On Monday, in Ramos v. Louisiana, the Supreme Court ruled that the Constitution’s criminal jury requirement—which appears in both Section 2 of Article III and the Sixth Amendment and which applies to the states via the Fourteenth Amendment—forbids states from convicting defendants except by a unanimous jury. In so holding the Court overruled the 1972 decision in Apodaca v. Oregon. In that case, four justices had said that the Constitution does not require unanimous juries for state or federal criminal cases, and one justice—Lewis Powell—had said that while unanimity is required in federal court, principles of federalism give states greater leeway.
The Ramos case will not have nationwide practical significance. Only two states—Louisiana and Oregon—have allowed non-unanimous felony convictions, and Louisiana recently abolished them on a prospective basis. Nonetheless, the disagreement between the lead opinion of Justice Gorsuch and the dissent of Justice Alito highlights an important consideration that bears on the Court’s ability to set aside erroneous precedents: limitations on non-retroactivity.
The Disagreement in Ramos
The bottom-line vote in Ramos was 6-3, with a nonideological division. Joining Justice Gorsuch in the majority were Justices Thomas, Ginsburg, Breyer, Sotomayor, and Kavanaugh, although Justice Thomas did not join any of Justice Gorsuch’s opinion and Justices Sotomayor and Kavanaugh each joined only in part. Joining Justice Alito’s dissent were Chief Justice Roberts (in full) and Justice Kagan (in part). Such nonideological divisions occur with some frequency, but the one in Ramos was unexpected because much of what the justices seemed to be disagreeing about concerned other hot-button issues, including illicit racial motives by legislators and abortion.
Although none of the opinions expressly mentioned abortion, they nonetheless hinted at its relevance through their discussion of the considerations that justify a departure from precedent. The point was especially clear in the dueling concurrences of Justices Sotomayor and Kavanaugh. Sotomayor tried to show how extraordinary Ramos was, thus implying that in a less extraordinary case—like one presenting the question whether to retain a constitutional right to abortion—precedent should prevail. Kavanaugh tried to show that the Court has often overruled its own decisions in setting important new precedents—thus implying that it should be willing to overrule the abortion decisions when given the opportunity.
Yet as important as the debate over the respect owed to precedent may be for forecasting results in abortion cases, it also played a key role in Ramos itself. Even the dissenters did not appear to disagree with the majority that the Constitution is best read to require jury unanimity in all felony trials. Had Ramos been a case of first impression, it probably would have been unanimous itself.
Why then did three justices dissent? The short answer is cost. Justice Gorsuch wrote that “prior convictions in only two States” would be “potentially affected by” the decision but then acknowledged in the very next sentence that as a result, hundreds “of nonunanimous felony convictions” in those two states would likely have to be reversed. However, that cost was worth bearing, Justice Gorsuch went on to say, because “new rules of criminal procedures usually . . . affect significant numbers of pending cases.” He compared the expected impact of Ramos with a 2005 case that necessitated new sentencing hearings for hundreds of federal prisoners.
Yet as Justice Alito explained in dissent, “the burden of resentencing cannot be compared with the burden of retrying cases. . . . In some cases, key witnesses may not be available, and it remains to be seen whether the criminal justice systems of Oregon and Louisiana have the resources to handle the volume of cases in which convictions will be reversed.” Justice Alito and the other dissenters thought the Court ought not to make Oregon and Louisiana pay such a steep price for their reasonable reliance on the Apodaca precedent.
Why Not Non-Retroactivity?
Readers might wonder why the justices did not settle on a seemingly obvious compromise: require unanimous juries in all future cases but do not reverse existing convictions that were obtained using non-unanimous juries prior to the Ramos decision. After all, that is the approach that legislatures generally take when adopting new laws; absent some special consideration, new laws apply prospectively but not retroactively. Non-retroactivity is even a constitutional requirement for new state and federal criminal laws, as enshrined in the Ex Post Facto clauses of Article I, Sections 9 and 10. Why not give states that relied on Supreme Court precedent similar protection?
At one time, the Court’s remedial doctrines expressly considered the cost of retroactivity. Under the approach most closely associated with the 1965 case of Linkletter v. Walker, when the Supreme Court changed the law, it made a judgment about whether to apply the new rule of law retroactively by considering “the prior history of the rule in question, its purpose and effect, and whether retrospective operation will further or retard its operation.” In Linkletter itself, the Court determined that its 1961 decision in Mapp v. Ohio—which applied the Fourth Amendment exclusionary rule to the states—should not be given retroactive effect. To rule otherwise, the Linkletter Court worried, would potentially result in thousands of new trials and thus “seriously disrupt the administration of justice.”
Non-retroactivity has another benefit: It permits a court to decide whether a rule of law is justified on a forward-looking basis without having to worry about its disruptive impact on cases already decided under a different approach. Considerations of retroactivity thus do not impede the law’s development.
