Respondent was convicted in a Texas state court of aggravated sexual assault and sentenced to life imprisonment and a $10,000 fine. After his conviction and sentence were affirmed on direct appeal, he applied for a writ of habeas corpus in state court, arguing that Texas law did not authorize both a fine and prison term for his offense, and thus that his judgment and sentence were void and he was entitled to a new trial. The court, bound by a State Court of Criminal Appeals' decision, recommended that the writ be granted. Before the writ was considered by the Court of Criminal Appeals, however, a new statute was passed allowing an appellate court to reform an improper verdict assessing a punishment not authorized by law. Thus, the Court of Criminal Appeals reformed the verdict by ordering that the fine be deleted and denied the request for a new trial. Arguing that the new Texas law's retroactive application violated the Ex Post Facto Clause of Art. 1, 10, of the Federal Constitution, respondent filed a writ of habeas corpus in Federal District Court, which was denied. The Court of Appeals reversed. Relying on the statement in Thompson v. Utah, 170 U.S. 343 , that retroactive procedural statutes violate the Ex Post Facto Clause unless they "leave untouched all the substantial protections with which existing law surrounds the . . . accused," the court held that respondent's right to a new trial under former Texas law was a "substantial protection."
REHNQUIST, C.J., delivered the opinion of the Court, in which WHITE, BLACKMUN, O'CONNOR, SCALIA, and KENNEDY, JJ., joined. STEVENS, J., filed an opinion concurring in the judgment, in which BRENNAN and MARSHALL, JJ., joined, post, p. 52.
Charles A. Palmer, Assistant Attorney General of Texas, argued the cause for petitioner. With him on the brief were Jim Mattox, Attorney General, Mary F. Keller, First Assistant Attorney General, and Michael P. Hodge, Assistant Attorney General. [497 U.S. 37, 39]
Jon R. Farrar argued the cause and filed a brief for respondent. *
[ Footnote * ] Solicitor General Starr, Assistant Attorney General Dennis, Deputy Solicitor General Bryson, and James A. Feldman filed a brief for the United States as amicus curiae urging reversal.
Arthur F. Mathews and Thomas F. Connell filed a brief for Wilbert Lee Evans as amicus curiae.
CHIEF JUSTICE REHNQUIST delivered the opinion of the Court.
The question presented in this case is whether the application of a Texas statute, which was passed after respondent's crime and which allowed the reformation of an improper jury verdict in respondent's case, violates the Ex Post Facto Clause of Art. I, 10. We hold that it does not.
Respondent Carroll Youngblood was convicted in a Texas court of aggravated sexual abuse. The jury imposed punishment of life imprisonment and a fine of $10,000. After his conviction and sentence were affirmed by the Texas Court of Criminal Appeals, Youngblood applied for a writ of habeas corpus in the State District Court. He argued that Texas Code of Criminal Procedure did not authorize a fine in addition to a term of imprisonment for his offense, and, thus, under the decision of the Court of Criminal Appeals in Bogany v. State, 661 S.W.2d 957 (Tex.Crim.App. 1983), the judgment and sentence were void, and he was entitled to a new trial. 1 In April, 1985, the District Court, feeling bound by Bogany, recommended that the writ be granted.
Before the habeas application was considered by the Texas Court of Criminal Appeals, which has the exclusive power under Texas law to grant writs of habeas corpus, see Tex. Code Crim.Proc.Ann., Art. 11.07 (Vernon 1977 and Supp. 1990), a new Texas statute designed to modify the Bogany [497 U.S. 37, 40] decision became effective. Article 37.10(b), as of June 11, 1985, allows an appellate court to reform an improper verdict that assesses a punishment not authorized by law. Tex.Code Crim.Proc.Ann., Art. 37-10(b) (Vernon Supp. 1990); see Ex parte Johnson, 697 S.W.2d 605 (Tex.Crim.App. 1985). Relying on that statute, the Court of Criminal Appeals reformed the verdict in Youngblood's case by ordering deletion of the $10,000 fine, and denied his request for a new trial.
