Henry Montgomery has survived the remarkable arc of the Supreme Court’s evolution juvenile sentencing. In 1970, Louisiana sentenced him to die in prison for the murder of a police officer, a crime he committed when he was seventeen years old. The sentence was mandatory, and it was perfectly legal. At that time it was also perfectly legal to execute juveniles. A generation later, the Supreme Court barred the execution of children under age sixteen in 1988, but the next year refused to extend the bar to all juveniles. Not until 2005 did the Court exempt all juveniles from the death penalty. In half a decade, the Court ruled that juveniles could not be imprisoned for life without any possibility of release for non-homicides. A mere two years later, yet forty-six years after Mr. Montgomery’s conviction, the Court declared, in Miller v. Alabama, that mandatory life sentences like Mr. Montgomery’s were unconstitutional.
Miller confirmed the lessons of these prior decisions that children’s youth and immaturity make them categorically different for sentencing purposes, and that life imprisonment without parole is akin to the death penalty for juveniles. Thus, automatically sentencing children to a lifetime of imprisonment “poses too great a risk of disproportionate punishment.” The Eighth Amendment’s protection against “cruel and usual punishments” therefore prohibits such sentences.
But Mr. Montgomery’s path to a hope for release was not yet complete. In fact, it was cut off by the Louisiana Supreme Court. That court ruled Mr. Montgomery could not benefit from Miller because, under the United States Supreme Court’s decision in Teague v. Lane, Miller did not apply retroactively to cases that were already final at the time of the decision. Mr. Montgomery’s case became final in 1984, thirty years too soon.
In Montgomery v. Louisiana, the Supreme Court reversed the Louisiana Supreme Court and held that Miller applies retroactively. The Court found that, by categorically prohibiting life sentences for the majority of juveniles whose crimes reflect “transient immaturity” rather than “irreparable corruption,” Miller announced a substantive rule of criminal law that is not subject to Teague’s general bar against retroactivity. Now, unless Louisiana can show that the crimes of those like Montgomery demonstrate “irreparable corruption,” it must grant them meaningful hope of “some years of life outside prison walls.”
As discussed below, Montgomery affirmed Court’s supremacy in declaring federal law while bolstering the significant limits that Miller places on states’ ability to condemn any juvenile to die in prison. But the Court left unresolved a critical question: how much hope for release is enough? Whatever the answer, it must account for Miller’s impact on the obligation of states to grant parole to juveniles facing lifelong incarceration. This article asserts that Miller cabins the state’s power to deny parole permanently to reformed juveniles. It does so by creating a modest, but absolute, liberty interest in release before death for rehabilitated youth. The Supreme Court, rather than state parole systems, must be the ultimate protector of this right.
I. How the Montgomery Court Decided It Could Decide
The central issue in Montgomery was whether Miller applied retroactively. However, the Court first had to determine whether it could even decide the case. In addition to granting certiorari on retroactivity, the Court sua sponte inserted the question of whether it had jurisdiction to review Louisiana’s determination that Miller was not retroactive. Both Mr. Montgomery and Louisiana agreed the Court had jurisdiction, mostly likely since the Louisiana Supreme Court expressly relied on the Supreme Court’s decision in Teague v. Lane, thus creating a federal question for Supreme Court review. Unsatisfied, the Court appointed a special amicus to argue against jurisdiction.
At stake was what it means for the Court to declare a rule retroactive. In Teague and subsequent cases, the Court identified the federal habeas statute as the source of its authority to apply rules retroactively. If that is true, then Teague only binds federal habeas courts. That leaves Congress free to alter Teague under its power to define the jurisdiction of the lower federal courts. That also leaves states free to fashion their own rules for applying new rules in their own courts. In fact, the Louisiana Supreme Court expressly declared that it was not obligated to follow Teague. However, if the Court’s power to announce and apply new rules retroactively derives from the Constitution, states have no discretion on applying new federal constitutional rules in their courts.
The Court partially addressed these issues in Danforth v. Minnesota. The Minnesota Supreme Court, under state law, gave retroactive effect to the Supreme Court’s decision in Crawford v. United States, which announced a new rule granting criminal defendants the right to cross-examine any testimonial statement at trial. Minnesota did so despite the fact that the Supreme Court had earlier held in Whorton v. Bockting that Crawford was not retroactive. While the Danforth Court acknowledged that Whorton did not require states to apply Crawford to cases on collateral review, it nonetheless held that states could choose to apply Crawford retroactively under state law.
