separate jury. Certainly Enmund v. Florida, 458 U.S. 782, 102 S. Ct. 3368, 73 L. Ed. 2d 1140 (1982), in which the Supreme Court reversed the death sentence of a defendant on the ground that he did not personally kill or intend to kill anyone, was not such a case.

* * * *

The dissent's references to Aristotle, Sir Francis Bacon, Sir Edward Coke, and English cases beginning in the year 1599 strike us as quite scholarly, even if only marginally relevant. Obviously the controlling law is that of the United States Supreme Court, not the King's Bench. What the dissent's historical exposition fails to cite is even a single instance in which the Supreme Court or any federal court has ever reversed one defendant's sentence or conviction based on another defendant's later acquittal by a separate jury. The dissent does not, because it cannot, explain how such a supposedly well-established rule has remained hidden within this country's federal jurisprudence for so long a time.

Now, for Judge Merritt's lead dissenting opinion (pp. 319-322):

I. The Invalidity of Inconsistent Verdicts in Prosecutions Based on A Criminal Agreement 

Since 1599 during the reign of Queen Elizabeth I, when Sir Edward Coke was Attorney General and the young philosopher-scientist, Sir Francis Bacon, was Queen's Counsel, the rule of Anglo-American law has been that "one cannot conspire alone," or alone commit a contract crime like murder for hire. This exact language was first enunciated in Marsh v. Vaughn, 78 Eng. Rep. 937 (Q.B. 1599). The opinion of the Queen's Bench states as follows: 

The defendants pleaded not guilty, and the one was found guilty and the other not. And it was hereupon moved, that the bill should abate; for it ought to be against two, and the one cannot conspire alone; and the one being acquitted, the other sole cannot be attainted. 

Id. (Emphasis added.) This is no judicial aberration. This is the way English law has dealt with such disproportionate punishment. This rule has been consistently followed in English law from that day to this. See, e.g., Harison v. Errington, 79 Eng. Rep. 1292 (K.B. 1627) (riot); Rex v. Grimes, 87 Eng. Rep. 142 (K.B. 1688) (two were charged with " confederationem" and "though one was acquitted, yet the jury had found the other guilty" requiring the court to quash the guilty verdict); Rex v. Kinnersley, 93 Eng. Rep. 467 (K.B. 1719) (same); Queen v. Thompson, 117 Eng. Rep. 1100 (Q.B. 1851) (same); Rex v. Plummer, 2 K.B. 339, 345 (1902) (court invalidated a conspiracy conviction after a guilty plea when the defendant's two alleged co-conspirators were acquitted). See also IV Blackstone's Commentaries on the Laws of England, ch. 10, P 15, p. 136 (1765) (Legal Classics Library 1983) (requiring conviction of two to constitute a criminal agreement "for there must be at least two to form a conspiracy"). This ancient rule of consistency and proportionality in punishment was legislatively enacted by Parliament in 1977, Criminal Law Act, 1977, ch. 45 § 5(8), which provides that when other persons charged with a criminal agreement "have been acquitted of conspiracy by reference to that agreement (whether after being tried with the person convicted or separately) the conviction shall be quashed if under all the circumstances of the case his conviction is inconsistent with the acquittal of the other." For more than four centuries, from 1599 until the present day, that has been the law. Since the time of Lord Coke, the English courts under this doctrine would never have let Getsy's conviction of murder for hire stand. 

In a supreme instance of legal legerdemain, the majority opinion in this case tries to spin the opinion of the Supreme Court in United States v. Powell, 469 U.S. 57, 105 S. Ct. 471, 83 L. Ed. 2d 461 (1984), into a rejection of this ancient rule and the rule's insistence on a measure of rationality, consistency and proportionality in punishment. That reliance is completely specious because Powell is not a multi-defendant case in which a defendant's conviction of a criminal agreement with another stands as the jury acquits his only alleged co-conspirator. Powell was simply a single defendant situation in which the jury convictions on separate counts were inconsistent under one reading of the charges made in the separate counts. There the Supreme Court, relying on language from an earlier opinion by Justice Holmes, held that the separate counts provided rough equity or fairness in holding that the Powell defendant must take the bad count with the good counts. See 469 U.S. at 62 (quoting and agreeing with Justice Holmes in Dunn v. United States, 284 U.S. 390, 393, 52 S. Ct. 189, 76 L. Ed. 356 (1932), in which Holmes relied on an English case from the Queen's Bench, Latham v. The Queen, 122 Eng. Rep. 968 (Q.B. 1864), for the proposition that "each Count in an indictment is regarded as if it was a separate indictment"). 

It is impossible to legitimately rely on Powell and Dunn here because two terms later in Morrison v. California, 291 U.S. 82, 54 S. Ct. 281, 78 L. Ed. 664 (1934), the Supreme Court in a unanimous opinion by Justice Cardozo, followed the ancient English rule that an inconsistent verdict of conviction in a multiple defendant case based on a criminal agreement must be quashed as a matter of due process. In that case the Court found in a state criminal case that the California Supreme Court erred in violation of the Due Process Clause in upholding a conspiracy verdict against one party to an illegal contract for the sale of land when the other party lacked the requisite element of intent and was, therefore, acquitted. Justice Cardozo explained: 

It is impossible in the nature of things for man to conspire with himself. Turinetti v. United States, 2 F. (2d) 15, 17. In California as elsewhere conspiracy imports a corrupt agreement between not less than two with guilty knowledge on the part of each. People v. Richards, 67 Cal. 412, 7 P. 828; People v. Kizer, 22 Cal. App. 10, 14, 133 P. 516; 134 id. 346; People v. Entriken, 106 Cal. App. 29, 288 Pac. 788; DSands v. Commonwealth, 62 Va. 871,21 Gratt. (Va.) 871, 899; Pettibone v. United States, 148 U.S. 197, 203, 205, 13 S. Ct. 542, 37 L. Ed. 419. . . . In such circumstances the conviction of Morrison because he failed to assume the burden of disproving a conspiracy was a denial of due process that vitiates the judgment as to him. Nor is that the only consequence. . . . . The conviction failing as to the one defendant must fail as to the other. Turinetti v. United States, supra; Williams v. United States, 282 Fed. 481, 484; Gebardi v. United States, supra. 