Youngblood then sought a writ of habeas corpus from the United States District Court for the Eastern District of Texas, arguing that the retroactive application of Article 37.10(b) violated the Ex Post Facto Clause of Art. I, 10, of the Federal Constitution. The District Court concluded that, since Youngblood's "punishment . . . was not increased (but actually decreased), and the elements of the offense or the ultimate facts necessary to establish guilt were not changed," there was no ex post facto violation. App. to Pet. for Cert. C-6.
The Court of Appeals reversed. Youngblood v. Lynaugh, 882 F.2d 956 (CA5 1989). It relied on the statement in this Court's decision in Thompson v. Utah, 170 U.S. 343 (1898), the retroactive procedural statutes violate the Ex Post Facto Clause unless they "`leave untouched all the substantial protections with which existing law surrounds the person accused of crime,'" Lynaugh, supra, at 959 (quoting 170 U.S., at 352 ). It held that Youngblood's right to a new trial under the Bogany decision was such a "substantial protection," and therefore ordered that a writ of habeas corpus be issued. We granted certiorari. 493 U.S. 1001 (1989).
Because respondent is before us on collateral review, we are faced with a threshold question whether the relief sought by Youngblood would constitute a "new rule," which would not apply retroactively under our decisions in Teague v. Lane, 489 U.S. 288 (1989), and Butler v. McKellar, 494 U.S. 407 (1990). Generally speaking, "[r]etroactivity is [497 U.S. 37, 41] properly treated as a threshold question, for, once a new rule is applied to the defendant in the case announcing the rule, even-handed justice requires that it be applied retroactively to all who are similarly situated." Teague, supra, at 300. The State of Texas, however, did not address retroactivity in its petition for certiorari or its briefs on the merits, and, when asked about the issue at oral argument, counsel answered that the State had chosen not to rely on Teague. Tr. of Oral Arg. 4-5. Although the Teague rule is grounded in important considerations of federal-state relations, we think it is not "jurisdictional" in the sense that this Court, despite a limited grant of certiorari, must raise and decide the issue sua sponte. Cf. Patsy v. Board of Regents of Fla., 457 U.S. 496, 515 , n. 19 (1982) (Eleventh Amendment defense need not be raised and decided by the Court on its own motion). We granted certiorari to consider the merits of respondent's ex post facto claim, and we proceed to do so.
Although the Latin phrase "ex post facto" literally encompasses any law passed "after the fact," it has long been recognized by this Court that the constitutional prohibition on ex post facto laws applied only to penal statutes which disadvantage the offender affected by them. Calder v. Bull, 3 Dall. 386, 390-392 (1798) (opinion of Chase, J.); id., at 396 (opinion of Paterson, J.); id., at 400 (opinion of Iredell, J.). See Miller v. Florida, 482 U.S. 423, 430 (1987). 2 As early opinions in this Court explained, "ex post facto law" was a term of art with an established meaning at the time of the framing of the Constitution. Calder, 3 Dall. at 391 (opinion of Chase, J.); id., at 396 (opinion of Paterson, J.). Justice Chase's now familiar opinion in Calder expounded those legislative [497 U.S. 37, 42] acts which in his view implicated the core concern of the Ex Post Facto Clauses:
The Beazell formulation is faithful to our best knowledge of the original understanding of the Ex Post Facto Clause: Legislatures may not retroactively alter the definition of crimes or increase the punishment for criminal acts. Several early state constitutions employed this definition of the term, and they appear to have been a basis for the Framers' understanding of the provision. See The Federalist No. 44, p. 301 (J. Cooke ed. 1961) (J. Madison); 2 M. Farrand, Records of the Federal Convention of 1787, p. 376 (1911); (Calder, 3 Dall. at 391-392 (opinion of Chase, J.); id., at 396-397 (opinion of Paterson, J.). The Constitutions of Maryland and North Carolina, for example, declared that "retrospective laws, punishing facts committed before the existence of such laws, and by them only declared criminal, are oppressive, unjust, and incompatible with liberty; wherefore no ex post facto law ought to be made." See Constitution of Maryland, Declaration of Rights, Art. XV (1776); Constitution of North Carolina, Declaration of Rights, Art. XXIV (1776). Other state constitutions, though not using the phrase "ex post facto," included similar articles. See Declaration of Rights and Fundamental Rules of the Delaware State, 11 (1776); Constitution or Form of Government for the Commonwealth of Massachusetts, Declaration of Rights, Art. XXIV (1780). [497 U.S. 37, 44]
Another historical reference, Blackstone's Commentaries, which was discussed by the Framers during debates on the Ex Post Facto Clause, see 2 M. Farrand, Records of the Federal Convention of 1787, p. 448-449 (1911), and deemed an authoritative source of the technical meaning of the term in Calder, see 3 Dall. at 391 (opinion of Chase, J.); id., at 396 (opinion of Paterson, J.), buttresses this understanding. According to Blackstone, a law is ex post facto "when after an action (indifferent in itself) is committed, the legislator then for the first time declares it to have been a crime, and inflicts a punishment upon the person who has committed it." 1 W. Blackstone, Commentaries *46. Although increased punishments are not mentioned explicitly in the historical sources, the Court has never questioned their prohibition, apparently on the theory that "[t]he enhancement of a crime, or penalty, seems to come within the same mischief as the creation of a crime or penalty." Calder, supra, at 397 (opinion of Paterson, J.). The Beazell definition, then, is faithful to the use of the term "ex post facto law" at the time the Constitution was adopted.