Central to Danforth was the Court’s framing of retroactivity as deciding what remedies are available for constitutional violations. Under this view, Teague identifies the remedies available in federal habeas. But states are free to provide broader remedies under their own law. Shrugging off concerns that this approach undermined the uniform enforcement of federal law—a key value in Teague—the Court justified the result as a necessary consequence of the states’ independent sovereignty. Thus, “the remedy a state court chooses to provide its citizens for violations of the Federal Constitution is primarily a question of state law.”
Chief Justice Roberts, in a dissent joined by Justice Kennedy, countered that retroactivity of constitutional rules is always an issue of federal law. The Supremacy Clause therefore binds states to the Court’s retroactivity decisions. Roberts argued that retroactivity does not involve a choice of available remedies, but a choice of law, specifically, whether the old law or the new law applies. Only the Supreme Court may make this choice for constitutional rules. Roberts dismissed the majority’s grounding of Teague in the federal habeas statute as irrelevant, since Congress’ authority over the lower federal courts cannot usurp the Court’s prerogative to decide federal law.
Despite the majority’s broad language empowering states to decide what relief they afford to constitutional rights, it limited Danforth to situations where states provide broader relief on collateral review than what is required under Teague. The Court declined to resolve if states must apply retroactive rules in post-conviction, or if Congress could amend Teague by statute.
The debate in Danforth over the meaning of retroactivity helps explain why the Court insisted on addressing its jurisdiction in Montgomery. The Court, in an opinion written by Danforth dissenter Justice Kennedy, held that states must give retroactive effect to new substantive rules, those which prohibit states from criminalizing certain conduct or imposing certain punishments for a class of defendants. The Court explained that new substantive rules categorically barring penalties are required by the Constitution itself. Any penalty that violates a substantive rule is therefore void and unenforceable, regardless of when the defendant’s conviction became final.
Though the majority did not mention the “choice of remedies” versus “choice of law” conflict that animated Danforth, Montgomery unmistakably adopts the choice of law view, at least for substantive rules. The majority also did not address whether Congress could overrule Teague. Nonetheless, Congress clearly may not, again, at least with respect to new substantive rules. Montgomery thus preserves the Court’s supremacy vis-a-vis the states and Congress in declaring new substantive rules
Beyond settling these fundamental issues of federalism and the separation of powers, the practical import of Montgomery’s jurisdictional ruling is elusive. Had the Court held it did not have jurisdiction in Montgomery, it could have eventually addressed Miller’s retroactivity via a federal habeas case. Several such cases were pending when the Court issued Montgomery, and all of the Court’s post-Teague retroactivity decisions arose from federal habeas. A decision out of federal habeas would have had the same effect as a decision out of state collateral review: states would have to apply Miller if it was retroactive, or they would be left to decide whether to apply Miller if it was not.
II. The Montgomery Court’s Shaping of Miller into a Categorical Rule
Having established its authority to decide whether Miller announced a retroactive substantive rule, the Court turned to this main issue. The Court first determined that substantive rules must categorically limit the states authority to either define crimes or impose a punishment on a class of persons. By this rubric, Miller’s chances of qualifying as substantive were grim. Miller pointedly stated that it was not categorically barring a sentence for a class of offenders, and that it was only mandating that states follow a certain process—considering a juvenile’s youth and attendant circumstances—before imposing punishment. For this reason, many who argued that Miller was substantive relied on the Court’s earlier suggestion that substantive rules include, but are not necessarily limited to, categorical guarantees. To assert otherwise would seemingly require the Court to contradict itself.
But that is essentially what the Court did. The Court minimized its statement in Miller as too general. It claimed that, while Miller did not bar a penalty for all juveniles, Miller did ban life imprisonment without release for all juveniles whose crimes do not reflect irreparable corruption. Fortunately, the Court’s innovation has precedent, even if Miller is not that precedent. As the Montgomery Court suggested, creating a categorical bar is an exercise in line drawing. Roper drew a line banning the death penalty for juveniles. Graham drew a more specific line prohibiting life imprisonment for juveniles convicted of non-homicide offenses. In the death penalty, the Court has drawn a host of lines exempting those with a mental disability, those who are insane, those who have not committed a homicide, and those who did not intend to commit a homicide. The Court has extended this line drawing beyond extreme sentencing to prohibit any punishment for “status offenses,” such as being addicted to drugs or too poor to pay a fine.
