When the Supreme Court announces a constitutional rule — such as Miranda’s warning or Gideon’s right to counsel — courts must decide whether to undo already-final criminal convictions that would violate the new rule. Teague v. Lane1 created a general presumption that new rules are not retroactive on collateral review, but with two exceptions: (1) substantive rules — those preventing criminalization of particular conduct2 — and (2) “watershed” procedural rules.3 In Miller v. Alabama,4 the Court prohibited mandatory life-without-parole (LWOP) sentences for juveniles.5 After Miller, courts divided over whether the new rule qualified for retroactivity.6 Last Term, in Montgomery v. Louisiana,7 the Supreme Court held that states are constitutionally required to give retroactive effect to new substantive rules and that Miller announced a substantive rule. But Miller is more naturally read as a procedural rule of individualized sentencing for juveniles. Montgomery’s treatment of Miller illuminates Teague’s central failing: an impossible bar for procedural rules. To recalibrate retroactivity doctrine to better balance fairness against finality, the Court might have instead held that Miller was animated by a watershed principle — that “kids are different” under the Eighth Amendment — and should apply retroactively for that reason.
The Eighth Amendment demands proportionality between crime and punishment.8 Roper v. Simmons,9 which banned the death penalty for juvenile offenders,10 launched a subset of proportionality analysis that focuses on the unique characteristics of youth: because of their impulsivity and inability to appreciate consequences, the death penalty is categorically disproportionate to their crimes.11 Then, in Graham v. Florida,12 the Court found it categorically cruel and unusual to sentence a juvenile to LWOP for a nonhomicide offense. By analogizing juvenile LWOP to the death penalty, Graham drew from the Eighth Amendment cases that mandated individualized sentencing in the death penalty context.13 Miller, in turn, extrapolated from Graham and ruled that mandatory LWOP for juveniles is cruel and unusual, even for juveniles convicted of homicide. Under Miller, as in death penalty sentencing, courts must consider mitigating factors, including youth.14 But the rule’s retroactivity was unclear. Enter Montgomery.
Henry Montgomery was seventeen at the time of his crime, a tenth-grade truant caught in a park by Charles Hurt, a deputy sheriff. While being frisked, Montgomery pulled out a pistol and shot and killed the deputy.15 He was initially sentenced to death, but because “public prejudice had prevented a fair trial” of the black teenager, he was retried.16 In the second trial, the verdict was “guilty without capital punishment,” automatically triggering an LWOP sentence.17 Thus, no separate sentencing phase offered Montgomery a chance to introduce mitigating evidence,18 such as his age’s effect on both his impulse control and ability to reform. After Miller, Montgomery, then in his sixties, sought collateral review in state court. State courts denied his petition, refusing to apply Miller retroactively.19 The U.S. Supreme Court granted certiorari to review the Louisiana Supreme Court’s retroactivity analysis.
The Court held that the Federal Constitution requires retroactive application of new substantive rules and that Miller’s prohibition on mandatory LWOP for juveniles was substantive. Writing for the majority, Justice Kennedy20 first analyzed the Court’s jurisdiction to review the retroactivity analysis on state habeas, an inquiry that boiled down to whether retroactivity is a constitutional mandate. Danforth v. Minnesota21 suggested that Teague’s prescription “was an exercise of [the] Court’s power to interpret the federal habeas statute,”22 and by implication not a constitutional mandate. But Justice Kennedy argued that Danforth considered only Teague’s general presumption against retroactivity and never addressed the distinct question whether its two exceptions were constitutionally required.23 Now the Court held that Teague’s first exception for substantive rules is a constitutional mandate.24 The Court reasoned that unlike procedural rules, which “are designed to enhance the accuracy of a conviction or sentence by regulating ‘the manner of determining the defendant’s culpability,’”25 substantive rules are “categorical constitutional guarantees that place certain criminal laws and punishments altogether beyond the State’s power to impose.”26 And whether a conviction became final before a new substantive rule was announced is irrelevant: “There is no grandfather clause that permits States to enforce punishments the Constitution forbids.”27 The principle applies to states by virtue of the Supremacy Clause; if a state allows collateral review, it must apply substantive constitutional rules retroactively.28
Having established its jurisdiction, the majority turned to Miller’s retroactivity. First, the majority situated Miller as a natural extension of Roper and Graham, grounded in the idea that children are different: their immaturity leads to impulsivity, they are more easily influenced by family and peers and less able to avoid criminogenic social settings, and above all, they have greater capacity for rehabilitation.29 These factors combine to reduce culpability and to lessen the force of the traditional justifications for the harshest punishments — retribution, deterrence, and incapacitation. Miller held that, because of the special characteristics of youth, mandatory LWOP sentences for juveniles “pos[e] too great a risk of disproportionate punishment”; thus, the Eighth Amendment requires individualized sentencing.30 But, according to the majority, beyond a procedural requirement of individualized sentencing, Miller “established that the penological justifications for life without parole collapse in light of ‘the distinctive attributes of youth’” and that “sentencing a child to life without parole is excessive for all but ‘the rare juvenile offender whose crime reflects irreparable corruption.’”31 Thus, Miller was substantive: “[I]t rendered life without parole an unconstitutional penalty for ‘a class of defendants because of their status’ — that is, juvenile offenders whose crimes reflect the transient immaturity of youth.”32
Justice Kennedy acknowledged that Miller seemed plainly procedural, that it explicitly required only “a certain process” and did not categorically bar LWOP for juveniles.33 He argued, however, that the seemingly procedural rule becomes substantive when one focuses on the relevant class for which Miller does categorically bar LWOP: “[A]ll but the rarest of juvenile offenders, those whose crimes reflect permanent incorrigibility.”34 Justice Kennedy held that Miller’s procedural mandate merely operationalized a substantive rule by identifying that subset class. Finally, he explained that Miller’s retroactivity would not overburden states, which may simply offer parole hearings.
