13 Cal.5th 152 (2022)
509 P.3d 975
293 Cal. Rptr. 3d 623

IN THE SUPREME COURT OF CALIFORNIA

THE PEOPLE,

Plaintiff and Respondent,

v.

MARIO SALVADOR PADILLA,

Defendant and Appellant.

S263375

Second Appellate District, Division Four

B297213

Los Angeles County Superior Court

TA051184

May 26, 2022

Justice Liu authored the opinion of the Court, in which Justices Kruger, Groban, and Jenkins concurred.

Justice Corrigan filed a dissenting opinion in which Chief Justice Cantil-Sakauye and Justice Perren[1]* concurred.

Counsel:

Jonathan E. Demson, under appointment by the Supreme Court, for Defendant and Appellant.

Cyn Yamashiro, Markéta Sims; Susan Lynn Burrell and L. Richard Braucher for Independent Juvenile Defender Program and Pacific Juvenile Defender Center as Amici Curiae on behalf of Defendant and Appellant.

Xavier Becerra and Rob Bonta, Attorneys General, Matthew Rodriquez, Acting Attorney General, Lance E. Winters, Chief Assistant Attorney General, Susan Sullivan Pithey, Assistant Attorney General, Michael R. Johnsen, David E. Madeo, Lindsay Boyd and Daniel J. Hilton, Deputy Attorneys General, for Plaintiff and Respondent.

Opinion of the Court by Liu, J.

In 2016, the voters of California enacted Proposition 57, a measure that amended the law governing the punishment of juvenile offenses in adult criminal court by requiring hearings to determine whether the offenses should instead be heard in juvenile court. Adjudicating these offenses in juvenile court typically results in less severe punishment for the juvenile offender. (People v. Superior Court (Lara) (2018) 4 Cal.5th 299, 306–307 (Lara).)

Our precedent holds that “new laws that reduce the punishment for a crime are presumptively to be applied to defendants whose judgments are not yet final.” (People v. Conley (2016) 63 Cal.4th 646, 656 (Conley), citing In re Estrada (1965) 63 Cal.2d 740 (Estrada).) When that presumption applies, its retroactivity rule extends to all “nonfinal judgments.” (People v. Esquivel (2021) 11 Cal.5th 671, 677 (Esquivel).) Applying that rule, we unanimously concluded two years after Proposition 57 passed that the initiative “ameliorated the possible punishment for a class of persons, namely juveniles.” (Lara, supra, 4 Cal.5th at p. 308.) We held that “Estrada’s inference of retroactivity applies” to the proposition’s juvenile provisions, making those provisions applicable to all cases in which the judgment was not final when the proposition went into effect. (Lara, at p. 309.)

The question here is whether Proposition 57 applies during resentencing when a criminal court sentence imposed on a juvenile offender before the initiative’s passage has since been vacated. Defendant Mario Salvador Padilla was originally sentenced before Proposition 57 was enacted, but his judgment later became nonfinal when his sentence was vacated on habeas corpus and the case was returned to the trial court for imposition of a new sentence. Consistent with our decisions articulating the scope of the Estrada presumption, we hold that Proposition 57 applies to his resentencing.

I.

When Padilla was 16 years old, he stabbed his mother to death and conspired with a cousin to kill his stepfather. Following a hearing “at which he was determined not fit to be dealt with under juvenile court law,” Padilla was convicted in adult criminal court and was sentenced to life without the possibility of parole. (People v. Padilla (2020) 50 Cal.App.5th 244, 248 (Padilla); see Welf. & Inst. Code, former § 707 [fitness hearing procedure].) After the United States Supreme Court held in Miller v. Alabama (2012) 567 U.S. 460 (Miller) that mandatory life without parole sentences for juveniles violate the federal Constitution, he petitioned for a writ of habeas corpus seeking resentencing in light of the high court’s holding. (Padilla, at p. 248.) The trial court vacated his sentence, reconsidered it in light of Miller, and again imposed life without the possibility of parole. (Padilla, at p. 248.) While Padilla’s appeal from his new sentence was pending, the United States Supreme Court decided Montgomery v. Louisiana (2016) 577 U.S. 190 (Montgomery), which clarified the analysis that must precede a sentence of life without the possibility of parole for a juvenile defendant. (See id. at pp. 208–210.) The Court of Appeal vacated Padilla’s second sentence in light of Montgomery and again remanded his case to the trial court for resentencing. (Padilla, at p. 248.)

About two weeks after Padilla’s second sentence was vacated, California voters approved Proposition 57. As relevant here, Proposition 57 requires all criminal charges against minors to be filed in juvenile courts. Under the proposition, minors may be tried and sentenced in criminal courts “ ‘only after a juvenile court judge conducts a transfer hearing to consider various factors such as the minor’s maturity, degree of criminal sophistication, prior delinquent history, and whether the minor can be rehabilitated.’ ” (Lara, supra, 4 Cal.5th at p. 305, quoting People v. Vela (2017) 11 Cal.App.5th 68, 72.) As discussed below, this transfer hearing differs in significant ways from the fitness hearing Padilla received.

The trial court again imposed life imprisonment without the possibility of parole (LWOP). Padilla appealed, arguing that he was entitled to a transfer hearing under Proposition 57 because his case became nonfinal once his sentence was vacated. (Padilla, supra, 50 Cal.App.5th at p. 248.) The Court of Appeal agreed and remanded Padilla’s case once more to the trial court with directions to refer the matter to juvenile court for a transfer hearing. (Id. at p. 256.) We granted the Attorney General’s petition for review and now affirm.

II.

Section 3 of the Penal Code instructs that no part of that code applies retroactively, which we have taken to mean that new criminal laws do not govern prosecutions initiated before the law went into effect. (See Estrada, supra, 63 Cal.2d at pp. 746–748.) But we have recognized an exception to this rule for new laws that mitigate punishment; in Estrada, we held that such laws are presumed to apply to cases charged before the law’s enactment but not yet final. (Id. at p. 745.) Absent evidence to the contrary, we presume that when the Legislature “amends a statute so as to lessen the punishment,” it “must have intended that the new statute imposing the new lighter penalty now deemed to be sufficient should apply to every case to which it constitutionally could apply.” (Ibid.) Because the Legislature has “determined that its former penalty was too severe,” the only reason to apply that penalty in pending cases would be “a desire for vengeance,” a motivation we decline to attribute to our lawmakers. (Ibid.) This presumption applies to ameliorative laws enacted by ballot proposition as well. (See Conley, supra, 63 Cal.4th at p. 656.)

We recently held that the Estrada presumption applies to the juvenile provisions of Proposition 57. (Lara, supra, 4 Cal.5th at p. 309; see id. at p. 303 [explaining that although “Estrada is not directly on point[,] . . . its rationale does apply”].) Before the proposition passed, “prosecutors were permitted, and sometimes required, to file charges against a juvenile directly in criminal court, where the juvenile would be treated as an adult.” (Id. at p. 305.) Proposition 57 eliminated that direct filing procedure, reestablishing the historical rule that charges against juveniles must be brought in juvenile court. (Lara, at p. 305.) If the case is retained by the juvenile court after a transfer hearing, and if the court finds that the minor committed the charged offense, the court then conducts a dispositional hearing, where potential custody commitments are less lengthy than in criminal court. (See Welf. & Inst. Code, § 607; see also id., § 730, subd. (a)(2).) Because Proposition 57 reduced “the possible punishment for a class of persons, namely juveniles,” we determined that it made “an ‘ameliorative change[] to the criminal law’ that we infer the legislative body intended ‘to extend as broadly as possible.’ ” (Lara, at pp. 308, 309, quoting Conley, supra, 63 Cal.4th at p. 657.) We accordingly held that “this part of Proposition 57 applies to all juveniles charged directly in adult court whose judgment was not final at the time it was enacted.” (Lara, at p. 304.)

III.

Our cases indicate that the range of judgments affected by Estrada is delimited by constitutional constraints; as we said in Estrada itself, a law lessening punishment is understood to apply “to every case to which it constitutionally could apply.” (Estrada, supra, 63 Cal.2d at p. 745.) We have not had occasion to delineate the parameters of “the Legislature’s power to intervene in judicial decisionmaking.” (Esquivel, supra, 11 Cal.5th at p. 678.) But we have indicated that any restrictions on that power would attach at “the conclusion of a criminal proceeding as a whole” — i.e., when “ ‘the last word of the judicial department with regard to a particular case or controversy’ ” has issued. (Ibid., quoting Plaut v. Spendthrift Farm, Inc. (1995) 514 U.S. 211, 227 (Plaut).)

On this question, we have consulted high court precedent interpreting the principle of separation of powers to provide that when the judicial department has concluded its judgment in a particular case, “Congress may not declare by retroactive legislation that the law applicable to that very case was something other than what the courts said it was.” (Plaut, supra, 514 U.S. at p. 227.) Congress may not direct “findings or results under old law,” but it may “compel[] changes in law.” (Robertson v. Seattle Audubon Soc. (1992) 503 U.S. 429, 438.) Consistent with this view, we have approved laws that alter indisputably final cases when they create new rules or procedures by which a defendant may seek relief. (See Esquivel, supra, 11 Cal.5th at p. 677.)