Respondent concedes that Tex. Code Crim. Proc. Ann., Art. 37.10(b) (Vernon Supp. 1990), does not fall within any of the Beazell categories and, under that definition, would not constitute an ex post facto law as applied to him. The new statute is a procedural change that allows reformation of improper verdicts. It does not alter the definition of the crime of aggravated sexual abuse, of which Youngblood was convicted, nor does it increase the punishment for which he is eligible as a result of that conviction. Nevertheless, respondent maintains that this Court's decisions have not limited the scope of the Ex Post Facto Clause to the finite Beazell categories, but have stated more broadly that retroactive legislation contravenes Art. I, 10 if it deprives an accused of a "substantial protection" under law existing at the time of the crime. He argues that the new trial guaranteed him by former Texas law is such a protection. [497 U.S. 37, 45]
Several of our cases have described as "procedural" those changes which, even though they work to the disadvantage of the accused, do not violate the Ex Post Facto Clause. Dobbert v. Florida, supra, at 292-293, and n. 6; Beazell v. Ohio, 269 U.S., at 171 ; Mallett v. North Carolina, 181 U.S. 589, 597 (1901). While these cases do not explicitly define what they mean by the word "procedural," it is logical to think that the term refers to changes in the procedures by which a criminal case is adjudicated, as opposed to changes in the substantive law of crimes. Respondent correctly notes, however, that we have said a procedural change may constitute an ex post facto violation if it "affect[s] matters of substance," Beazell, supra, at 171, by depriving a defendant of "substantial protections with which the existing law surrounds the person accused of crime," Duncan v. Missouri, 152 U.S. 377, 382 -383 (1894), or arbitrarily infringing upon "substantial personal rights." Malloy v. South Carolina, 237 U.S. 180, 183 (1915); Beazell, supra, at 171.
We think this language from the cases cited has imported confusion into the interpretation of the Ex Post Facto Clause. The origin of the rather amorphous phrase "substantial protections" appears to lie in a nineteenth century treatise on constitutional law by Professor Thomas Cooley. T. Cooley, Constitutional Limitations *272. According to Cooley, who notably assumed the Calder construction of the Ex Post Facto Clause to be correct, Constitutional Limitations *265, a legislature "may prescribe altogether different modes of procedure in its discretion, though it cannot lawfully, we think, in so doing, dispense with any of those substantial protections with which the existing law surrounds the person accused of crime." Id., at *272.