291 U.S. at 92-93 (emphasis added). Under the Due Process Clause of the Fourteenth Amendment, the Supreme Court reversed the conspiracy conviction and followed the English rule in existence when our Constitution was framed. The Court has never questioned the validity of its unanimous due process holding in Morrison . It has never retracted or narrowed the constitutional holding quoted above derived directly from the ancient English rule. See, e.g., Hartzel v. United States, 322 U.S. 680, 682 n. 3, 64 S. Ct. 1233, 88 L. Ed. 1534 (1944) (the Court described two other defendants as "the only co-conspirators of petitioner named in the indictment and the setting aside of their convictions makes it impossible to sustain petitioner's conviction upon the basis of count 7 of the conspiracy count"). In Powell, relied on in error by the majority, the Supreme Court does not even mention Morrison or the traditional rule -- obviously considering it unrelated to Powell's single-defendant, separate count inconsistency, just as the English courts considered the two rules completely unrelated, as Justice Holmes recognized in his opinion in Dunn. Powell is, therefore, entirely irrelevant to the problem before us and cannot be legitimately spun as a justification by the majority in favor of upholding Getsy's execution. 

It is equally misguided for the majority to say that a clear, unanimous constitutional holding in 1934 in Morrison, never overruled or questioned since, does not meet the standard of "clearly established law" found in AEDPA. If a clear rule of law four centuries old, adopted as a matter of Due Process 70 years ago by the Supreme Court, will not meet the AEDPA test, nothing will. 

In response to this dissenting opinion, the majority has attempted to distinguish the Morrison case. It says that Morrison does not apply because (a) one of the two alleged co-conspirators, Santine, was acquitted by a jury rather than by a court, and because (b) the inconsistent punishment in this case was imposed "by different juries in separate trials" instead of in a joint trial. The majority would create a brand new "ancient" rule that is incompatible with the original common law rule and with the Morrison case. They gut the ancient rule of consistent and proportional punishment adopted in Morrison by limiting it to situations where only a judge rather than a jury has acquitted one of the two alleged co-conspirators, and then only after a joint trial. These two exceptions were explicitly rejected by the English common law and by Parliament, as the cases and parliamentary action discussed above clearly demonstrate. The majority refuses to acknowledge that the English common law rule adopted in Morrison applies to jury acquittals in separate trials. 

Justice Cardozo's unanimous opinion in Morrison states the rule it adopts using the same language as the English courts: "It is impossible in the nature of things for a man to conspire with himself . . . . The conviction failing as to one defendant must fail as to the other." 291 U.S. at 93. This language states a general rule and leaves no room for the majority's two exceptions. The Morrison rule does not turn on fortuitous circumstances like whether the trial judge granted a severance and tried the defendant separately, or granted a motion for acquittal rather than letting the case go to the jury. In the present case, Getsy and Santine were indicted jointly but severed for trial. The majority makes the question of life or death in this case turn on the granting of a severance. My colleagues in the majority refuse to carry out the basic purpose of the rule: the elimination of inconsistent and disproportionate punishment among alleged co-conspirators. 

In addition, and equally important, the effect of the majority's exception for separate trials is to make Morrison and the ancient rule completely inapplicable to all modern death penalty cases. The states that continue to use the death penalty bifurcate such trials by conducting a guilt phase trial and then a second trial for imposing the punishment. E.g., Ohio Rev. Code § 2929.03 (describing trial jury's role in determining the sentence for a capital defendant). As a result, trial judges in capital cases now grant a severance and try defendants separately rather than jointly. The bifurcation of all capital trials, together with extensive voir dire of jurors and the present requirements for jury findings of individual aggravating circumstances, makes the conduct of multi-defendant capital trials too complex. Therefore, the current practice in capital cases is to grant a severance and try defendants separately, as in the case of Getsy and Santine. The majority's exceptions mean that the ancient rule of consistency and proportionality of punishment no longer applies in death penalty cases because such trials are not conducted as joint trials. 

The ancient rule was adopted and applied when there were almost 200 crimes in addition to murder carrying the death penalty -- robbery, larceny, burglary, rape, assault, treason, sedition, blasphemy, sodomy, and many others. The ancient rule was designed to eliminate some of the harshness and arbitrariness of the death penalty by introducing a common sense rule of consistency and proportionality among the participants in the same criminal episode. It is ironic, indeed, that the majority has now eliminated the rule in capital cases. What was true for four centuries in such cases -- "it is impossible in the nature of things for a man to conspire with himself" -- is no longer true. The majority is willing to destroy the ancient rule, but the judges are unable to cite a single capital case supporting their position from the entire history of Anglo-American law. No such case has ever suggested, much less applied, the majority's exceptions. The Morrison rule has existed for four centuries only to be effectively overruled today in capital cases by the majority of this Court.