No similar constitutional concern arises when the Legislature or electorate enacts new laws altering nonfinal judgments. (Esquivel, supra, 11 Cal.5th at pp. 678–679.) As a result, Padilla’s case does not come near whatever limits there may be on the power of lawmakers to impose their commands retroactively. He was sentenced to life imprisonment without the possibility of parole before the United States Supreme Court held in Miller and Montgomery that such a sentence is unconstitutional when imposed on a juvenile unless the court has considered whether the sentence is appropriate in light of the minor’s age and potential for rehabilitation. After petitioning for habeas relief on that basis, his sentence was vacated, a new term was imposed, and then that sentence was vacated too. The decision below followed Padilla’s appeal from his second resentencing. (See Padilla, supra, 50 Cal.App.5th at pp. 253–254.)

The Attorney General concedes that the vacatur of Padilla’s sentence made the judgment in his case nonfinal. We agree. A case is final when “the criminal proceeding as a whole” has ended (Esquivel, supra, 11 Cal.5th at p. 678) and “the courts can no longer provide a remedy to a defendant on direct review” (In re Spencer (1965) 63 Cal.2d 400, 405 (Spencer)). When Padilla’s sentence was vacated, the trial court regained the jurisdiction and duty to consider what punishment was appropriate for him, and Padilla regained the right to appeal whatever new sentence was imposed. His judgment thus became nonfinal, and it remains nonfinal in its present posture because the Court of Appeal ordered a second resentencing, from which the Attorney General now appeals. There is no “constitutional obstacle” to applying the Estrada presumption to his case. (Esquivel, at p. 679.)

The Attorney General nonetheless asks us to distinguish for Estrada purposes between cases that are nonfinal because the defendant is undergoing retrial or resentencing and those in a newly coined procedural stance — cases “not yet final on initial review.” But Estrada made no such distinction. The Estrada presumption stems from our understanding that when the Legislature determines a lesser punishment is appropriate for a particular offense or class of people, it generally does not wish the previous, greater punishment — which it now deems “too severe” — to apply going forward. (Estrada, supra, 63 Cal.2d at p. 745.) We presume the Legislature intends the reduced penalty to be used instead in all cases in which there is no judgment or a nonfinal one, and in which it is constitutionally permissible for the new law to control. (See ibid.; Esquivel, supra, 11 Cal.5th at p. 677.)

The Legislature may write statutes that provide for a different or more limited form of retroactivity, or for no retroactivity at all. This includes the prerogative to disclaim the application of a new ameliorative law to proceedings that occur after a defendant’s conviction or sentence has been vacated. But we have not presumed from statutory silence any retroactive intent less than that described in Estrada — i.e., absent a discernable intent to the contrary, ameliorative criminal laws apply to all nonfinal cases. (Estrada, supra, 63 Cal.2d at p. 745.) Proposition 57 reflects a decision by California’s voters that the range of punishments meted out in criminal court is too severe for most juvenile offenders. In accord with Estrada, our presumption is that the voters wanted that reduction in punishment to stretch “ ‘as broadly as possible, distinguishing only as necessary between sentences that are final and sentences that are not.’ ” (Lara, supra, 4 Cal.5th at p. 308, quoting Conley, supra, 63 Cal.4th at p. 657.) Nothing about this presumption is undermined when a case is nonfinal because the defendant’s sentence has been vacated rather than because the initial review of the sentence has not yet concluded.

Under our precedent and the high court’s, a judgment becomes final “ ‘where the judgment of conviction was rendered, the availability of appeal exhausted, and the time for petition for certiorari ha[s] elapsed.’ ” (Spencer, supra, 63 Cal.2d at p. 405, quoting Linkletter v. Walker (1965) 381 U.S. 618, 622, fn. 5, disapproved on another ground in Teague v. Lane (1989) 489 U.S. 288.) Once that process ends, the judgment may be challenged on collateral review. Merely filing a collateral attack does not make the judgment nonfinal. As the high court has explained, collateral review is distinct from direct review in that it seeks to unwind a judgment that has been affirmed on appeal. (Brecht v. Abrahamson (1993) 507 U.S. 619, 634.) For that reason, “ ‘ “an error that may justify reversal on direct appeal will not necessarily support a collateral attack on a final judgment.” ’ ” (Ibid., quoting United States v. Frady (1982) 456 U.S. 152, 165.) But once a court has determined that a defendant is entitled to resentencing, the result is vacatur of the original sentence, whereupon the trial court may impose any appropriate sentence.

It is clear that Padilla’s present appeal from his resentencing is part of direct review of a nonfinal judgment, not collateral review of a final judgment. The court had the power to impose any sentence available for his crime, including life without the possibility of parole if it found that sentence appropriate in light of Padilla’s “ ‘youth and its attendant characteristics.’ ” (Jones v. Mississippi (2021) ___ U.S. ___ [141 S.Ct. 1307, 1317].) Indeed, while collateral review is an attack on a final judgment, that is plainly not the posture here. When Padilla’s new sentence was imposed, there was no final judgment to attack because his prior sentence had been vacated.

IV.

Our dissenting colleagues have filed a lengthy opinion objecting to today’s holding. The dissent repeatedly asserts that the Estrada presumption applies only to nonfinal judgments. (Dis. opn., post, at pp. 4?8, 16–17.) No one disagrees. The question here is whether Estrada’s applicability to nonfinal judgments means it applies to a resentencing that occurs after a defendant’s original sentence is vacated in a habeas corpus proceeding. The dissent also devotes several pages to showing that our past cases have not addressed whether a judgment like the one before us is nonfinal. (Dis. opn., post, at pp. 20?24.) No one disagrees with that either; we granted review to decide a question that our cases have not had occasion to address.

On the question presented, the dissent declares without citation to authority that “a case has either become final on direct appeal or it has not.” (Dis. opn., post, at p. 5.) Once a judgment has become final on direct appeal, the dissent says, that finality cannot be “ignored because of a later-brought collateral attack.” (Ibid.)

As an initial matter, we note that the dissent’s thesis has not been urged by any party in this case. The Attorney General concedes the judgment before us is nonfinal — his briefing says he “does not challenge the Court of Appeal’s observation that the judgment in this case became nonfinal when appellant was resentenced” — and instead argues that Estrada does not reach all nonfinal judgments. The dissent, by contrast, acknowledges that Estrada reaches all nonfinal judgments and — directly contrary to the Attorney General’s position — argues that Padilla’s judgment remains final. This is not a difference in “nomenclature.” (Dis. opn., post, at p. 12, fn. 8.)

Novelty aside, the dissent’s approach fails to persuade because the notion that a criminal judgment’s finality may be interrupted by a subsequent habeas action is unexceptional. When a habeas court vacates a prior judgment and orders a new trial or new sentencing hearing, the prior judgment — now ineffective — can no longer be a final one. The high court has indicated that when a “new trial proceeding” is conducted after a collateral attack vacates a defendant’s judgment, an appeal from that new proceeding is part of direct rather than collateral review. (McKinney v. Arizona (2020) 589 U.S. __, __, fn. * [140 S.Ct. 702, 709, fn. *] (McKinney).) That is exactly what happened here: Padilla’s sentence was vacated, a new sentencing hearing occurred, and he took the present appeal from that resentencing.

The dissent says McKinney supports its position that Padilla’s initial judgment remains final because “[t]he procedural posture of McKinney and Padilla’s case seem the same.” (Dis. opn., post, at p. 15.) In the dissent’s view, Padilla’s resentencing in light of Miller and Montgomery is no different from a reweighing of aggravating and mitigating circumstances under Clemons v. Mississippi (1990) 494 U.S. 738 (Clemons) when a capital jury has relied on an invalid aggravating circumstance or, as in McKinney, when a capital jury has failed to properly consider relevant mitigating evidence. (Dis. opn., post, at p. 15 [characterizing Padilla’s resentencing as “reweigh[ing] the Miller/Montgomery considerations to determine whether an LWOP sentence was appropriate”].) The dissent suggests that Padilla’s resentencing, like a Clemons reweighing, “ ‘is akin to harmless-error review’ that is ‘routinely conduct[ed] . . . in collateral proceedings.’ ” (Dis. opn., post, at p. 15, quoting McKinney, supra, 589 U.S. at p. __ [140 S.Ct. at p. 709].) This understanding of Miller and Montgomery leads the dissent to assert that no proceeding in this case “constituted a determination that Padilla’s LWOP sentence was illegal” and that “[i]n reality, the [trial] court concluded that the LWOP term was properly imposed.” (Dis. opn., post, at pp. 10, 12.)