This Court's decision in Duncan v. Missouri, supra, subsequently adopted that phraseology:
We think the best way to make sense out of this discussion in the cases is to say that, by simply labeling a law "procedural," a legislature does not thereby immunize it from scrutiny under the Ex Post Facto Clause. See Gibson v. Mississippi, 162 U.S. 565, 590 (1896). Subtle ex post facto violations are no more permissible than overt ones. In Beazell, supra, we said that the constitutional prohibition is addressed to laws, "whatever their form," which make innocent acts criminal, alter the nature of the offense, or increase the punishment. Id., at 170. But the prohibition which may not be evaded is the one defined by the Calder categories. See Duncan, supra, at 382; Malloy, supra, at 183-184. The references in Duncan and Malloy to "substantial protections" and "personal rights" should not be read to adopt without explanation an undefined enlargement of the Ex Post Facto Clause. [497 U.S. 37, 47]
Two decisions of this Court, relied upon by respondent, do not fit into this analytical framework. In Kring v. Missouri, 107 U.S. 221 (1883), the Court said "it is not to be supposed that the opinion in [Calder v. Bull] undertook to define, by way of exclusion, all the cases to which the constitutional provision would be applicable. Id., at 228. It defined an ex post facto law, inter alia, as one which, "`in relation to the offence or its consequences, alters the situation of a party to his disadvantage.'" Id., at 228-229 (quoting United States v. Hall, 26 F.Cas. 84, 86 (No. 15,285) (D Pa. 1809)) (emphasis deleted). And in Thompson v. Utah, 170 U.S. 343 (1898), the Court held that a change in Utah law reducing the size of juries in criminal cases from 12 persons to 8 deprived Thompson of "a substantial right involved in his liberty," and violated the Ex Post Facto Clause. Id., at 352.
Neither of these decisions, in our view, is consistent with the understanding of the term "ex post facto law" at the time the Constitution was adopted. Nor has their reasoning been followed by this Court since Thompson was decided in 1898. These cases have caused confusion in state and lower federal courts about the scope of the Ex Post Facto Clause, as exemplified by the opinions of the District Court and Court of Appeals in this case. See also Murphy v. Kentucky, 465 U.S. 1072, 1073 (1984) (WHITE, J., dissenting from denial of certiorari) (noting "the evident confusion among lower courts concerning the application of the Ex Post Facto Clause to changes in rules of evidence and procedure"); United States v. Kowal, 596 F.Supp. 375, 377 (Conn. 1984) (Supreme Court jurisprudence applying ex post facto prohibition to retroactive procedural changes "is not all of one piece"); L. Tribe, American Constitutional Law 638 (2d ed. 1988) (procedural changes upheld by the Court "can hardly be distinguished in any functional way from those invalidated").
The earlier decision, Kring v. Missouri, was a capital case with a lengthy procedural history. Kring was charged with first-degree murder, but, pursuant to a plea agreement, he [497 U.S. 37, 48] pleaded guilty to second-degree murder. The plea was accepted by the prosecutor and the trial court, and he was sentenced to 25 years in prison. He appealed the judgment, however, on the ground that his plea agreement provided for a sentence of no more than 10 years. The State Supreme Court reversed the judgment and remanded for further proceedings. In the trial court, Kring refused to withdraw his guilty plea to second-degree murder and refused to renew his plea of not guilty to first-degree murder, insisting instead that the acceptance of his earlier plea constituted an acquittal on the greater charge. The trial court, over Kring's objection, directed a general plea of not guilty to be entered, and, upon retrial, he was convicted of first-degree murder and sentenced to death.
At the time the crime was committed, Missouri law provided that a defendant's plea of guilty to second-degree murder, if accepted by the prosecutor and the court, served as an acquittal of the charge of first-degree murder. After the crime, but before Kring made his plea, a new Missouri Constitution abrogated that rule. The State was thus free, as a matter of Missouri law, to retry Kring for first-degree murder after his conviction and the 25-year sentence for second-degree murder were vacated. The Supreme Court of Missouri held that the new law did not violate the Ex Post Facto Clause, because it effected only a change in criminal procedure.
This Court reversed by a vote of 5 to 4. As support for the view that Calder did not define an exclusive list of legislative acts falling within the constitutional prohibition, Justice Miller's opinion for the Court quoted a jury charge given by Justice Washington sitting in the District Court: "`[A]n ex post facto law is one which, in its operation, makes that criminal which was not so at the time the action was performed; or which increases the punishment, or, in short, which, in relation to the offence or its consequences, alters the situation of a party to his disadvantage.'" Kring, supra, at 228-229 [497 U.S. 37, 49] (quoting United States v. Hall, supra, at 86) (emphasis in original). Applying that test, the Court concluded that, because the new Missouri Constitution denied Kring the benefit of an implied acquittal which the previous law provided, it "altered the situation to his disadvantage," and his conviction for first-degree murder was void. Kring, supra, at 235-236.
