443 U.S. 307
99 S.Ct. 2781
61 L.Ed.2d 560
James A. JACKSON, Petitioner,
Commonwealth of VIRGINIA et al.
Argued March 21, 1979.
Decided June 28, 1979.
Rehearing Denied Oct. 1, 1979.
See 444 U.S. 890, 100 S.Ct. 195.
Petitioner was convicted of first-degree murder after a bench trial in a Virginia court, and his motion and petition in the state courts to set aside the conviction on the ground that there was insufficient evidence of premeditation, a necessary element of first-degree murder, were denied. He then brought a habeas corpus proceeding in Federal District Court, which, applying the “no evidence” criterion of Thompson v. Louisville, 362 U.S. 199, 80 S.Ct. 624, 4 L.Ed.2d 654, found the record devoid of evidence of premeditation and granted the writ. Applying the same criterion, the Court of Appeals reversed, holding that there was some evidence that petitioner had intended to kill the victim.
1. A federal habeas corpus court must consider not whether there was any evidence to support a state-court conviction, but whether there was sufficient evidence to justify a rational trier of fact to find guilt beyond a reasonable doubt. In re Winship, 397 U.S. 358, 90 S.Ct. 1068, 25 L.Ed.2d 368. Pp. 313-324.
(a) In re Winship presupposes as an essential of the due process guaranteed by the Fourteenth Amendment that no person shall be made to suffer the onus of a criminal conviction except upon sufficient proof?defined as evidence necessary to convince a trier of fact beyond a reasonable doubt of the existence of every element of the offense. Pp. 316-320.
(b) After In re Winship, the critical inquiry on review of the sufficiency of the evidence to support a criminal conviction must be not simply to determine whether the jury was properly instructed on reasonable doubt, but to determine whether the record evidence could reasonably support a finding of guilt beyond a reasonable doubt. The relevant question is whether after viewing the evidence in the light most favorable to the prosecution, any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt. The Thompson “no evidence” rule is simply inadequate to protect against misapplications of the constitutional standard of reasonable doubt. Pp. 313-324.
(c) In a challenge to a state conviction brought under 28 U.S.C. ? 2254, which requires a federal court to entertain a state prisoner’s claim that he is being held in “custody in violation of the Constitution or laws or treaties of the United States,” the applicant is entitled to habeas corpus relief if it is found that upon the evidence adduced at the trial no rational trier of fact could have found proof of guilt beyond a reasonable doubt. Pp. 320-324.
2. A review of the record in this case in the light most favorable to the prosecution shows that a rational factfinder could have found petitioner guilty beyond a reasonable doubt of first-degree murder under Virginia law. Pp. 324-326.
4th Cir., 580 F.2d 1048, affirmed.
Carolyn J. Colville, Richmond, Va., for petitioner, pro hac vice, by special leave of Court.
Marshall Coleman, Richmond, Va., for respondents.
Mr. Justice STEWART delivered the opinion of the Court.
The Constitution prohibits the criminal conviction of any person except upon proof of guilt beyond a reasonable doubt. In re Winship, 397 U.S. 358, 90 S.Ct. 1068, 25 L.Ed.2d 368. The question in this case is what standard is to be applied in a federal habeas corpus proceeding when the claim is made that a person has been convicted in a state court upon insufficient evidence.
* The petitioner was convicted after a bench trial in the Circuit Court of chesterfield County, Va., of the first-degree murder of a woman named Mary Houston Cole.1 Under Virginia law, murder is defined as “the unlawful killing of another with malice aforethought.” Stapleton v. Commonwealth, 123 Va. 825, 96 S.E. 801. Premeditation, or specific intent to kill, distinguishes murder in the first from murder in the second degree; proof of this element is essential to conviction of the former offense, and the burden of proving it clearly rests with the prosecution. Shiflett v. Commonwealth, 143 Va. 609, 130 S.E. 777; Jefferson v. Commonwealth, 214 Va. 432, 201 S.E.2d 749.