As the dissent acknowledges, however, “ ‘Clemons itself . . . stated that an appellate reweighing is not a sentencing proceeding . . . .’ ” (Dis. opn, post, at p. 15, quoting McKinney, supra, 589 U.S. at p. __ [140 S.Ct. at p. 708].) Clemons made clear that “the invalidation of one aggravating circumstance does not necessarily require an appellate court to vacate a death sentence and remand to a jury.” (Clemons, supra, 494 U.S. at p. 745.) Indeed, McKinney’s sentence was never vacated (McKinney, at p. __ [140 S.Ct. at p. 706]), unlike Padilla’s. And Padilla’s sentence was vacated because, contrary to what the dissent says, it had been improperly imposed and was illegal under Miller and Montgomery.

In Miller, the high court held that “the Eighth Amendment forbids a sentencing scheme that mandates life in prison without possibility of parole for juvenile offenders.” (Miller, supra, 567 U.S. at p. 479.) In addition, before issuing an LWOP sentence to a juvenile offender, a sentencing court is “require[d] . . . to take into account how children are different, and how those differences counsel against irrevocably sentencing them to a lifetime in prison.” (Id. at p. 480.) In Montgomery, the high court held that Miller announced a substantive principle of constitutional law — a rule that “place[d] certain . . . punishments altogether beyond the State’s power to impose.” (Montgomery, supra, 577 U.S. at p. 201.) “It follows that when a State enforces a proscription or penalty barred by the Constitution, the resulting conviction or sentence is, by definition, unlawful.” (Ibid.) The court explained that Miller “bar[red] life without parole . . . for all but the rarest of juvenile offenders, those whose crimes reflect permanent incorrigibility.” (Montgomery, at p. 209.) To separate out those juveniles who may constitutionally be sentenced to LWOP from those who may not, “Miller requires a sentencer to consider a juvenile offender’s youth and attendant characteristics before determining that life without parole is a proportionate sentence.” (Montgomery, at pp. 209–210.)

Miller and Montgomery do not contemplate a harmless error-type assessment of a defendant’s youth on collateral review and affirmance of an existing LWOP sentence in the manner envisioned by the dissent. The cases instruct that an LWOP sentence cannot be imposed except in a sentencing hearing in which the defendant’s “ ‘youth and its attendant characteristics’ are considered as sentencing factors,” and they declare that an LWOP sentence imposed on a juvenile without prior consideration of these factors is “not just erroneous but contrary to law and, as a result, void.” (Montgomery, supra, 577 U.S. at pp. 210, 203.) Moreover, Montgomery gave only two options for states to remedy a Miller violation: “permitting juvenile homicide offenders to be considered for parole” or “resentencing them.” (Montgomery, at p. 212.)

As Montgomery requires, and contrary to how the dissent characterizes the procedure for remedying Miller error, courts faced with Miller violations routinely vacate the unlawful sentence and order resentencings that comply with the high court’s instructions. (See, e.g., People v. Watson (2017) 8 Cal.App.5th 496, 503; In re Berg (2016) 247 Cal.App.4th 418, 425; People v. Blackwell (2016) 3 Cal.App.5th 166, 173–174; People v. Lozano (2016) 243 Cal.App.4th 1126, 1129–1130; U.S. v. Delgado (2d Cir. 2020) 971 F.3d 144, 159–160; U.S. v. Friend (4th Cir. 2021) 2 F.4th 369, 374; Jackson v. Vannoy (5th Cir. 2020) 981 F.3d 408, 411–412; U.S. v. Sparks (5th Cir. 2019) 941 F.3d 748, 752–753; Wright v. U.S. (8th Cir. 2018) 902 F.3d 868, 871; U.S. v. Pete (9th Cir. 2016) 819 F.3d 1121, 1126; State v. Montgomery (La. 2016) 194 So.3d 606, 606–607 [on remand from Montgomery].) Indeed, this court in In re Kirchner (2017) 2 Cal.5th 1040 affirmed an order granting a new sentencing hearing when the juvenile’s original LWOP sentence was imposed in violation of Miller. We held that “the possibility that a resentencing that accounts for the Miller factors will occur” under a pre?Miller statute allowing juvenile LWOP sentences to be recalled in certain circumstances “does not represent an adequate substitute for the timely and certain resentencing hearings that Miller and Montgomery require.” (In re Kirchner, at p. 1056, citations omitted.) Against this uniform body of case law, we are not aware of any authority — and the dissent cites none — suggesting that a Miller violation can be remedied by a reweighing process akin to harmless error review.

In sum, because a resentencing to remedy a Miller violation bears no resemblance procedurally to a Clemons reweighing, the dissent’s analogy to McKinney fails. Indeed, the dissent tellingly minimizes a fact that readily distinguishes McKinney: Padilla’s sentence, unlike McKinney’s, was vacated. Indeed, it was vacated twice: first by the trial court when it resentenced him after Miller, then by the Court of Appeal when it ordered a second resentencing after Montgomery. It was vacated because both the original sentence and the sentence that replaced it were invalid — not because LWOP was categorically out of bounds for his offense, but because his sentence had not been lawfully imposed in light of Miller and Montgomery. To suggest there is any ambiguity as to whether Padilla’s sentence was vacated (see dis. opn., post, at p. 10 [“Even assuming the trial court vacated defendant’s LWOP sentence . . . .”]) “simply ignores the facts and the procedural posture of the case” (id. at p. 8).

In addition, the dissent notes that Padilla was originally tried before the enactment of the direct filing regime that preceded Proposition 57 and argues that this is a “crucial distinction” between his case and those to which Proposition 57 retroactively applies. (Dis. opn., post, at p. 20.) Padilla’s case was initially brought in juvenile court and removed to criminal court after a “fitness hearing.” (Welf. & Inst. Code, former § 707.) But, as the Court of Appeal explained, there are significant differences between the fitness hearing envisioned by the prior law and the transfer hearing provided by Proposition 57. “Notably, under prior law, juveniles age 16 or older who were accused of certain offenses, including murder, were subject to a rebuttable presumption that they were unfit for juvenile court treatment. [Citation.] No such presumption applies in transfer hearings under Proposition 57, and the People have the burden to show that the juvenile should be treated as an adult.” (Padilla, supra, 50 Cal.App.5th at p. 249.) Furthermore, the prior law permitted the juvenile court to retain jurisdiction only if it found the minor suitable for juvenile court adjudication under each of five statutory criteria. (Id. at pp. 249–250.) Those criteria are now merely factors for the juvenile court to consider in exercising “broad discretion” as to whether to retain jurisdiction. (Id. at p. 250; see Welf. & Inst. Code, § 707, subd. (a)(3).) In short, Proposition 57 is ameliorative within the meaning of Estrada, whether compared to the direct filing regime or the fitness hearing scheme that preceded it.

Moreover, the law under which Padilla was originally tried does not change how the presumption we recognized in Estrada applies to Proposition 57. Under our precedent, we presume the electorate intended the proposition to apply to all nonfinal cases — that is, “to every case to which it constitutionally could apply.” (Estrada, supra, 63 Cal.2d at p. 745.) We have never suggested that limits on a new law’s application may flow from the legal regime under which a defendant whose judgment is nonfinal was originally tried. And we have applied new, ameliorative laws where the initial disposition took place under a version of the law several iterations back. (See People v. Vieira (2005) 35 Cal.4th 264, 305.) We take the same approach here.

The dissent also complains that our decision “means that a man who is now 40 years of age will be given a new juvenile transfer hearing” under Proposition 57. (Dis. opn., post, at p. 15.) It objects that “the juvenile court would be forced to determine, over 20 years after the fact, whether Padilla should have been treated as a juvenile in 1999.” (Id. at p. 17.) The Attorney General similarly argues that a transfer hearing in Padilla’s case will likely “present challenges given the passage of time” because some of the criteria that juvenile courts assess during those hearings may be difficult to apply to a defendant who is “past the age of juvenile court jurisdiction.”

Whatever concern our dissenting colleagues may have about the result in this case, it must be observed that the dissent’s proposed rule is not limited to defendants beyond a certain age. The dissent does not dispute that under its view Padilla could not receive a transfer hearing even if he had been 17 years old when his original sentence was vacated, so long as direct review of that initial sentence had concluded before Proposition 57 became effective. By calling attention to Padilla’s age, the dissent obscures the fact that its categorical rule would apply equally to individuals within or near the age of juvenile court jurisdiction.

More generally, we do not doubt that “the appropriate remedy can be somewhat complex” when new laws are applied retroactively in the juvenile context because of the consequences for those proceedings of the passage of time. (Lara, supra, 4 Cal.5th at p. 313.) But Lara considered those complexities and determined they do not bar retroactive application of Proposition 57 to nonfinal cases. Because of our decision in Lara, the law already requires some defendants who exceed the age of juvenile court jurisdiction to have their amenability to juvenile adjudication considered retrospectively under the new standards of Proposition 57. (See People v. Ramirez (2019) 35 Cal.App.5th 55, 59–60 [affirming order for transfer hearing for defendant over the age of 25].)