The Court's departure from Calder's explanation of the original understanding of the Ex Post Facto Clause was, we think, unjustified. The language in the Hall case, heavily relied upon in Kring and repeated in other decisions thereafter, does not support a more expansive definition of ex post facto laws.
In Hall, a vessel owner was sued by the United States for forfeiture of an embargo bond obliging him to deliver certain cargo to Portland. As a legal excuse, the defendant argued that a severe storm had disabled his vessel and forced him to land in Puerto Rico, where he was forced by the Puerto Rican government to sell the cargo. In dicta, Justice Washington hypothesized that, according to the law in effect at the time Hall forfeited the cargo, an "unavoidable accident" was an affirmative defense to a charge of failing to deliver cargo. His jury instruction then explained that a subsequent law imposing an additional requirement for the affirmative defense - that the vessel or cargo actually be lost at sea as a result of the unavoidable accident - would deprive Hall of a defense of his actions available at the time he sold the cargo, and thus be an invalid ex post facto law.
This analysis is consistent with the Beazell framework. A law that abolishes an affirmative defense of justification or excuse contravenes Art. I, 10, because it expands the scope of a criminal prohibition after the act is done. It appears, therefore, that Justice Washington's reference to laws "relat[ing] to the offence, or its consequences," was simply shorthand for legal changes altering the definition of an offense or increasing a punishment. His jury charge should not be read to mean that the Constitution prohibits retrospective [497 U.S. 37, 50] laws, other than those encompassed by the Calder categories, which "alter the situation of a party to his disadvantage." Nothing in the Hall case supports the broad construction of the ex post facto provision given by the Court in Kring.
It is possible to reconcile Kring with the numerous cases which have held that "procedural" changes do not result in ex post facto violations by saying that the change in Missouri law did not take away a "defense" available to the defendant under the old procedure. But this use of the word "defense" carries a meaning quite different from that which appears in the quoted language from Beazell, where the term was linked to the prohibition on alterations in "the legal definition of the offense" or "the nature or amount of the punishment imposed for its commission." Beazell, 269 U.S., at 169 -170. The "defense" available to Kring under earlier Missouri law was not one related to the definition of the crime, but was based on the law regulating the effect of guilty pleas. Missouri had not changed any of the elements of the crime of murder, or the matters which might be pleaded as an excuse or justification for the conduct underlying such a charge; it had changed its law respecting the effect of a guilty plea to a lesser included offense. The holding in Kring can only be justified if the Ex Post Facto Clause is thought to include not merely the Calder categories but any change which "alters the situation of a party to his disadvantage." We think such a reading of the Clause departs from the meaning of the Clause as it was understood at the time of the adoption of the Constitution, and is not supported by later cases. We accordingly overrule Kring.
The second case, Thompson v. Utah, must be viewed in historical context. Thompson was initially charged with his crime - grand larceny committed by stealing a calf - in 1895, when Utah was a Territory. He was tried by a jury of 12 persons and convicted. A new trial was subsequently granted, however, and in the meantime Utah was admitted [497 U.S. 37, 51] into the Union as a State. The Constitution of the State of Utah provided that juries in noncapital cases would consist of eight persons, not twelve, and Thompson was retried and convicted by a panel of eight.
This Court reversed the conviction. It reasoned first that while Utah was a Territory, the Sixth Amendment applied to actions of the territorial government and guaranteed Thompson a right to a 12-person jury. 170 U.S., at 349 -350. The Court then held that "the State did not acquire upon its admission into the Union the power to provide, in respect of felonies committed within its limits while it was a Territory, that they should be tried otherwise than by a jury such as is provided by the Constitution of the United States. Id., at 350-351. Because the State Constitution "deprive[d] him of a substantial right involved in his liberty" and "materially alter[ed] the situation to his disadvantage," the Court concluded that Thompson's conviction was prohibited by the Ex Post Facto Clause. Id., at 352-353.