That the petitioner had shot and killed Mrs. Cole was not in dispute at the trial. The State’s evidence established that she had been a member of the staff at the local county jail, that she had befriended him while he was imprisoned there on a disorderly conduct charge, and that when he was released she had arranged for him to live in the home of her son and daughter-in-law. Testimony by her relatives indicated that on the day of the killing the petitioner had been drinking and had spent a great deal of time shooting at targets with his revolver. Late in the afternoon, according to their testimony, he had unsuccessfully attempted to talk the victim into driving him to North Carolina. She did drive the petitioner to a local diner. There the two were observed by several police officers, who testified that both the petitioner and the victim had been drinking. The two were observed by a deputy sheriff as they were preparing to leave the diner in her car. The petitioner was then in possession of his revolver, and the sheriff also observed a kitchen knife in the automobile. The sheriff testified that he had offered to keep the revolver until the petitioner sobered up, but that the latter had indicated that this would be unnecessary since he and the victim were about to engage in sexual activity.
Her body was found in a secluded church parking lot a day and a half later, naked from the waist down, her slacks beneath her body. Uncontradicted medical and expert evidence established that she had been shot twice at close range with the petitioner’s gun. She appeared not to have been sexually molested. Six cartridge cases identified as having been fired from the petitioner’s gun were found near the body.
After shooting Mrs. Cole, the petitioner drove her car to North Carolina, where, after a short trip to Florida, he was arrested several days later. In a postarrest statement, introduced in evidence by the prosecution, the petitioner admitted that he had shot the victim. He contended, however, that the shooting had been accidental. When asked to describe his condition at the time of the shooting, he indicated that he had not been drunk, but had been “pretty high.” His story was that the victim had attacked him with a knife when he resisted her sexual advances. He said that he had defended himself by firing a number of warning shots into the ground, and had then reloaded his revolver. The victim, he said, then attempted to take the gun from him, and the gun “went off” in the ensuing struggle. He said that he fled without seeking help for the victim because he was afraid. At the trial, his position was that he had acted in self-defense. Alternatively, he claimed that in any event the State’s own evidence showed that he had been too intoxicated to form the specific intent necessary under Virginia law to sustain a conviction of murder in the first degree.2
The trial judge, declaring himself convinced beyond a reasonable doubt that the petitioner had committed first-degree murder, found him guilty of that offense.3 The petitioner’s motion to set aside the judgment as contrary to the evidence was denied, and he was sentenced to serve a term of 30 years in the Virginia state penitentiary. A petition for writ of error to the Virginia Supreme Court on the ground that the evidence was insufficient to support the conviction was denied.4
The petitioner then commenced this habeas corpus proceeding in the United States District Court for the Eastern District of Virginia, raising the same basic claim.5 Applying the “no evidence” criterion of Thompson v. Louisville, 362 U.S. 199, 80 S.Ct. 624, 4 L.Ed.2d 654, the District Court found the record devoid of evidence of premeditation and granted the writ. The Court of Appeals for the Fourth Circuit reversed the judgment.6 The court noted that a dissent from the denial of certiorari in a case in this Court had exposed the question whether the constitutional rule of In re Winship, 397 U.S. 358, 90 S.Ct. 1068, 25 L.Ed.2d 368, might compel a new criterion by which the validity of a state criminal conviction must be tested in a federal habeas corpus proceeding. See Freeman v. Zahradnick, 429 U.S. 1111, 97 S.Ct. 1150, 51 L.Ed.2d 566 (dissent from denial of certiorari). But the appellate court held that in the absence of further guidance from this Court it would apply the same “no evidence” criterion of Thompson v. Louisville that the District Court had adopted. The court was of the view that some evidence that the petitioner had intended to kill the victim could be found in the facts that the petitioner had reloaded his gun after firing warning shots, that he had had time to do so, and that the victim was then shot not once but twice. The court also concluded that the state trial judge could have found that the petitioner was not so intoxicated as to be incapable of premeditation.
We granted certiorari to consider the petitioner’s claim that under In re Winship, supra, a federal habeas corpus court must consider not whether there was any evidence to support a state-court conviction, but whether there was sufficient evidence to justify a rational trier of the facts to find guilt beyond a reasonable doubt. 439 U.S. 1001, 99 S.Ct. 609, 58 L.Ed.2d 676.