Under Lara, such defendants must receive a transfer hearing; their sentence will be reinstated if the court finds criminal adjudication appropriate, or else their convictions will be “ ‘treat[ed] . . . as juvenile adjudications.’ ” (Lara, supra, 4 Cal.5th at p. 310.) For a defendant over the age of 25, a juvenile court generally will not be able to retain continuing jurisdiction if it finds juvenile adjudication proper. (Welf. & Inst. Code, § 607, subds. (c), (h)(2).) We made clear in Lara that the complexity and possible outcomes of this remedial approach are “no reason to deny the [transfer] hearing.” (Lara, at p. 313.)

We note that some odd results are inevitable with any rule of retroactivity. (Cf. Dorsey v. United States (2012) 567 U.S. 260, 280–281.) As the Attorney General argues, applying ameliorative laws to proceedings like Padilla’s resentencing may yield different outcomes in certain instances for defendants whose cases were initially similar. But the decision we reach today “properly rests on considerations of finality in the judicial process.” (Shea v. Louisiana (1985) 470 U.S. 51, 59–60.) When a defendant’s sentence has been vacated, the parties’ interests in repose and finality are necessarily diminished; at that point, the countervailing interest in effectuating current legislative policy decisions may appropriately control. The dissent’s and the Attorney General’s positions, by contrast, would require sentencing courts in such cases to apply statutes that the Legislature or electorate has changed upon finding them “too severe” — excessively punitive, unwise, or even constitutionally infirm. (Estrada, supra, 63 Cal.2d at p. 745.)

Of course, courts may assess the practical operation of an ameliorative law in determining whether it was intended to apply retroactively to all nonfinal cases, as Estrada presumes. Having undertaken such an assessment in Lara, we concluded that Estrada’s “inference of retroactivity should apply” to Proposition 57. (Lara, supra, 4 Cal.5th at p. 308.) We might have drawn a different conclusion in a case involving a different statutory scheme. But the dissent’s view that “once final” means “final forever” is not specific to Proposition 57 or to juvenile laws; it would apply to any offender, at any age, regardless of the nature of the ameliorative change at issue. The dissent underscores that “Padilla received a juvenile fitness hearing under prior law” and suggests that Proposition 57 may not be sufficiently ameliorative in his case to trigger the Estrada presumption. (Dis. opn., post, at pp. 19–20.) But under the dissent’s approach to finality, Estrada would be inapplicable even if Proposition 57 had capped the punishment for Padilla’s offense at 25 years of imprisonment; despite such an ameliorative change, he could still be resentenced to LWOP. This cannot be squared with the “inevitable inference that the Legislature must have intended that the new statute imposing the new lighter penalty now deemed to be sufficient should apply to every case to which it constitutionally could apply.” (Estrada, supra, 63 Cal.2d at p. 745.)

The dissent further contends that the Proposition 57 ballot materials told voters “the changes enacted by Prop. 57 would be prospectively applied.” (Dis. opn., post, at p. 16.) But we reviewed those ballot materials in Lara and unanimously found they were “inconclusive” and “silent on the question” of retroactivity. (Lara, supra, 4 Cal.5th at p. 309.) In the face of such silence, we followed Estrada’s instruction to “infer the legislative body intended [the ameliorative change] ‘to extend as broadly as possible.’ ” (Lara, at p. 309.) The dissent offers no reason why we should reconsider Lara’s analysis.

Finally, we find unpersuasive two arguments made by the Attorney General. He points to a recent amendment to the firearm enhancement statutes providing that the new discretion courts have to dismiss these enhancements “applies to any resentencing that may occur pursuant to any other law.” (Pen. Code, § 12022.5, subd. (c).) The Attorney General says this supports the view that the Legislature does not generally intend ameliorative laws to apply when a defendant’s sentence has been vacated. But the Legislature was entitled to take a belt?and?suspenders approach to ensuring that the firearm enhancement reform it passed would apply broadly. Relying on legislative silence to infer an intent to limit the retroactive application of ameliorative laws would invert Estrada’s basic principle that we presume from legislative silence an intent to apply new laws as broadly as constitutional boundaries permit. (Esquivel, supra, 11 Cal.5th at p. 677.)

The Attorney General also contends that applying Proposition 57 to defendants whose sentences are vacated would be inconsistent with “principles that generally limit the scope of subsequent modification of a judgment after initial finality.” In support, the Attorney General argues that vacatur of a defendant’s sentence “does not allow a resentencing court to consider new claims or affect any part of the judgment other than the sentence.” But the right and remedy we recognize today does not allow Padilla to raise claims unrelated to his sentence. The relief that applies to him is the same as what we approved in Lara for juveniles whose cases were pending when that measure passed: He must receive a transfer hearing in a juvenile court, where the court will decide whether criminal adjudication is appropriate for the murder of his mother and conspiracy to kill his stepfather. Whatever potential that hearing may have for reducing his punishment (the nonfinal part of his judgment), it does not authorize or constitute relitigation of guilt.

CONCLUSION

Because the judgment in Padilla’s case became nonfinal when his sentence was vacated on habeas corpus, Proposition 57 applies to his resentencing. We affirm the judgment of the Court of Appeal and remand the case for further proceedings consistent with this opinion.

LIU, J.

We Concur:

KRUGER, J.

GROBAN, J.

JENKINS, J.

Dissenting Opinion by Justice Corrigan

In 1998, when he was 16 years old, defendant Mario Salvador Padilla and his cousin devised a plan to kill his mother and stepfather and steal money from them. With his cousin’s assistance, Padilla stabbed his mother 45 times while she sat in the family living room, took money intended for his newborn stepsister, and fled. Padilla’s mother identified her son as her attacker before she died from her wounds. Padilla was arrested the same day. Also in 1998, Padilla was charged “as an adult, following a hearing at which he was determined not fit to be dealt with under juvenile court law.” (People v. Padilla (2020) 50 Cal.App.5th 244, 248.) The following year, he was convicted of the first degree murder of his mother and conspiracy to kill his stepfather.[2] A robbery-murder special circumstance was found true,[3] and he was sentenced to life without the possibility of parole (LWOP). His case became final in 2001 when we denied his petition for review of the Court of Appeal’s judgment, and he did not petition the United States Supreme Court for a writ of certiorari.

Eleven years later, Miller v. Alabama (2012) 567 U.S. 460 held “mandatory life without parole for those under the age of 18 at the time of their crimes violates the Eighth Amendment’s prohibition on ‘cruel and unusual punishments’ ” (id. at p. 465), and concluded that “[a]lthough we do not foreclose a sentencer’s ability to make that judgment in homicide cases, we require it to take into account how children are different, and how those differences counsel against irrevocably sentencing them to a lifetime in prison” (id. at p. 480). In 2014, 13 years after Padilla’s case had become final, he filed a habeas corpus petition seeking relief under Miller. His habeas petition did not challenge the adjudication of his guilt, nor the determination that he should be tried as an adult. The trial court held a hearing in compliance with Miller and concluded an LWOP term was appropriate. Defendant appealed.

While that appeal was pending, the high court returned to the subject in Montgomery v. Louisiana (2016) 577 U.S. 190 and clarified: “Miller announced a substantive rule of constitutional law” that was fully retroactive. (Id. at p. 212.) Montgomery held that Miller “did more than require a sentencer to consider a juvenile offender’s youth before imposing life without parole.” (Id. at p. 208.) “Because Miller determined that sentencing a child to life without parole is excessive for all but ‘ “the rare juvenile offender whose crime reflects irreparable corruption,” ’ [citation], it 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.” (Ibid.) In light of this development, the Court of Appeal remanded for a second hearing to comply with Montgomery. (See People v. Padilla (2016) 4 Cal.App.5th 656, 673–674.) The trial court again concluded that LWOP was an appropriate sentence because Padilla’s crimes did not stem from transient youthful immaturity.

Under this procedural posture, the majority concludes Padilla should receive the retroactive benefit of Proposition 57 (Prop. 57). That initiative contained no expression of intent for retroactive application and was passed 15 years after Padilla’s direct appeal became final. Yet, under the majority’s reasoning, Padilla, now 40 years old, is entitled to a new juvenile transfer hearing under Prop. 57, even though he had already received a fitness hearing under existing law and the trial court, after a habeas collateral attack, had twice concluded that an LWOP sentence was appropriate under the guidance of Miller and Montgomery. The majority’s application of In re Estrada (1965) 63 Cal.2d 740 (Estrada) to these facts announces an expanded and unsound rule. It fails to honor the distinction between a judgment that has become final on appeal and a new remedy sought by collateral attack. The distinction is important. Estrada created an exception to the statutory presumption that a new statute is presumed to apply prospectively absent an express declaration to the contrary.