The result in Thompson v. Utah foreshadowed our decision in Duncan v. Louisiana, 391 U.S. 145 (1968), which held that the Sixth Amendment right to trial by jury - then believed to mean a jury of 12, see, e.g., Patton v. United States, 281 U.S. 276, 288 -289 (1930) - was incorporated and made applicable by the Fourteenth Amendment against the states. The Court held that, since Utah was a territory when Thompson's crime was committed, and therefore obligated to provide a 12-person jury by the Sixth Amendment, the Ex Post Facto Clause prevented the State from taking away that substantial right from him when it became a State and was no longer bound by the Sixth Amendment as then interpreted. The right to jury trial provided by the Sixth Amendment is obviously a "substantial" one, but it is not a right that has anything to do with the definition of crimes, defenses, or punishments, which is the concern of the Ex Post Facto Clause. To the extent that Thompson v. Utah [497 U.S. 37, 52] rested on the Ex Post Facto Clause and not the Sixth Amendment, we overrule it. 4
The Texas statute allowing reformation of improper verdicts does not punish as a crime an act previously committed which was innocent when done, nor make more burdensome the punishment for a crime after its commission, nor deprive one charged with crime of any defense available according to law at the time when the act was committed. Its application to respondent therefore is not prohibited by the Ex Post Facto Clause of Art. I, 10.
The judgment of the Court of Appeals is
[ Footnote 2 ] Although there has been some debate within the Court about the accuracy of the historical discussion in Calder v. Bull, see Satterlee v. Matthewson, 2 Pet. 380, 381 (1829) (note by Johnson, J.), the Court has consistently adhered to the view expressed by Justices Chase, Paterson, and Iredell in Calder that the Ex Post Facto Clause applies only to penal statutes.
[ Footnote 3 ] The Beazell definition omits the reference by Justice Chase in Calder v. Bull, 3 Dall. 386 (1798), to alterations in the "legal rules of evidence." See also Hopt v. Utah, 110 U.S. 574, 590 (1884) (approving procedural changes "leaving untouched the nature of the crime and the amount or degree of proof essential to conviction"). As cases subsequent to Calder make clear, this language was not intended to prohibit the application of new evidentiary rules in trials for crimes committed before the changes. Thompson v. Missouri, 171 U.S. 380, 386 -387 (1898) (rejecting ex post facto challenge to retroactive application of statute making admissible handwritten documents as handwriting exemplars); Hopt, supra, at 588-590 (upholding retroactive application of statute making felons competent to testify).
JUSTICE STEVENS, with whom JUSTICE BRENNAN and JUSTICE MARSHALL join, concurring in the judgment.
The "Ex Post Facto" Clause of the Constitution 1 has been construed to embrace any law that deprives a person accused of crime of a "substantial protection" that the law afforded at the time of the alleged offense. Thus, the Clause prohibits not only the retroactive creation of new criminal offenses and more harsh penalties but also substantial changes in procedure that are designed to protect the defendant from a wrongful conviction. The question in this case is whether a law that changed a post-conviction remedy for an erroneous sentence - by conforming it to the law in effect at the time of the offense instead of affording the defendant a new trial on all issues - effected a "substantial" deprivation within the meaning of our cases. I agree with the Court's conclusion [497 U.S. 37, 53] that the new law did not violate the Ex Post Facto Clause, but I believe that conclusion is entirely consistent with our precedents.
Respondent committed a crime that was punishable by a maximum sentence of life imprisonment, and was convicted on March 17, 1982. Under Texas law, it was the jury's task to impose sentence as well as to determine guilt or innocence. By consequence of a faulty instruction, respondent was improperly sentenced to life imprisonment and a fine of $10,000. The following year, in Bogany v. State, 661 S.W.2d 957 (1983), the Texas Court of Criminal Appeals held in a somewhat similar case that the fine was not authorized by law, and that no reviewing court had authority to correct such an erroneous sentence. Instead, the entire judgment was deemed "void" and the defendant was entitled to a new trial. 2 Understandably, the Texas legislature recognized that corrective legislation was in order, for it is difficult to understand why an error in sentencing should necessitate a second trial on the issue of guilt or innocence. [497 U.S. 37, 54]
Theoretically, the legislature might have remedied the situation in either of two ways. It might have authorized a punishment of both life imprisonment and a $10,000 fine for respondent's offense or, alternatively, it might have authorized a court to correct the sentence by eliminating the fine. The former option would plainly have violated the Ex Post Facto Clause, because it would have increased the respondent's punishment beyond the penalty authorized at the time of his offense. The second option, which the Texas legislature adopted, is not subject to that defect, nor does it criminalize previously innocent conduct or make any change in the procedures used to convict or to sentence respondent. It created a new remedy designed to conform respondent's sentence to that authorized by law at the time of his offense. Such legislation does not violate the Ex Post Facto Clause.