Our inquiry in this case is narrow. The petitioner has not seriously questioned any aspect of Virginia law governing the allocation of the burden of production or persuasion in a murder trial. See Mullaney v. Wilbur, 421 U.S. 684, 95 S.Ct. 1881, 44 L.Ed.2d 508; Patterson v. New York, 432 U.S. 197, 97 S.Ct. 2319, 53 L.Ed.2d 281. As the record demonstrates, the judge sitting as factfinder in the petitioner’s trial was aware that the State bore the burden of establishing the element of premeditation, and stated that he was applying the reasonable-doubt standard in his appraisal of the State’s evidence. The petitioner, moreover, does not contest the conclusion of the Court of Appeals that under the “no evidence” rule of Thompson v. Louisville, supra, his conviction of first-degree murder is sustainable. And he has not attacked the sufficiency of the evidence to support a conviction of second-degree murder. His sole constitutional claim, based squarely upon Winship, is that the District Court and the Court of Appeals were in error in not recognizing that the question to be decided in this case is whether any rational factfinder could have concluded beyond a reasonable doubt that the killing for which the petitioner was convicted was premeditated. The question thus raised goes to the basic nature of the constitutional right recognized in the Winship opinion.
This is the first of our cases to expressly consider the question whether the due process standard recognized in Winship constitutionally protects an accused against conviction except upon evidence that is sufficient fairly to support a conclusion that every element of the crime has been established beyond a reasonable doubt. Upon examination of the fundamental differences between the constitutional underpinnings of Thompson v. Louisville, supra, and of In re Winship, supra, the answer to that question, we think, is clear.
It is axiomatic that a conviction upon a charge not made or upon a charge not tried constitutes a denial of due process. Cole v. Arkansas, 333 U.S. 196, 201, 68 S.Ct. 514, 517, 92 L.Ed. 644; Presnell v. Georgia, 439 U.S. 14, 99 S.Ct. 235, 58 L.Ed.2d 207. These standards no more than reflect a broader premise that has never been doubted in our constitutional system: that a person cannot incur the loss of liberty for an offense without notice and a meaningful opportunity to defend. E. g., Hovey v. Elliott, 167 U.S. 409, 416-420, 17 S.Ct. 841, 844-846, 42 L.Ed. 215. Cf. Boddie v. Connecticut, 401 U.S. 371, 377-379, 91 S.Ct. 780, 785-787, 28 L.Ed.2d 113. A meaningful opportunity to defend, if not the right to a trial itself, presumes as well that a total want of evidence to support a charge will conclude the case in favor of the accused. Accordingly, we held in the Thompson case that a conviction based upon a record wholly devoid of any relevant evidence of a crucial element of the offense charged is constitutionally infirm. See also Vachon v. New Hampshire, 414 U.S. 478, 94 S.Ct. 664, 38 L.Ed.2d 666; Adderley v. Florida, 385 U.S. 39, 87 S.Ct. 242, 17 L.Ed.2d 149; Gregory v. Chicago, 394 U.S. 111, 89 S.Ct. 946, 22 L.Ed.2d 134; Douglas v. Buder, 412 U.S. 430, 93 S.Ct. 2199, 37 L.Ed.2d 52. The “no evidence” doctrine of Thompson v. Louisville thus secures to an accused the most elemental of due process rights: freedom from a wholly arbitrary deprivation of liberty.
The Court in Thompson explicitly stated that the due process right at issue did not concern a question of evidentiary “sufficiency.” 362 U.S., at 199, 80 S.Ct., at 625. The right established in In re Winship, however, clearly stands on a different footing. Winship involved an adjudication of juvenile delinquency made by a judge under a state statute providing that the prosecution must prove the conduct charged as delinquent?which in Winship would have been a criminal offense if engaged in by an adult?by a preponderance of the evidence. Applying that standard, the judge was satisfied that the juvenile was “guilty,” but he noted that the result might well have been different under a standard of proof beyond a reasonable doubt. In short, the record in Winship was not totally devoid of evidence of guilt.