When our Penal Code was enacted in 1872, it provided that it would take effect on January 1, 1873 (Pen. Code, § 2) and that “[n]o part of it is retroactive, unless expressly so declared.” (Pen. Code, § 3, italics added.) This direct limitation on retroactivity remains a part of the code to this day. “We have previously construed the statute to mean ‘[a] new statute is generally presumed to operate prospectively absent an express declaration of retroactivity or a clear and compelling implication that the Legislature intended otherwise.’ ” (People v. Alford (2007) 42 Cal.4th 749, 753, quoting People v. Hayes (1989) 49 Cal.3d 1260, 1274.) There is a general presumption that if a new law is silent as to retroactively, it was intended to apply prospectively only. Estrada recognized an exception to this general rule when a new law reduces the punishment for a crime. But in doing so it repeatedly limited the exception to cases that had yet to become final on direct appeal.

Estrada’s opening passage presented the issue: “A criminal statute is amended after the prohibited act is committed, but before final judgment, by mitigating the punishment. What statute prevails as to the punishment — the one in effect when the act was committed or the amendatory act? That is the question presented by this petition.” (Estrada, supra, 63 Cal.2d at p. 742, italics added.) In answering that question, Estrada explained: “When the Legislature amends a statute so as to lessen the punishment it has obviously expressly determined that its former penalty was too severe and that a lighter punishment is proper as punishment for the commission of the prohibited act. It is an inevitable inference that the Legislature must have intended that the new statute imposing the new lighter penalty now deemed to be sufficient should apply to every case to which it constitutionally could apply. The amendatory act imposing the lighter punishment can be applied constitutionally to acts committed before its passage provided the judgment convicting the defendant of the act is not final.” (Id. at p. 745, italics added.) This court held that Estrada was entitled to retroactive application of the new rule precisely because his case was not yet final.

The Estrada holding created an exception to the statutory requirement that a retroactive intent be “expressly so declared.” (Pen. Code, § 3.) However, the Estrada court explicitly cabined the exception it created. Estrada emphasized: “The key date is the date of final judgment. If the amendatory statute lessening punishment becomes effective prior to the date the judgment of conviction becomes final then, in our opinion, it, and not the old statute in effect when the prohibited act was committed, applies.” (Estrada, supra, 63 Cal.2d at p. 744, italics added.) In addressing the application of Penal Code section 3, Estrada reasoned that “[i]n the instant case there are . . . other factors that indicate the Legislature must have intended that the amendatory statute should operate in all cases not reduced to final judgment at the time of its passage.” (Estrada, at p. 746, italics added.)

It is important to note that Estrada recognizes a presumption about legislative or electoral intent regarding retroactivity in the face of silence. Estrada could not have been more explicit: When a new law is enacted that reduces punishment, courts will presume that the Legislature or the electorate intended the new provision should apply not only to all future cases, but also to all pending cases before finality of judgment. However, once a case does become final, we can no longer infer from silence that the Legislature or electorate intended the new law should apply. “The key date is the date of final judgment.” (Estrada, supra, 63 Cal.2d at p. 744.) There is simply no suggestion in Estrada that, for purposes of applying its presumption about legislative or electoral intent, there could be multiple relevant dates of finality or that finality may be ignored because of a later-brought collateral attack. Estrada makes no provision for the reopening of a judgment that has become final after direct review. In other words, a case has either become final on direct appeal or it has not. As we have previously recognized, the Legislature or electorate may expressly enact laws that apply even to final judgments.[4] But the courts may not infer from their silence an intent to do so.

Until this case, we have consistently understood Estrada’s rule to apply to a case that had not been reduced to a final judgment. People v. Esquivel (2021) 11 Cal.5th 671, 675 (Esquivel) observed that the Estrada presumption “has been a fixture of our criminal law for more than 50 years.” In People v. McKenzie (2020) 9 Cal.5th 40 (McKenzie), we recently concluded that the defendant’s case was not yet final for Estrada purposes when he was placed on probation with imposition of sentence suspended and an ameliorative law was later enacted during his appeal from a sentence imposed following a probation revocation. We rejected the People’s argument that the defendant’s case became final when he failed to appeal from the initial grant of probation.

We observed that a criminal action “ ‘continues into and throughout the period of probation’ and expires only ‘when th[e] [probation] period ends.’ ” (McKenzie, supra, 9 Cal.5th at p. 47.) McKenzie’s case was not final and had never become so. By virtue of its grant of probation, the sentencing court retained jurisdiction, which included the authority to impose a sentence should defendant violate probation. His exposure to a state prison sentence remained active during the period of his probation. When his probation was revoked and he was sentenced to prison, he took a direct appeal from the judgment ordering that sentence, which was imposed under the jurisdiction the court retained. That direct appeal was pending when the new provision at issue went into effect.

McKenzie repeated the Estrada rule: “[I]n Estrada, we also referred to the cutoff point for application of ameliorative amendments as the date when the ‘case[]’ [citation] or ‘prosecution[]’ is ‘reduced to final judgment’ [citation]. And in [People v.] Rossi [(1976)]18 Cal.3d [295,] 304, we stated that an amendatory statute applies in ‘ “any [criminal] proceeding [that], at the time of the supervening legislation, has not yet reached final disposition in the highest court authorized to review it.” ’ It cannot be said that this criminal prosecution or proceeding concluded before the ameliorative legislation took effect.” (McKenzie, supra, 9 Cal.5th at p. 46.) After McKenzie, we recently repeated that Estrada “continues to stand for the proposition that (i) in the absence of a contrary indication of legislative intent, (ii) legislation that ameliorates punishment (iii) applies to all cases that are not yet final as of the legislation’s effective date.” (Esquivel, supra, 11 Cal.5th at p. 675, italics added.)[5]

Numerous other cases have made similar statements regarding Estrada’s application to cases not yet reduced to final judgment.[6] None of these cases support the proposition that, once a case becomes final for Estrada purposes, the finality of that case may be later revisited as the result of collateral attack. As McKenzie reasoned, “the cutoff point for application of ameliorative amendments” under Estrada is “the date when the ‘case[]’ . . . is ‘reduced to final judgment.’ ” (McKenzie, supra, 9 Cal.5th at p. 46, citation omitted.) Padilla’s case reached that cutoff point in 2001.

In the face of clear precedent, the majority struggles to find a way to say that this long-final judgment has somehow been rendered not final. The majority suggests that “Padilla’s present appeal from his resentencing is part of direct review of a nonfinal judgment, not collateral review” because the trial court “had the power to impose any sentence available for his crime,” and “while collateral review is an attack on a final judgment, that is plainly not the posture here. When Padilla’s new sentence was imposed, there was no final judgment to attack because his prior sentence had been vacated.” (Maj. opn., ante, at pp. 8–9.) That assertion simply ignores the facts and the procedural posture of the case.

The only reason defendant received a new sentencing hearing for consideration of the Miller/Montgomery factors was because defendant collaterally challenged his long-final judgment through a petition for a writ of habeas corpus. His appeal from the sentence imposed after such hearing did not transform a collateral attack on a final judgment into what the majority characterizes as a new “direct review of a nonfinal judgment . . . .” (Maj. opn., ante, at p. 8.) Brecht v. Abrahamson (1993) 507 U.S. 619 (Brecht), cited by the majority, does not suggest otherwise. That case addressed the question of what harmless error standard should apply to a claim on habeas that the prosecution had committed error under Doyle v. Ohio (1976) 426 U.S. 610.[7] The high court concluded the harmless beyond a reasonable doubt standard of Chapman v. California (1967) 386 U.S. 18, applicable on direct review, did not apply. Rather, the court applied “a less onerous harmless-error standard on habeas.” (Brecht, at p. 623.) That standard asks whether the error “ ‘had substantial and injurious effect or influence in determining the jury’s verdict.’ ” (Ibid.) In Brecht, the high court firmly maintained the distinction between direct and collateral review: “The principle that collateral review is different from direct review resounds throughout our habeas jurisprudence. [Citations.] Direct review is the principal avenue for challenging a conviction. ‘When the process of direct review . . . comes to an end, a presumption of finality and legality attaches to the conviction and sentence. The role of federal habeas proceedings, while important in assuring that constitutional rights are observed, is secondary and limited. Federal courts are not forums in which to relitigate state trials.’ [Citation.] [¶] In keeping with this distinction, the writ of habeas corpus has historically been regarded as an extraordinary remedy, ‘a bulwark against convictions that violate “fundamental fairness.” ’ ” (Brecht, at pp. 633?634.)