Most importantly, Montgomery applied Miller’s line drawing to all life without parole sentences imposed on juveniles, though Miller specifically addressed the constitutionality of mandatory life without parole for juveniles. This likely means that states must provide collateral review for all juveniles serving life without the possibility of parole, regardless of whether those sentences were mandatory. Think of it this way: the Court’s prohibition in Atkins v. Virginia on executing the mentally disabled is a retroactive, substantive rule. It gave every death row inmate the right to establish they belonged to the new category of individuals exempt from execution. By this same reasoning, Montgomery grants every juvenile serving life imprisonment without parole the right to establish that their crimes reflected “transient immaturity,” the category of individuals now exempt from a death-in-prison sentence.
III. The Adequacy of Parole Under Miller
The Montgomery majority anticipated the chaos of states holding sentencing hearings that require the state to prove that a juvenile was irreparably corrupt at the time of an offense from decades ago. It therefore suggested states could avoid these difficulties by granting juveniles parole. This conclusion necessarily assumes that access to parole would protect juveniles from dying in prison by allowing them to demonstrate their maturity and capacity for change.
Yet, if the central premise of Miller is that, regardless of their crimes, nearly every juvenile must have a meaningful opportunity for some years outside prison walls, it is not at all clear that parole always satisfies this standard, either doctrinally or practically. Doctrinally, the Supreme Court has refused to recognize a constitutional right to release before the end of a valid sentence. States are not obligated to even have a parole system, and, when they do, they are free to define the factors relevant to release. One’s expectation of release is thus entirely up to the state.
Practically speaking, individuals convicted of violent offenses stand little chance of being granted parole in many states. As Professor Sarah French Russell explains, “[t]he nature of the crime of conviction is often the driving force in parole decisions.” This status quo is deadly for juveniles whose only chance at release may be their state’s parole system. It demonstrates that merely offering parole cannot reliably guarantee that a juvenile whose crime did not reflect irreparable corruption might still die in prison.
For Miller truly to restore juveniles’ hope for release, the ruling must regulate not only the state’s sentencing authority, but also its obligation to grant juveniles parole. Though Miller most directly guarantees juveniles an opportunity at release, the decision contains the corollary that juveniles who demonstrate maturity and rehabilitation as adults must eventually be released. Reformed juveniles could of course be denied release for some minimum amount of time to serve the penological interests of retribution, deterrence, and rehabilitation. But at some point those interests must yield to the juvenile’s right to release. Determining that point cannot be left to the whim of state executive branches. It would require judicial review of whether a parole system complied with Miller. Without this check on a state’s ability to deny parole, Miller would become an empty promise.
At a minimum, Miller’s liberty interest in release before death also allows juveniles to challenge deficient parole procedures. For instance, Miller likely limits a state’s ability to maintain release factors that facially undermine the decision’s intent. These include allowing a parole board to deny release based on the charge alone and without considering a juvenile’s age and immaturity at the time of the offense, or to treat a juvenile’s age at the time of the offense as an aggravating, rather than a mitigating, circumstance.
A form of this challenge has succeeded in Michigan, a state with approximately 360 juveniles serving mandatory life sentences without parole, 334 of which were on collateral review. In 2013, as part of a civil rights action brought under 42 U.S.C. § 1983, a federal district court invalidated Michigan’s mandatory sentencing scheme as applied to juveniles. Rather than ordering resentencings—a remedy unavailable in § 1983 civil rights suits—the district court ordered the parties to propose a parole system that would afford juveniles a meaningful opportunity for eventual release. The case is pending at the Sixth Circuit.
However, the case’s fate is uncertain. Because of Montgomery, all Michigan juveniles formerly serving mandatory life imprisonment are entitled to resentencing under Michigan law. These juveniles now face a sentencing range between a minimum term of twenty-five to forty years, and a maximum term of not less than sixty years. They may only receive life without parole if the prosecutor moves to seek the sentence and the court holds an individualized sentencing hearing. This relief under state law may moot the §1983 action. Regardless, the case could offer a blueprint for Miller to reform juvenile parole procedures.