Linkletter and Ramos are both illustrative. The justices who decided Mapp might have decided against applying the Fourth Amendment exclusionary rule to the states if doing so would have led to the retrial or even release of thousands of prisoners whose guilt was not in doubt. Likewise, the Ramos dissenters likely would have joined the majority were it not for the worry about the costs imposed on Louisiana and Oregon.
The Linkletter approach had its own difficulties, however. Like other federal courts, the U.S. Supreme Court does not decide abstract questions of law. It only has an occasion to announce a new rule of constitutional law in a criminal case when a convicted defendant challenges some aspect of the trial or other proceedings. Yet if the Court withholds a retroactive remedy, no such defendant will have any incentive to bring the challenge, because no such defendant will benefit from the prospective-only ruling.
Hence, under the Linkletter approach, the Court applied the new rule retroactively to the first defendant to win but not to others. Mapp got the benefit of Mapp v. Ohio even though her trial (obviously) occurred before the Supreme Court ruling in her case; other prisoners whose trials occurred before Mapp v. Ohio did not get the benefit of Mapp. Thus, to create an incentive for litigants to seek new rules, the Linkletter approach treated similarly situated defendants unequally.
Accordingly, in a pair of cases—Griffith v. Kentucky in 1987 and Teague v. Lane in 1989—the Court abandoned Linkletter in favor of the current regime. Under Griffith, the Court applies its new rulings of criminal law and procedure to any defendant whose conviction has not yet become final, which includes defendants whose trials are over but whose appeals remain pending. Under Teague, the Court does not generally give state prisoners who bring federal court habeas corpus petitions the benefit of a rule that was announced after the proceedings in state court were final.
Yet while Griffith and Teague promote one kind of fairness, they can seem arbitrary in their own way. The date on which a conviction becomes final does not necessarily correlate with when the underlying crime allegedly occurred or even when the trial occurred.
Suppose that Jones and Smith are each tried and convicted on the same day as one another for crimes they allegedly committed on the same day as one another. They file simultaneous appeals, but Jones’s case is assigned to a panel that moves swiftly, whereas Smith’s case languishes because one of the appellate judges on his panel is ill. Consequently, Smith’s but not Jones’s appeal is still pending when the Supreme Court announces a new rule that would benefit each of them. Under Griffith, Smith but not Jones receives the benefit of the new rule, even though the respective alleged conduct and trials were simultaneous. Smith and Jones are, in all relevant respects, similarly situated, yet the Griffith rule treats them differently.
What about federal habeas? Teague and a related federal statutory provision enacted in 1996 generally bar the application of new rules to federal habeas petitioners, but Teague articulated two exceptions—one for rules that place primary conduct beyond the reach of the criminal law entirely and another for “watershed rules of criminal procedure” without which there is “an impermissibly large risk that the innocent will be convicted.”
Ramos did not involve the first exception; no one would contend that murder (the crime of conviction) is beyond the reach of the criminal law. The case did, however, arguably implicate the second Teague exception. In his lead opinion, Justice Gorsuch specifically avoided deciding whether the juror unanimity requirement is the sort of watershed rule that would be available in a habeas corpus proceeding. By contrast, in his concurrence, Justice Kavanaugh said that juror unanimity is not a watershed rule, but his analysis is not controlling. Accordingly, as the dissenters noted, the case leaves open the second Teague exception question.
There is some uncertainty about whether the watershed-procedural-rule exception to the Teague doctrine is available in most habeas cases, because the key statutory provision does not expressly include that exception. However, a related statutory provision does seem to assume that some new rules apply even in federal habeas proceedings. Meanwhile, in a 2016 case the Court held that the primary-conduct exception is constitutionally required, while leaving open the question whether the watershed-procedural-rule exception is also constitutionally required.
Future litigation will determine whether prisoners in Louisiana and Oregon whose convictions are final but whose time to file federal habeas petitions has not yet expired may do so in reliance on Ramos. Whatever the answer to that question, it is clear that under the Griffith doctrine, prisoners whose convictions by non-unanimous juries are not yet final will be entitled to new trials.
Is that a problem? Maybe. In his dissent, Justice Alito speculated that “[i]n many cases, if a unanimous vote had been needed, the jury would have continued to deliberate and the one or two holdouts might well have ultimately voted to convict.” Perhaps, but many other cases might have resulted in hung juries and even an occasional acquittal if one or two holdouts were given their due.
The dissenters may well be wrong in Ramos. Jury unanimity is a very important principle that could be said to justify imposing the considerable cost of retroactive application to several hundred not-yet-final cases in Louisiana and Oregon.
Yet such a calculation will sometimes come out the other way. There will be cases in which the best rule going forward imposes too great a cost when applied retroactively. The Court’s inability to choose the best rule in such cases is itself a cost of our constitutional structure. In some U.S. states and a great many foreign countries, courts can issue advisory opinions with purely prospective effect. For better or (in my view, mostly) worse, our Supreme Court does not think it has that power, and so it sometimes must choose between the best rule and the best rule available given retroactivity’s cost.