Under a Brecht analysis, defendant’s full, direct review ended in 2001 when we denied his petition for review, and he did not seek a writ of certiorari before the United States Supreme Court. Contrary to the majority’s suggestion, the purpose of the 2014 habeas proceeding was limited: to ensure that defendant’s sentence complied with Miller and Montgomery. Even assuming the trial court vacated defendant’s LWOP sentence before then reimposing that same term after consideration of the factors outlined by the high court, no portion of Padilla’s sentence or conviction was overturned or rendered invalid. Indeed, the majority acknowledges that an LWOP term was not “categorically out of bounds for his offense” (maj. opn., ante, at p. 14) and makes no suggestion the court improperly reimposed an LWOP term after consideration of the Miller/Montgomery factors. Despite this circumstance, the majority asserts the court’s act of vacating Padilla’s sentence alone rendered his case “nonfinal” for Estrada purposes. (See maj. opn., ante, at pp. 13–14.) The majority’s attempt to focus on how the issue came before the court ignores the substance of the proceedings. In reality, the court concluded that the LWOP term was properly imposed. Padilla received the remedy his writ sought: consideration by the court of the youth factors outlined in Miller and, later, in Montgomery. As the high court has observed, “habeas corpus is, at its core, an equitable remedy.” (Schlup v. Delo (1995) 513 U.S. 298, 319; see Brecht, supra, 507 U.S. at p. 633.) The court’s consideration of the Miller/Montgomery factors on habeas did not serve to reopen direct review regardless of whether or not the court first vacated defendant’s sentence. The juvenile court initially determined that defendant was unfit for juvenile treatment under the existing law. Prop. 57, as a matter of policy, operated to change how such a determination is to be made. There is no indication, however, that the original legal determination which defendant received violated fundamental fairness. (See Brecht, at p. 633.) There has been no suggestion that the “historical rule” permitting trial of a minor in adult court after a judicial determination of unfitness was or is constitutionally infirm.[8] (People v. Superior Court (Lara) (2018) 4 Cal.5th 299, 305 (Lara).)

Brecht helpfully serves to clarify the difference between direct and collateral review. Its further utility is somewhat limited here because, as we have recognized, “the Estrada rule reflects a presumption about legislative intent, rather than a constitutional command . . . .” (Conley, supra, 63 Cal.4th at p. 656.) However, it is worth noting that Estrada did not involve a new constitutional rule. Instead, it focused on discerning the enactors’ intent as to retroactivity in the face of their silence on that matter. For the reasons discussed, Estrada did not contemplate a rule where we may infer from silence a legislative or electoral intent to apply new laws enacted long after a case is final or that a case may be rendered “not final” for these purposes.

The majority’s suggestion that a long-final case can subsequently become “nonfinal” under Estrada essentially treats “finality” like a switch that can be toggled on and off.[9] This conclusion is contrary to Estrada’s reasoning and our decades of subsequent Estrada jurisprudence. Under the majority’s approach, no criminal judgment could ever truly be considered final, because some future collateral habeas attack might arise. Indeed, such collateral attack need not even establish the illegality of a defendant’s sentence before rendering a judgment “nonfinal.” It would only require that “the trial court regained the jurisdiction and duty to consider what punishment was appropriate” and that defendant “regained the right to appeal whatever new sentence was imposed.” (Maj. opn., ante, at p. 6.) Here, neither the trial court’s initial resentencing in light of Miller nor consideration after the Montgomery remand constituted a determination that Padilla’s LWOP sentence was illegal. Instead, those two hearings involved reconsideration of the LWOP sentence in light of the standards set out in Miller and Montgomery.[10]

Under the majority’s reasoning, any collateral attack on a conviction that simply results in a new sentencing proceeding, without any prior determination that the sentence or underlying conviction was illegal would be enough to render a long-final case “nonfinal” for purposes of applying the presumption of Estrada. Further, although the majority only discusses the application of Prop. 57 to Padilla’s case, nothing in its reasoning would limit the application of any and all statutory amendments reducing punishment enacted in the 18 years between the finality of Padilla’s sentence in 2001 and his second Miller/Montgomery hearing in 2019. And, as noted, this outcome would be embraced in the face of legislative and electoral silence regarding retroactivity. This clear expansion of the Estrada doctrine is both unwarranted and unworkable.

The majority maintains “the notion that a criminal judgment’s finality may be interrupted by a subsequent habeas action is unexceptional,” and “[t]he high court has indicated that when a ‘new trial proceeding’ is conducted after a collateral attack vacates a defendant’s judgment, an appeal from that new proceeding is part of direct rather than collateral review,” citing McKinney v. Arizona (2020) 589 U.S. __ [140 S.Ct. 702] (McKinney). (Maj. opn., ante, at p. 10.) In fact, McKinney illustrates just why Padilla’s case does not involve a renewed direct review. A jury convicted McKinney of two murders in 1992. The trial court found aggravating circumstances for both murders and imposed the death penalty. The Arizona Supreme Court affirmed the judgment in 1996. (McKinney, at p. __ [140 S.Ct. at p. 706].) Twenty years later, the Ninth Circuit Court of Appeals granted habeas relief on the ground that “the Arizona courts had failed to properly consider McKinney’s posttraumatic stress disorder (PTSD)” as a relevant mitigating circumstance. (Ibid.) On remand, the Arizona Supreme Court reweighed the aggravating and mitigating circumstances, including the PTSD evidence, under the procedure allowed in Clemons v. Mississippi (1989) 494 U.S. 738, 744–750, and affirmed the death sentence.

As relevant here, McKinney argued his death sentence ran afoul of Ring v. Arizona (2002) 536 U.S. 584 and Hurst v. Florida (2016) 577 U.S. 92 because the trial court, and not the jury, found true the aggravating circumstances before imposing a death judgment. The high court observed “[t]he hurdle is that McKinney’s case became final on direct review in 1996, long before Ring and Hurst. Ring and Hurst do not apply retroactively on collateral review. [Citation.] Because this case comes to us on state collateral review, Ring and Hurst do not apply.” (McKinney, supra, 589 U.S. at p. __ [140 S.Ct. at p. 708].) Similarly to the majority’s reasoning here, McKinney argued that “the Arizona Supreme Court’s 2018 decision reweighing the aggravators and mitigators constituted a reopening of direct review” to which Ring and Hurst should apply. (Ibid.) The high court rejected the claim, reasoning that “the premise of that argument is wrong because the Arizona Supreme Court’s reweighing of the aggravating and mitigating circumstances occurred on collateral review, not direct review.” (Ibid.) Although the defendant protested that “the state label of collateral review cannot control the finality question,” McKinney observed that “Clemons itself . . . stated that an appellate reweighing is not a sentencing proceeding that must be conducted by a jury,” such a reweighing “is akin to harmless-error review” that is “routinely conduct[ed] . . . in collateral proceedings. [Citation.] There is no good reason — and McKinney supplies none — why state courts may not likewise conduct a Clemons reweighing on collateral review.” (Id. at pp. 708–709, fn. omitted.)

The procedural posture of McKinney and Padilla’s case seem the same. In McKinney, the court reviewed aggravating and mitigating circumstances to determine whether the death penalty was appropriate. Here the court reweighed the Miller/Montgomery considerations to determine whether an LWOP sentence was appropriate. Thus, McKinney does not support the conclusion the majority draws from it. As discussed, no case applying Estrada has suggested that a once-final case may be reopened for purposes of applying its exception to the general rule that new laws apply only prospectively.

The majority’s holding means that a man who is now 40 years of age will be given a new juvenile transfer hearing under Prop. 57. It infers from silence the electorate’s intent to permit such a result. Yet that inference would be inconsistent with the legislative analyst’s description of the new transfer hearing procedure in the voter information guide. The description is worded prospectively and nowhere suggests that adults would receive such hearings after the fact: “The measure changes state law to require that, before youths can be transferred to adult court, they must have a hearing in juvenile court to determine whether they should be transferred. As a result, the only way a youth could be tried in adult court is if the juvenile court judge in the hearing decides to transfer the youth to adult court. Youths accused of committing certain severe crimes would no longer automatically be tried in adult court and no youth could be tried in adult court based only on the decision of a prosecutor. In addition, the measure specifies that prosecutors can only seek transfer hearings for youths accused of (1) committing certain significant crimes listed in state law (such as murder, robbery, and certain sex offenses) when they were age 14 or 15 or (2) committing a felony when they were 16 or 17. As a result of these provisions, there would be fewer youths tried in adult court.” (Voter Information Guide, Gen. Elec. (Nov. 8, 2016) analysis of Prop. 57 by Legis. Analyst, p. 56, italics added.) It is dubious at best to argue that a voter who read that description would assume the new procedure would be applied to a case involving an adult like Padilla.[11] Instead, the clear implication, based upon what the voters were told, was that the changes enacted by Prop. 57 would be prospectively applied such that there would be fewer youths tried in adult court.

We should recall that, when interpreting their intent, the enactors are presumed to know the state of the law. (See People v. Shabazz (2006) 38 Cal.4th 55, 65, fn. 8; Anderson v. Superior Court (1995) 11 Cal.4th 1152, 1161.) The clear and settled state of the law was that, even when a retroactive intent is judicially inferred, that inference will not apply to judgments that are final. Again, it is the final judgment rule of Estrada that lies at the heart of this case. Only by devising a way around that longstanding safeguard can the majority’s outcome stand.