The argument to the contrary is based on our cases holding that the Clause applies to procedural, as well as substantive, changes that deprive a defendant of "substantial personal rights" and a claim that respondent's right to a new trial after an erroneous sentence was such a right. The argument misreads our precedents and overlooks the critical importance of evaluating the procedural right at issue by reference to the time of the offense.
In Kring v. Missouri, 107 U.S. 221 (1883), the Court rejected the argument that the Ex Post Facto Clause has no application to procedural changes. At the time of Kring's offense, Missouri law provided that the acceptance of a plea of guilty to second-degree murder constituted an acquittal of first-degree murder. A subsequent amendment to the Missouri Constitution abrogated that rule, and Kring was thereafter convicted of first-degree murder and sentenced to death. The Missouri Supreme Court held that there was no violation of the Ex Post Facto Clause, because the retroactive [497 U.S. 37, 55] amendment was merely a procedural change. 3 This Court's reversal of that holding demonstrates that the Clause applies to some procedural changes, but our decision rested on the fact that the change had deprived the defendant of a complete defense to the charge of first-degree murder and to the imposition of the death penalty. We wrote:
The foregoing cases make it clear that the mere fact that this case involves a procedural change in Texas law is not sufficient to exclude it from the coverage of the Ex Post Facto Clause. But it is equally clear that our analysis should focus on the impact of the change upon the "right belonging to [the defendant] when the offense was committed." Thompson, 170 U.S., at 351 . In this case, neither the defendant's right to a fair trial nor his right to be protected against unauthorized or excessive punishment has been impaired in the slightest by the new Texas rule.
This conclusion follows immediately from an observation which is both and evident from precedent: a procedural protection is likely to be substantial, when viewed from the time of the commission of the offense, only if it affects the modes of procedure by which a valid conviction or sentence may be imposed. The claims in Kring and Thompson both satisfy this threshold test. In Kring, the procedural change - which deprived Kring of a defense based upon an earlier trial or plea - made it easier for the State to obtain a first-degree murder charge against a defendant who had never been subject to any valid conviction for the crime in question, much less a valid conviction for first-degree murder. In Thompson, the reduction in the size of the jury made it easier for the State to obtain a unanimous verdict [497 U.S. 37, 59] against a defendant who, before the verdict, likewise had not been convicted.
Mallett v. North Carolina, 181 U.S. 589 (1901), is, however, distinguishable from Kring and Thompson, because it fails to meet the threshold test. In Mallett, a valid conviction had been obtained against the defendant. Under the defendant's theory in that case, however, the State would have been prohibited from relying upon this conviction, because it had been vacated by an intermediate appellate court. Although the North Carolina Supreme Court reinstated the conviction, Mallett claimed that it lacked power to do so. At the time Mallett committed his crime, the State was prohibited by state law from appealing the adverse decisions of intermediate appellate courts in criminal cases. This restriction had been removed, but Mallett contended that the State had thereby enacted an ex post facto law. As the case came to this Court, it was conceded that Mallett was convicted after a trial which afforded him all the procedural and substantive protections guaranteed by North Carolina law at the time he committed his offense. Nevertheless, according to Mallett's theory, the State was prohibited from relying upon his conviction because of the combination of an intervening - and, for this Court's purposes, erroneous - appellate decision and a restriction upon the State's access to the appellate processes. Not surprisingly, we rejected this claim.