In dissent, Justice Scalia35 denounced the majority’s “astonishing” jurisdictional analysis as the “conscription into federal service of state postconviction courts.”36 The history of retroactivity provided no support for the idea that Teague’s rules are constitutionally compelled,37 and the Supremacy Clause argument merely restated the question whether old or new federal law should govern.38 He criticized the majority’s “sleight of hand” and use of “cherry picked” dicta, concluding that “[a]ll that remains to support the majority’s conclusion is that all-purpose Latin canon: ipse dixit.”39 Justice Scalia argued that Teague’s central aim was to protect finality and that the majority’s jurisdictional analysis undermined that policy entirely.40 By finding that Teague’s substantive exception is a constitutional requirement, the majority “not only forecloses Congress from eliminating this expansion of Teague in federal courts, but also foists this distortion upon the States.”41
Justice Scalia then turned to the merits, arguing that “[h]aving distorted Teague, the majority simply proceeds to rewrite Miller.”42 He highlighted Miller’s explicit procedural language and rejected the majority’s argument that Miller made juvenile LWOP unconstitutional in all but the rarest cases.43 Justice Scalia lamented that the “rewriting” of Miller means that even if a sentencing court determines that a crime does reflect permanent incorrigibility, a juvenile defendant may continue to protest that the sentencing court simply got it wrong, that the crime reflected only “transient immaturity,” and therefore that the sentence violates the Eighth Amendment.44 He criticized the majority’s “not-so-subtle invitation” to remedy Miller violations not via resentencing, which would be a “practical impossibility,” but instead by offering parole hearings.45 He contended that a justification for Roper’s elimination of the death penalty was states’ ability to sentence juveniles to LWOP and accused the majority of disingenuously refusing to outright ban the sentence for juveniles, while effectively making that sentence impossible to impose.46
Montgomery constitutionally embeds Teague while simultaneously illuminating a twofold problem: the fiction of a clear substance-procedure divide and the impossibility of the watershed exception. While the Court rightly applied Miller retroactively, it failed to acknowledge these doctrinal difficulties. To rebalance fairness concerns with long-dominant finality concerns and to acknowledge the shortcomings of the substance-procedure divide, the Court should pry open the impossible retroactivity standard for procedural rules. The Montgomery Court, for instance, might have acknowledged that Miller is most naturally read as a procedural rule and held that where a watershed principle of fairness revealed in cumulative cases — such as that “kids are different” — animates a significant procedural rule,47 that rule should apply retroactively.