Further, it should be noted that the Legislature has provided specific procedures for relief to minors receiving LWOP terms. For example, Padilla is eligible for a youth offender parole hearing during his “25th year of incarceration.” (Pen. Code, § 3051, subd. (b)(4).) Such a hearing “shall provide for a meaningful opportunity to obtain release” (Pen. Code, § 3051, subd. (e)) but would also include considerations of public safety. (See Pen. Code, §§ 3041, subd. (b)(1), 3051, subd. (d).) In addition, a defendant who received an LWOP term and committed his offense when under 18 may, upon serving 15 years, file a petition to recall the sentence wherein he describes his “remorse and work towards rehabilitation. . . .” (Pen. Code, § 1170, subd. (d)(2).)

The majority’s retroactive application of Prop. 57 here would short circuit procedures intended to evaluate whether a defendant in Padilla’s circumstance has successfully been rehabilitated or still presents a danger to public safety if released. Instead, these balanced procedures would be replaced by a new juvenile transfer hearing wherein the juvenile court would be forced to determine, over 20 years after the fact, whether Padilla should have been treated as a juvenile in 1999. If it were to so conclude, the juvenile court could no longer assert jurisdiction over him. His immediate release would be required, regardless of any sign of rehabilitation or consideration of public safety. It seems highly unlikely that voters intended, by silence, to dispense with these carefully crafted procedures for the treatment of youth offenders facing LWOP terms.

On its face, Estrada’s exception to the general rule does not apply here. Padilla’s case was long final, and the change enacted in Prop. 57 did not reduce punishment for a prohibited act. (See Lara, supra, 4 Cal.5th at p. 308.) For Padilla, finality occurred in 2001. Proposition 57 was enacted in 2016, 15 years later. The majority avoids this conclusion, however, by reasoning that Padilla’s 2014 habeas petition seeking relief and the trial court’s decision to maintain his LWOP sentence reopened his case and transformed a final case to one that “became nonfinal.” (Maj. opn., ante, at p. 2.) According to the majority, Padilla’s subsequent appeal from that sentencing, and the Court of Appeal’s later remand for a new hearing under Montgomery, constituted a new “direct review of a nonfinal judgment” to which the Estrada rule applied. (Maj. opn., ante, at p. 8.) Such an analysis ignores the fact that Padilla had already received direct appellate review and that his current petition is a collateral attack on a judgment using the extraordinary equitable remedy of habeas corpus. That approach was rejected by the high court in McKinney.

To defend its extension of Estrada, the majority switches the focus away from Estrada’s finality doctrine to a broader consideration of the concept of amelioration. Under Estrada, amelioration of punishment is a threshold criterion. But it only comes into play to support an unspoken retroactive intent for cases not final on appeal. To bolster its position, the majority points to Lara, supra, 4 Cal.5th 299. The reliance is misplaced. Lara noted that “ ‘[h]istorically, a child could be tried in criminal court only after a judicial determination, before jeopardy attached, that he or she was unfit to be dealt with under juvenile court law.’ ” (Id. at p. 305.) That changed between 1999 and 2000, when new laws permitted, and sometimes required, prosecutors in specified circumstances “to file charges against a juvenile directly in criminal court, where the juvenile would be treated as an adult.” (Ibid.) In 2016, under these provisions, Lara, who committed his offenses at ages 14 and 15, was charged directly in adult court. But Lara’s case did not involve the finality of direct review principle. Before he was tried, the electorate enacted Prop. 57, which “largely returned California to the historical rule” requiring a judicial juvenile transfer hearing and eliminating the direct filing of criminal cases involving minors. (Lara, at p. 305.) In this context, we concluded the rationale of Estrada applied. While Prop. 57 did not reduce punishment (see Lara, at p. 308), the changes it enacted were sufficiently ameliorative to permit application of Estrada’s presumption as to the voters’ unspoken retroactive intent. The conclusion extended the inference about reduction of punishment that Estrada relied upon. “The possibility of being treated as a juvenile in juvenile court — where rehabilitation is the goal — rather than being tried and sentenced as an adult can result in dramatically different and more lenient treatment. Therefore, Proposition 57 reduces the possible punishment for a class of persons, namely juveniles.” (Lara, at p. 303, italics added.) In other words, Lara reasoned that Prop. 57 constituted a reduction in punishment for minors subject to direct filing of charges in adult court because the new law granted them a juvenile transfer hearing and, thus, the possibility of juvenile treatment that they did not have. Lara, by its facts, applied Prop. 57 to juveniles whose cases were still pending in adult court. Lara does not resolve this case.

Here, Padilla’s case is so old that it predated the direct filing scheme that Prop. 57 sought to overturn. Padilla received a juvenile fitness hearing under prior law. He suggests that Lara is dispositive with respect to whether Prop. 57 ameliorated punishment under Estrada. However, Lara did not consider the pre-direct filing scheme at issue in Padilla’s case. Its reasoning depended upon minors being granted the chance for juvenile treatment that had previously been unavailable. That circumstance does not obtain here. Padilla’s case began in juvenile court and was only transferred to adult court after a judge determined that Padilla was not a fit candidate for juvenile treatment. The majority suggests that although Prop.  57 does not reduce punishment, it was nevertheless sufficiently ameliorative of punishment to fall under Estrada’s rationale. It reasons that the transfer hearing prescribed under Prop. 57 is qualitatively different from the fitness hearing Padilla received under prior law because the new law generally made it harder to transfer a case to adult court. (See maj. opn., ante, at pp. 14–15; People v. Padilla, supra, 50 Cal.App.5th at pp. 249–250.) The analysis overlooks a crucial distinction. Lara got his previously denied chance at juvenile treatment because his case had not been adjudicated before Prop. 57’s passage. In Lara there was no final judgment. In fact, there was no judgment at all. Conversely, Padilla got his chance at juvenile treatment. There was no direct adult court filing by the prosecution. The juvenile court determined he was not a fit candidate and ordered his transfer.

Regardless of whether Prop. 57 would constitute an amelioration of Padilla’s punishment, Lara does not support the majority’s conclusion that a case that has become final may become un-finalized for Estrada purposes. There was no question there that Lara’s case was not yet final and the court had no occasion to comment on the finality issue here.

The same was true in Conley, supra, 63 Cal.4th 646, on which the majority relies. (See maj. opn., ante, at pp. 1, 4–5, 8.) In that case, the law changed after Conley, an adult, committed his drunk driving offense. It was also alleged that he had four similar prior convictions, as well as two strike offenses (Pen. Code, §§ 667, subd. (d), 1170.12, subd. (b)) for a residential burglary and stabbing a victim multiple times. Conley was sentenced to a third strike term of 25 years to life based on his new drunk driving conviction. While his appeal from that sentence was pending, voters passed Proposition 36, the Three Strikes Reform Act of 2012 (Reform Act), which reduced “the punishment prescribed for certain third strike defendants.” (Conley, at p. 651.) Conley did not involve Prop. 57, but, more importantly, the case was indisputably not yet final when the Reform Act was passed.[12] Conley did not involve the question of retroactivity at issue here, and it certainly says nothing about the majority’s expansion of Estrada.

Similarly, the majority’s reliance on People v. Vieira (2005) 35 Cal.4th 264 is misplaced. The majority cites that case for the proposition that “we have applied new, ameliorative laws where the initial disposition took place under a version of the law several iterations back.” (Maj. opn., ante, at p. 15, citing Vieira, at p. 305.) That observation is true, but it overlooks the key fact that Vieira’s case was not yet final. Vieira, a capital defendant, claimed on appeal that he should receive the benefit of a 1992 statutory amendment requiring consideration of an ability to pay before imposing a restitution fine. Vieira observed that “[d]efendant is not entitled to benefit from the 1992 amendment; it was repealed in 1994.” (Vieira, at p. 305.) However, Vieira reasoned that the defendant should receive the benefit of the then-current version of the statute, “which provide[d] detailed guidance to the trial court in setting a restitution fine, including consideration of a defendant’s ability to pay. ‘The key date is the date of final judgment. If the amendatory statute lessening punishment becomes effective prior to the date the judgment of conviction becomes final then, in our opinion, it, and not the old statute in effect when the prohibited act was committed, applies.’ ” (Ibid., quoting Estrada, supra, 63 Cal.2d at p. 744.) Vieira concluded that the case was not yet final because it was still pending on direct appeal. (Vieira, at p. 306.)

Although it is true that Vieira applied a new law ameliorating punishment to a defendant subject to a version of the law that had been amended several times, it still involved a case not yet final on direct appeal. Vieira quoted Estrada’s observation that the key date is that of final judgment. At no time did Vieira suggest that finality may be reopened once that date has passed.