This case is comparable to Mallett. Respondent does not claim that he was denied any procedural protections relevant to the determination of his guilt or innocence. Nor does he claim that his life sentence was unauthorized by law, or that it was the consequence of improper procedures. Finally, he does not argue that he has been deprived of any avenue of review for correcting errors that may have vitiated the validity of his conviction or sentence. For example, respondent does not contend - and we do not see how he could plausibly contend, that the State has deprived him of any opportunity to challenge his conviction on the ground that the improper [497 U.S. 37, 60] sentencing instruction somehow infected the jury's deliberations about his guilt or about the propriety of life imprisonment. Respondent instead claims, as did the defendant in Mallett, that an unrelated error must bar the State from relying upon his concededly valid conviction, and predicates this claim solely on a restriction upon the State's access to appellate - or, more precisely in this case, post-conviction - remedies. 4 Unlike the defendants in Thompson and Kring, Youngblood wishes to have a new trial according to the same procedures, regulated by the same laws, open to the. same evidence, and capped by the same sentencing limitations that resulted in his conviction and his life sentence. 5
Obviously, as our decision in Beazell itself makes clear, a procedural protection does not become substantial merely because it meets the low threshold that I have discussed. It does, however, become insubstantial by failing to do so. Whatever else may be said of the factors that determine whether a procedural protection affects substantial rights, it is difficult to imagine how a retroactive law could, when viewed from the standpoint of the date the offense was committed, implicate substantial rights of any defendant if the law does no more than expand the flexibility of postconviction processes available to the State with respect to a defendant [497 U.S. 37, 61] who is subject to a valid conviction and sentence. Indeed, respondent has barely even attempted to articulate any justification for the Texas procedure that the legislature abolished. The mere possibility of a capricious and unlikely windfall is not the sort of procedural protection that could reasonably be judged substantial from the perspective of the defendant at the time the offense was committed.
Accordingly, I concur in the Court's judgment, but not in its opinion.
[ Footnote 1 ] Art. I, 10 of the Constitution provides in part:
[ Footnote 2 ] At the time of respondent's offense, it apparently was well established under Texas law that, as a general proposition, when a criminal jury rendered a verdict not authorized by law the verdict was void at its inception. See, e.g., Ex parte McIver, 586 S.W.2d 851, 854 (Tex.Crim.App. 1979); Ocker v. State, 477 S.W.2d 288, 290 (Tex.Crim.App. 1972). However, until the Court of Criminal Appeals decided Bogany, there was some doubt both as to whether that general rule would apply to the error in this case, and as to whether the sentence imposed by the jury in this case was in fact unlawful. See, e.g., Adams v. State, 642 S.W.2d 211, 213-14 (Tex.App. 1982) (reforming jury's sentence); Bogany v. State, 646 S.W.2d 663, 664-665 (Tex.App.) (stating that jury's sentence could be reformed), rev'd, 661 S.W.2d 957 (Tex.Crim.App. 1983) id., at 960 (McCormick, J., dissenting) (contending that supplementary fine was authorized by law). For purposes of this opinion, I assume that both the substantive limitation upon respondent's sentence and the procedural limitation on the remedial powers of reviewing courts were law at the time that respondent's offense was committed.
[ Footnote 3 ] The Missouri Supreme Court relied upon the reasoning of the St. Louis Court of Appeals. See State v. Kring, 74 Mo. 612, 631 (1881). The relevant passage from the Court of Appeals opinion was quoted (and then disavowed) by this Court in Kring v. Missouri, 107 U.S. 221, 223 -224 (1883):
[ Footnote 4 ] In Mallett v. North Carolina, 181 U.S. 589 (1901), the unrelated intervening error was an incorrect decision by the intermediate appellate court; in this case, it was the imposition of a supplementary fine in addition to the life sentence. In Mallett, the restriction upon the review process prohibited the State from taking an appeal; in this case, it prohibits the courts from saving the conviction and sentence by removing the improper supplement.
[ Footnote 5 ] Indeed, this case is a fortiori by comparison to Mallett. In that case, the defendant would benefit from an evidentiary exclusion at the secondary trial, although that exclusion would be entirely the consequence of the appellate court's incorrect interpretation of State law, and not a consequence of the trial procedures established by North Carolina law in effect at the time of the offense. By contrast, in this case, the procedures at the second trial would be in all relevant respects identical. [497 U.S. 37, 62]