The constitutional problem addressed in Winship was thus distinct from the stark problem of arbitrariness presented in Thompson v. Louisville. In Winship, the Court held for the first time that the Due Process Clause of the Fourteenth Amendment protects a defendant in a criminal case against conviction “except upon proof beyond a reasonable doubt of every fact necessary to constitute the crime with which he is charged.” 397 U.S., at 364, 90 S.Ct., at 1073. In so holding, the Court emphasized that proof beyond a reasonable doubt has traditionally been regarded as the decisive difference between criminal culpability and civil liability. Id., at 358-362, 90 S.Ct., at 1068-1072. See Davis v. United States, 160 U.S. 469, 16 S.Ct. 353, 40 L.Ed. 499; Brinegar v. United States, 338 U.S. 160, 174, 69 S.Ct. 1302, 1310, 93 L.Ed. 1879; Leland v. Oregon, 343 U.S. 790, 72 S.Ct. 1002, 96 L.Ed. 1302; 9 J. Wigmore, Evidence ? 2495, pp. 307-308 (3d ed. 1940). Cf. Woodby v. INS, 385 U.S. 276, 285, 87 S.Ct. 483, 487, 17 L.Ed.2d 362. The standard of proof beyond a reasonable doubt, said the Court, “plays a vital role in the American scheme of criminal procedure,” because it operates to give “concrete substance” to the presumption of innocence to ensure against unjust convictions, and to reduce the risk of factual error in a criminal proceeding. 397 U.S., at 363, 90 S.Ct., at 1072. At the same time by impressing upon the factfinder the need to reach a subjective state of near certitude of the guilt of the accused, the standard symbolizes the significance that our society attaches to the criminal sanction and thus to liberty itself. Id., at 372, 90 S.Ct., at 1076 (Harlan, J., concurring).
The constitutional standard recognized in the Winship case was expressly phrased as one that protects an accused against a conviction except on “proof beyond a reasonable doubt . . . .” In subsequent cases discussing the reasonable-doubt standard, we have never departed from this definition of the rule or from the Winship understanding of the central purposes it serves. See, e. g., Ivan V. v. City of New York, 407 U.S. 203, 204, 92 S.Ct. 1951, 1952, 32 L.Ed.2d 659; Lego v. Twomey, 404 U.S. 477, 486-487, 92 S.Ct. 619, 625-626, 30 L.Ed.2d 618; Mullaney v. Wilbur, 421 U.S. 684, 95 S.Ct. 1881, 44 L.Ed.2d 508; Patterson v. New York, 432 U.S. 197, 97 S.Ct. 2319, 53 L.Ed.2d 281; Cool v. United States, 409 U.S. 100, 104, 93 S.Ct. 354, 357, 34 L.Ed.2d 335. In short, Winship presupposes as an essential of the due process guaranteed by the Fourteenth Amendment that no person shall be made to suffer the onus of a criminal conviction except upon sufficient proof?defined as evidence necessary to convince a trier of fact beyond a reasonable doubt of the existence of every element of the offense.
Although several of our cases have intimated that the factfinder’s application of the reasonable-doubt standard to the evidence may present a federal question when a state conviction is challenged, Lego v. Twomey, supra, 404 U.S., at 487, 92 S.Ct., at 625; Johnson v. Louisiana, 406 U.S. 356, 360, 92 S.Ct. 1620, 1623, 32 L.Ed.2d 152, the Federal Courts of Appeals have generally assumed that so long as the reasonable-doubt instruction has been given at trial, the no-evidence doctrine of Thompson v. Louisville remains the appropriate guide for a federal habeas corpus court to apply in assessing a state prisoner’s challenge to his conviction as founded upon insufficient evidence. See, e. g., Cunha v. Brewer, 511 F.2d 894 (CA8).7 We cannot agree.
The Winship doctrine requires more than simply a trial ritual. A doctrine establishing so fundamental a substantive constitutional standard must also require that the factfinder will rationally apply that standard to the facts in evidence.8 A “reasonable doubt,” at a minimum, is one based upon “reason.”9 Yet a properly instructed jury may occasionally convict even when it can be said that no rational trier of fact could find guilt beyond a reasonable doubt, and the same may be said of a trial judge sitting as a jury. In a federal trial, such an occurrence has traditionally been deemed to require reversal of the conviction. Glasser v. United States, 315 U.S. 60, 80, 62 S.Ct. 457, 469, 86 L.Ed. 680; Bronston v. United States, 409 U.S. 352, 93 S.Ct. 595, 34 L.Ed.2d 568. See also, e. g., Curley v. United States, 81 U.S.App.D.C. 389, 392-393, 160 F.2d 229, 232-233.10 Under Winship, which established proof beyond a reasonable doubt as an essential of Fourteenth Amendment due process, it follows that when such a conviction occurs in a state trial, it cannot constitutionally stand.