The Court of Appeal below suggested that “a collateral proceeding may reopen the finality of a sentence for retroactivity purposes, even while the conviction remains final” (People v. Padilla, supra, 50 Cal.App.5th at p. 253), ascribing this rule to People v. Jackson (1967) 67 Cal.2d 96. Its reading of Jackson is overbroad and does not assist Padilla. Jackson was convicted of special circumstances murder and sentenced to death. He successfully filed a habeas petition, which ultimately led to the reversal of the death sentence but not to the judgment of guilt. (See In re Jackson (1964) 61 Cal.2d 500, 501–508.) After a penalty phase retrial, Jackson again received the death penalty. On appeal from that second death sentence, Jackson argued that a new constitutional rule announced[13] after his initial conviction became final should retroactively apply to him and result in reversal of both the guilt and penalty judgments. Jackson rejected the defendant’s argument as to the guilt phase judgment: “The scope of this retrial is a matter of state procedure under which the original judgment on the issue of guilt remains final during the retrial of the penalty issue and during all appellate proceedings reviewing the trial court’s decision on that issue.” (People v. Jackson, at p. 99, italics added.) However, Jackson observed as to the penalty phase retrial that “[a]lthough defendant’s conviction was final before June 22, 1964, when Escobedo was decided, his retrial on the issue of penalty occurred after that date.” (Id. at p. 100.) Thus, Jackson applied Escobedo only to the penalty phase retrial but not to the judgment of guilt. (Ibid.)

Contrary to the Court of Appeal’s suggestion, Jackson was not an example of a case where finality of judgment was “reopened.” Rather, that case involved a reversal of the penalty judgment, resulting in a later retrial of that phase. To the extent the defendant argued for retroactive application of Escobedo, Jackson rejected that argument because his judgment of guilt was already final at the time Escobedo was decided. Jackson applied Escobedo to the penalty phase retrial prospectively because Escobedo predated that new penalty trial. Thus, Jackson’s sentence was not merely vacated, but the penalty judgment was reversed, rendering it invalid. Thereafter the penalty phase was retried. No analogous proceedings occurred in Padilla’s case. Further, it should be noted that Jackson did not involve an interpretation of Estrada and provides little guidance on the limits of Estrada’s presumption regarding legislative or electoral intent.

Here we find ourselves in new territory. Estrada and Lara do not squarely dispose of the case. The question for us here is whether we can say that the facts are sufficient for us to discern that the voters intended to have Prop. 57 apply retroactively, not to cases not yet final, but to grant relief to a 40-year-old whose case is long final. The ballot materials run counter to such a conclusion. They speak repeatedly in the future tense and repeatedly refer to juveniles, a status Padilla left long ago.

At bottom the “not really final” analysis begs the question: What kind of review can the collateral habeas corpus attack be said to reopen? It did not reopen the verdict of guilt, the finding of special circumstances, nor, critically, the finding of unfitness/transfer. All the two habeas corpus hearings considered was whether an LWOP should have been mandatory for Padilla following his adult conviction. It is important to remember that, even before Miller/Montgomery, Penal Code section 190.5, subdivision (b) gave the trial court discretion to impose a 25 years to life sentence to a minor, rather than LWOP.

The two Miller/Montgomery hearings took place, the trial court applied their standards, now as a matter of constitutional mandate, and still determined, exercising its discretion, that LWOP was appropriate. Even if the majority’s notion of a renewed direct review is well founded, which it is not, the court never reconsidered the juvenile fitness question, which was not an issue raised by the collateral attack. The majority does not acknowledge that procedural posture. Even if Padilla had won at the Miller/Montgomery hearing, the remedy to which he was entitled was a sentence of 25 years to life, instead of LWOP. The habeas corpus proceedings never encompassed whether he was entitled, as a 40 year old, to go back and be treated as a juvenile, which was jurisdictionally impossible. The LWOP sentencing question and the juvenile treatment question are, and always were, distinct. The majority blurs their distinction to create a bridge to their proposed rule.

In sum, Estrada stated an exception to the general rule that a new law which is silent as to retroactivity was intended to apply prospectively only. Estrada reasoned that, despite silence on the matter, a court may presume the enactor’s intent for retroactive application under the limited circumstances that a new law reduces punishment and a final judgment has not been rendered. The majority now expands this presumption to cases that have already become final because, following a collateral attack by way of habeas corpus, the court engages in proceedings that touch upon a defendant’s potential sentence. In such a posture, the majority holds the original case has been reopened, even if those habeas proceedings ultimately do not invalidate any aspect of the prior sentence or conviction. The majority’s expansion of Estrada has no support in the language or reasoning of that case or its progeny. The majority’s reasoning also improperly ascribes to the voters who enacted Prop. 57 an intent, through silence, to apply its provisions to long-final cases, resulting in juvenile transfer hearings for adults who are well past the age at which they can be treated under juvenile law. The majority’s holding significantly undermines the finality rule which all prior cases relied upon as a safeguard and which “has been a fixture of our criminal law for more than 50 years.” (Esquivel, supra, 11 Cal.5th at p. 675.) We should not, on the basis of unsound analysis, drag this Trojan Horse within Estrada’s carefully crafted walls. Accordingly, I respectfully dissent.

CORRIGAN, J.

We Concur:

CANTIL-SAKAUYE, C. J.

PERREN, J.*

_______________________

* Associate Justice of the Court of Appeal, Second Appellate District, Division Six, assigned by the Chief Justice pursuant to article VI, section 6 of the California Constitution.

Footnotes

  1. * Associate Justice of the Court of Appeal, Second Appellate District, Division Six, assigned by the Chief Justice pursuant to article VI, section 6 of the California Constitution. ?
  2. Penal Code sections 187, subdivision (a), 189, subdivision (a); 182, subdivision (a)(1). ?
  3. Penal Code section 190.2, subdivision (a)(17)(A). ?
  4. See, e.g., People v. Gentile (2020) 10 Cal.5th 830, 851–859 (Gentile) (Pen. Code, § 1170.95 [petition procedure for resentencing of homicide conviction based on the natural and probable consequences doctrine]); People v. DeHoyos (2018) 4 Cal.5th 594, 600–606 (Pen. Code, § 1170.18, added by Prop. 47, § 14, as approved by voters (Gen. Elec. (Nov. 4, 2014))); People v. Conley (2016) 63 Cal.4th 646, 656–662 (Conley) (Pen. Code, § 1170.126, added by Prop. 36, § 6, as approved by voters (Gen. Elec. (Nov. 6, 2012))). ?
  5. Esquivel applied an identical analysis as to finality for a defendant placed on probation with execution of a specific state prison sentence suspended. (See Esquivel, supra, 11 Cal.5th at pp. 677–680.) ?
  6. (See, e.g., Gentile, supra, 10 Cal.5th at p. 852; People v. Stamps (2020) 9 Cal.5th 685, 699; People v. Frahs (2020) 9 Cal.5th 618, 624–625; People v. Valenzuela (2019) 7 Cal.5th 415, 424; People v. Lara (2019) 6 Cal.5th 1128, 1134; People v. Buycks (2018) 5 Cal.5th 857, 888; People v. Floyd (2003) 31 Cal.4th 179, 184–185; People v. Rossi, supra, 18 Cal.3d at p. 304; People v. Francis (1969) 71 Cal.2d 66, 75–76.) ?
  7. Doyle held “that the use for impeachment purposes of petitioners’ silence, at the time of arrest and after receiving Miranda warnings, violated the Due Process Clause of the Fourteenth Amendment.” (Doyle v. Ohio, supra, 426 U.S. at p. 619; see Miranda v. Arizona (1966) 384 U.S. 436.) ?
  8. Indeed, this court in Manduley v. Superior Court (2002) 27 Cal.4th 537 rejected constitutional challenges even to an approach, now modified by Prop. 57, that allowed some juveniles to be tried directly in adult court without judicial review. (Manduley, at pp. 551–573; see discussion post, at pp. 18–19.) ?
  9. The majority suggests the People have conceded Padilla’s sentence had been rendered “nonfinal” because the court had vacated his sentence and that “the dissent’s thesis has not been urged by any party in this case.” (Maj. opn., ante, at p. 9.) That assertion mischaracterizes the People’s position. It is true the People have apparently adopted the Court of Appeal’s analysis that defendant’s collateral attack had reopened the finality of Padilla’s sentence. (See People v. Padilla, supra, 50 Cal.App.5th at pp. 253–254.) However, the Attorney General very much has not conceded that the Estrada rule should apply to the present case and, instead, has consistently argued that rule should apply only before a case has become final on an initial direct appeal. We ultimately share the People’s view of Estrada’s scope notwithstanding their use of different nomenclature. ?
  10. It is also important to recall here that neither Miller nor Montgomery forbade an LWOP sentence for a defendant like Padilla. They only require that the court consider the defendant’s youth before imposing such a sentence. ?
  11. As noted, Padilla was not “automatically” tried in adult court. His case was presented to a juvenile court judge who, applying the existing law, determined he should be tried as an adult. This decision was made under the historical rule which, with some modification, Prop. 57 was enacted to restore. (See Lara, supra, 4 Cal.5th at p. 305.) ?
  12. In fact, Conley held the Estrada presumption did not apply because “the Reform Act is not silent on the question of retroactivity.” (Conley, supra, 63 Cal.4th at pp. 657, 658.) ?
  13. The case in question, Escobedo v. Illinois (1964) 378 U.S. 478, predated Miranda v. Arizona, supra, 384 U.S. 436 and involved the admissibility of a defendant’s custodial statement made during a police interrogation. ?