While its potential for improving the parole process for juveniles is promising, Miller’s regulation of parole systems extends beyond procedure. Montgomery reading of Miller leaves states no room to disagree with the Court’s assessment that juvenile life without parole should be rare. Thus, even with appropriate procedures in place, a state’s parole system arguably still violates Miller if it fails to make irrevocable life sentences sufficiently rare. That means a juvenile could contest a parole system under the Eighth Amendment by demonstrating that too many youth still unduly risk serving death-in-prison sentences. Even if a state’s parole process succeeds in making life imprisonment rare for juveniles, it would remain vulnerable to attack if impermissible factors like race result in arbitrary parole outcomes.
Juveniles who wish to show they are trapped in an unconstitutional parole process face daunting challenges. The first is empirical. Establishing that a parole system does not reliably release juveniles requires building a robust record on parole release decisions. Ideally, that record would establish that deserving juveniles were routinely being denied relief. Since most states do not provide access too much or all of the information the parole board relies on to make decisions, this may prove an insurmountable barrier for most juveniles. The second barrier is temporal. Montgomery does not force states to release juveniles immediately, and it offers no guidance on how long states may delay the decision. As a result, release data for any single year, or even several years, may not suffice to challenge a parole system. A conservative release rate one year says nothing about release rates for the next year, especially if there are changes to the parole authority’s composition or political context. The third barrier is conceptual: how “rare” is “rare”? The Court provided no standards for this evaluation. It also did not require any particular fact-finding before a state denies release, leaving no objective basis for judging a parole system’s faithfulness to Miller. The Court’s vagueness in defining the contours of a new rule is not unusual, and future cases may force the Court to speak more clearly.
The fourth, and perhaps most daunting, barrier is institutional. Parole typically aims to gauge an inmate’s fitness to re-enter society. But prison is a terrible place to reform. This is especially true for children, most of whom enter prison with troubled backgrounds, only to be further traumatized by violence and abuse while in custody. Those not broken by this experience may carry a disciplinary record that bars them from enrichment programs. The problem of access is more acute for juveniles newly granted the opportunity for parole after Miller, many of whom are in states that deny reformative programming to those serving life sentences. Consequently, juveniles seeking parole must survive prison environments designed to obscure their path to redemption. The risk that prison will irreparably rob juveniles of their capacity to reform will cloud any inquiry into whether a parole board is properly releasing juveniles.
These obstacles emphasize the need for advocates to ensure that Montgomery’s nod to parole does not allow states to turn their backs on Miller. Restoring the hopes of juveniles for release demands efforts in the courts and state legislatures to reform parole procedures, gather data on release outcomes, concretize Miller’s meaning and improve prison conditions.
When Henry Montgomery stepped inside Angola forty-six years ago, he literally did not have the right to hope he would take another breath in the free world. Miraculously, now he does. To resolve the often-dry subjects of jurisdiction and retroactivity, Montgomery canvassed the meaning of federalism, the separation of powers, and ultimately, hope itself. Nonetheless, the decision will be of little import to those like Mr. Montgomery if the Court allows states simply to thwart juveniles’ opportunity for release with an unaccountable parole process. Miller and Montgomery therefore must extend to the parole process and its outcomes, an interpretation easily supported by Montgomery and the Court’s recent juvenile sentencing decisions. Realizing substantive limitations on parole, a process traditionally left to executive discretion, likely will not come any time soon, if at all. But Mr. Montgomery’s story provides some measure of the promise for hope to come.
* Brandon Buskey is a senior staff attorney for the Criminal Law Reform Project of the American Civil Liberties Union.
 State v. Montgomery, 242 So.2d 818 (La. 1970).
 Thompson v. Oklahoma, 487 U.S. 815 (1988).
 Stanford v. Kentucky, 492 U.S. 361 (1989).
 Roper v. Simmons, 543 U.S. 551 (2005).
 Graham v. Florida, 560 U.S. 48 (2010).
 Miller v. Alabama, 132 S. Ct. 2455 (2012).
 Id. at 2469.
 State v. Montgomery, 141 So.3d 264 (La. 2014).
 State v. Tate, 130 So.3d 829 (La. 2013).
 Teague v. Lane, 489 U.S. 288 (1989).
 Montgomery v. Louisiana, 136 S. Ct. 718 (2016).
 Id. at 736-37.
 Id. at 727.
 State v. Tate, 130 So.3d 829, 829 (La. 2013).
 See, e.g., Michigan v. Long, 463 U.S. 1032, 1040-41 (1983) (Douglas, J., dissenting).