A federal court has a duty to assess the historic facts when it is called upon to apply a constitutional standard to a conviction obtained in a state court. For example, on direct review of a state-court conviction, where the claim is made that an involuntary confession was used against the defendant, this Court reviews the facts to determine whether the confession was wrongly admitted in evidence. Blackburn v. Alabama, 361 U.S. 199, 205-210, 80 S.Ct. 274, 279-282, 4 L.Ed.2d 242. Cf. Drope v. Missouri, 420 U.S. 162, 174-175, and n. 10, 95 S.Ct. 896, 905-906, and n. 10, 43 L.Ed.2d 103. The same duty obtains in federal habeas corpus proceedings. See Townsend v. Sain, 372 U.S. 293, 318, 83 S.Ct. 745, 759, 9 L.Ed.2d 770; Brown v. Allen, 344 U.S. 443, 506-507, 73 S.Ct. 397, 445-446, 97 L.Ed. 469 (opinion of Frankfurter, J.).
After Winship the critical inquiry on review of the sufficiency of the evidence to support a criminal conviction must be not simply to determine whether the jury was properly instructed, but to determine whether the record evidence could reasonably support a finding of guilt beyond a reasonable doubt.11 But this inquiry does not require a court to “ask itself whether it believes that the evidence at the trial established guilt beyond a reasonable doubt.” Woodby v. INS, 385 U.S., at 282, 87 S.Ct., at 486 (emphasis added). Instead, the relevant question is whether, after viewing the evidence in the light most favorable to the prosecution, any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt. See Johnson v. Louisiana, 406 U.S., at 362, 92 S.Ct., at 1624-1625. This familiar standard gives full play to the responsibility of the trier of fact fairly to resolve conflicts in the testimony, to weigh the evidence, and to draw reasonable inferences from basic facts to ultimate facts. Once a defendant has been found guilty of the crime charged, the factfinder’s role as weigher of the evidence is preserved through a legal conclusion that upon judicial review all of the evidence is to be considered in the light most favorable to the prosecution.12 The criterion thus impinges upon “jury” discretion only to the extent necessary to guarantee the fundamental protection of due process of law.13
That the Thompson “no evidence” rule is simply inadequate to protect against misapplications of the constitutional standard of reasonable doubt is readily apparent. “[A] mere modicum of evidence may satisfy a ‘no evidence’ standard . . . .” Jacobellis v. Ohio, 378 U.S. 184, 202, 84 S.Ct. 1676, 1686, 12 L.Ed.2d 793 (Warren, C.J., dissenting). Any evidence that is relevant?that has any tendency to make the existence of an element of a crime slightly more probable than it would be without the evidence, cf. Fed.Rule Evid. 401?could be deemed a “mere modicum.” But it could not seriously be argued that such a “modicum” of evidence could by itself rationally support a conviction beyond a reasonable doubt. The Thompson doctrine simply fails to supply a workable or even a predictable standard for determining whether the due process command ofWinship has been honored.14
Under 28 U.S.C. ? 2254, a federal court must entertain a claim by a state prisoner that he or she is being held in “custody in violation of the Constitution or laws or treaties of the United States.” Under the Winship decision, it is clear that a state prisoner who alleges that the evidence in support of his state conviction cannot be fairly characterized as sufficient to have led a rational trier of fact to find guilt beyond a reasonable doubt has stated a federal constitutional claim. Thus, assuming that state remedies have been exhausted, see 28 U.S.C. ? 2254(b), and that no independent and adequate state ground stands as a bar, see Estelle v. Williams, 425 U.S. 501, 96 S.Ct. 1691, 48 L.Ed.2d 126; Francis v. Henderson, 425 U.S. 536, 96 S.Ct. 1708, 48 L.Ed.2d 149; Wainwright v. Sykes, 433 U.S. 72, 97 S.Ct. 2497, 53 L.Ed.2d 594; Fay v. Noia, 372 U.S. 391, 438, 83 S.Ct. 822, 848, 9 L.Ed.2d 837, it follows that such a claim is cognizable in a federal habeas corpus proceeding. The respondents have argued, nonetheless, that a challenge to the constitutional sufficiency of the evidence should not be entertained by a federal district court under 28 U.S.C. ? 2254.