Retroactivity is about the scope of postconviction review — whether a defendant may challenge a conviction or sentence with law announced after the conviction or sentence became final. The difficulty lies in balancing competing interests: states’ interests in finality and autonomy in the federal system against criminal defendants’ interests in accurate convictions and fair sentences.48 On the one hand, finality in criminal proceedings might be defensible as protective of both states’ and defendants’ rights. By conserving state resources, finality is fair to sovereign states in the federal system,49 and channeling resources to trials themselves, rather than to posttrial review, may increase both accuracy and efficiency.50 Further, this resource conservation may ultimately promote the recognition of new rights, which courts may otherwise hesitate to recognize if they fear expansive retroactivity’s burden on states.51 Yet weighty values also militate against strict finality. Above all, it is arguably unfair to punish in a manner that is now understood to be inconsistent with the Constitution, and it is likewise unfair to subject one defendant to punishment from which a similarly situated future defendant will be spared simply by virtue of timing.52
But the doctrine has veered too far toward finality and away from fairness. Before Teague, retroactivity was governed by a multifactor inquiry.53 Concerned about predictability, the Teague Court sought to create a bright-line framework for retroactivity on collateral review.54 The Court approximated a bright line by delineating substance from procedure and sharply limiting the retroactivity of procedural rules. The Teague Court envisioned that the narrow watershed exception would encompass only “procedures without which the likelihood of an accurate conviction is seriously diminished” and called it “unlikely that many such components of basic due process have yet to emerge.”55
Two and a half decades later, Teague has done more than paint bright lines on the retroactivity highway; it has closed down most of the lanes.56 In twenty-seven years, the Court has never declared a new procedural rule “watershed.”57 A year after Teague, the Court held that the rule announced in Caldwell v. Mississippi58 prohibiting prosecutors from characterizing the jury’s role in death penalty sentencing as merely advisory was not watershed.59 A rule going to accuracy in sentencing, in other words, was insufficient. Rather, to constitute a watershed, the rule must “‘alter our understanding of the bedrock procedural elements’ essential to the fairness of a proceeding.”60 More recently, in Schriro v. Summerlin,61 the Court held that the new rule that a jury, rather than a judge, must find the aggravating factors making a defendant eligible for the death penalty was not watershed.62 Writing for four dissenting Justices, Justice Breyer argued that the majority’s rote application of Teague failed to account for the competing interests that Teague attempted to balance.63 And academic criticism has likewise centered on the Court’s overly harsh restriction on the retroactivity of new procedural rules — that is, on the justifiability of the fairness-finality balance it has struck.64
Forced to characterize new rules as substantive in order to fairly grant retroactivity, the Court has engaged in unnatural readings of precedent, both opening itself to criticisms of “sleight of hand”65 and perpetuating an oversimplified dichotomy between substance and procedure. Miller, for example, resists easy classification. In one sense, the decision prohibited a type of punishment altogether for a class of persons — that is, if you consider mandatory LWOP to be the punishment and juveniles to be the class. In another sense — the one Justice Kennedy focused on — Miller prohibited a broader type of punishment (LWOP) for a narrower class of persons (non–permanently incorrigible juveniles). And in yet another sense, “mandatory” might simply be procedural scaffolding on “life without parole” — a punishment expressly not prohibited by Miller — and thus purely procedural. Given Miller’s declaration that it required only “a certain process,” it is most naturally read this third way.66 If a rule can be read as procedural, substantive, or both — as Miller seemingly allows — Teague provides insufficient guidance. And if the procedural category is functionally unavailable, the important retroactivity determination is dependent on whether the rule can be plausibly crammed into the substantive exception, not on whether fairness outweighs finality.
One way for the Montgomery Court to improve the Teague doctrine would have been to credit an animating watershed principle: for instance, “children are different.”67 This watershed principle reflects a greater appreciation of juvenile psychology.68 Neurological adolescence is associated with diminished capacity for “planning, motivation, judgment, and decision-making, including the evaluation of future consequences, the weighing of risk and reward, the perception and control of emotions, and the processing and inhibition of impulses.”69 As Miller recognized,70 family and home environments can exacerbate those problems.71 Troublingly, juveniles sentenced to LWOP disproportionately experience difficult childhoods: 79% regularly witnessed violence in their homes, 40% had been enrolled in special-needs education, and 47% were physically abused.72 Severe juvenile punishment is thus more difficult to justify under any penological theory. Retribution is less justified in light of diminished culpability, deterrence is less likely to be effective given what we now know about juvenile brain development and impulse control, and children’s greater capacity for change makes indefinite incapacitation simply unnecessary. “Miller demonstrate[s] that our societal understanding of fairness in the realm of juvenile sentencing has fundamentally changed.”73 Where the very foundations of fairness itself have so changed and a procedural rule is necessary to implement that watershed principle of fairness — as individualized sentencing is necessary to treat children like children, capable of change — fairness dictates that the rule apply retroactively.
While the Court got to the just result here, its uncharacteristically expansive jurisdictional analysis74 may have constitutionally embedded75 a broken doctrine. An overly harsh retroactivity standard undermines already incremental legal protections, and the law becomes twofold “a barrier to reform rather than a facilitator of it.”76 Applying a watershed-principle theory would have allowed the Court to acknowledge that Teague’s substance-procedure dichotomy is imperfect, to counsel lower courts to conduct the best categorization of a new rule, and to apply retroactively those procedural rules where fairness interests outweigh finality. “Miller’s central intuition,” the Court acknowledged, is that children who commit even heinous crimes are capable of change.”77 That premise is novel; its application in Miller, watershed.