 Danforth v. Minnesota, 552 U.S. 264, 277 (2008).
 State ex rel. Taylor v. Whitley, 606 So.2d 1292 (La. 1992).
 See Danforth, 552 U.S. at 282.
 Crawford v. United States, 541 U.S. 36 (2004).
 See Whorton v. Bockting, 549 U.S. 406, 409 (2007).
 See Danforth, 552 U.S. at 282.
 Id. at 275.
 Id. at 282.
 Id. at 288.
 Id. at 291 (Roberts, C.J., dissenting).
 Id. at 307.
 Id. at 308-09.
 Id. at 269 n.4 (majority opinion).
 Montgomery v. Louisiana, 136 S. Ct. 718, 729 (2016).
 Id. at 729-30.
 Id. at 731-32.
 See id. at 741 (Scalia, J., dissenting).
 See, e.g., Davis v. McCollum, 798 F.3d 1317 (10th Cir. 2015), cert. denied, No. 15-8005, 2016 WL 531245 (Jan. 22, 2016).
 See, e.g., Whorton v. Bockting, 549 U.S. 406, 413 (2007).
 Montgomery, 136 S. Ct. at 729.
 Miller v. Alabama, 132 S. Ct. 2455, 2471 (2012).
 See, e.g., Brandon Buskey & Daniel Korobkin, Elevating Substance Over Procedure: The Retroactivity of Miller v. Alabama Under Teague v. Lane, 18 CUNY L. Rev. 21, 26 (2014) (arguing that Miller announced a substantive obligation which should be found retroactive under Teague); Beth A. Colgan, Alleyne v. United States, Age as an Element, and the Retroactivity of Miller v. Alabama, 61 UCLA L. R. Disc. 262, 263 (2013).
 Montgomery, 136 S. Ct. at 734.
 Atkins v. Virginia, 536 U.S. 304, 305 (2002).
 Ford v. Wainwright, 477 U.S. 399, 400 (1986).
 Kennedy v. Louisiana, 554 U.S. 407, 421 (2008).
 Enmund v. Florida, 458 U.S. 782, 787 (1982).
 Robinson v. California, 370 U.S. 660, 678 (1962).
 Tate v. Short, 401 U.S. 395, 397 (1971).
 Atkins v. Virginia, 536 U.S. 304 (2002).
 Penry v. Lynaugh, 492 U.S. 302, 303 (1989), overruled by Atkins, 536 U.S. at 304.
 Montgomery, 136 S. Ct. at 735.
 Id. at 736.
 See Greenholtz v. Inmates of Neb. Penal & Corr. Complex, 442 U.S. 1, 7 (1979).
 Id. at 7-8.
 Id. at 11.
 Sarah F. Russell, Review for Release: Juvenile Offenders, State Parole Practices, and the Eighth Amendment, 89 Ind. L. J. 373, 397 (2014).
 In this way, Miller distinguishes juveniles from the typical adult seeking parole. Because states have no obligation to release adults before the end of their sentence, a valid conviction terminates an adult’s federal liberty interest in release. Not so with juveniles under Miller.
 Buskey & Korobkin, supra note 40, at 25.
 Hill v. Snyder, No. 10-14568, 2013 WL 364198 (E.D. Mich. Jan. 30, 2013).
 Heck v. Humphrey, 512 U.S. 477, 487 (1994).
 Mich. Comp. Laws Ann. § 769.25(6) (West 2016).
 Cf. Furman v. Georgia, 408 U.S. 238 (1972).
 Russell, supra note 57, at 397.
 Compare Gerstein v. Pugh, 420 U.S. 103 (1975) (holding that states must make probable cause determinations without undue delay for incarcerated arrestees), with Cnty. of Riverside v. McLaughlin, 500 U.S. 44 (1991) (holding that states must make probable cause determination without undue delay, but setting presumptive limit of 48 hours).
 See Jessica Lahey, The Steep Costs of Keeping Juveniles in Adult Prisons, Atlantic, Jan. 8, 2016, http://www.theatlantic.com/education/archive/2016/01/the-cost-of-keeping-juveniles-in-adult-prisons/423201/ [https://perma.cc/JC5H-FKHA].
 See Ashley Nellis, Throwing Away the Key: The Expansion of Life Without Parole Sentences in the United States, 23 Fed. Sent’g Rep. 1, 29 (2010).