In addition to the argument that a Winship standard invites replication of state criminal trials in the guise of ? 2254 proceedings?an argument that simply fails to recognize that courts can and regularly do gauge the sufficiency of the evidence without intruding into any legitimate domain of the trier of fact?the respondents have urged that any departure from the Thompson test in federal habeas corpus proceedings will expand the number of meritless claims brought to the federal courts, will duplicate the work of the state appellate courts, will disserve the societal interest in the finality of state criminal proceedings, and will increase friction between the federal and state judiciaries. In sum, counsel for the State urges that this type of constitutional claim should be deemed to fall within the limit on federal habeas corpus jurisdiction identified in Stone v. Powell, 428 U.S. 465, 96 S.Ct. 3037, 49 L.Ed.2d 1067, with respect to Fourth Amendment claims. We disagree.
First, the burden that is likely to follow from acceptance of the Winship standard has, we think, been exaggerated. Federal-court challenges to the evidentiary support for state convictions have since Thompson been dealt with under ? 2254. E. g., Freeman v. Stone, 444 F.2d 113 (CA9); Grieco v. Meachum, 533 F.2d 713 (CA1); Williams v. Peyton, 414 F.2d 776 (CA4). A more stringent standard will expand the contours of this type of claim, but will not create an entirely new class of cases cognizable on federal habeas corpus. Furthermore, most meritorious challenges to constitutional sufficiency of the evidence undoubtedly will be recognized in the state courts, and, if the state courts have fully considered the issue of sufficiency, the task of a federal habeas court should not be difficult. Cf. Brown v. Allen, 344 U.S., at 463, 73 S.Ct., at 410.15 And this type of claim can almost always be judged on the written record without need for an evidentiary hearing in the federal court.
Second, the problems of finality and federal-state comity arise whenever a state prisoner invokes the jurisdiction of a federal court to redress an alleged constitutional violation. A challenge to a state conviction brought on the ground that the evidence cannot fairly be deemed sufficient to have established guilt beyond a reasonable doubt states a federal constitutional claim. Although state appellate review undoubtedly will serve in the vast majority of cases to vindicate the due process protection that follows from Winship, the same could also be said of the vast majority of other federal constitutional rights that may be implicated in a state criminal trial. It is the occasional abuse that the federal writ of habeas corpus stands ready to correct. Brown v. Allen, supra, at 498-501, 73 S.Ct., at 441-443 (opinion of Frankfurter, J.).
The respondents have argued nonetheless that whenever a person convicted in a state court has been given a “full and fair hearing” in the state system?meaning in this instance state appellate review of the sufficiency of the evidence?further federal inquiry?apart from the possibility of discretionary review by this Court?should be foreclosed. This argument would prove far too much. A judgment by a state appellate court rejecting a challenge to evidentiary sufficiency is of course entitled to deference by the federal courts, as is any judgment affirming a criminal conviction. But Congress in ? 2254 has selected the federal district courts as precisely the forums that are responsible for determining whether state convictions have been secured in accord with federal constitutional law. The federal habeas corpus statute presumes the norm of a fair trial in the state court and adequate state postconviction remedies to redress possible error. See 28 U.S.C. ? 2254(b), (d). What it does not presume is that these state proceedings will always be without error in the constitutional sense. The duty of a federal habeas corpus court to appraise a claim that constitutional error did occur?reflecting as it does the belief that the “finality” of a deprivation of liberty through the invocation of the criminal sanction is simply not to be achieved at the expense of a constitutional right?is not one that can be so lightly abjured.
The constitutional issue presented in this case is far different from the kind of issue that was the subject of the Court’s decision in Stone v. Powell, supra. The question whether a defendant has been convicted upon inadequate evidence is central to the basic question of guilt or innocence. The constitutional necessity of proof beyond a reasonable doubt is not confined to those defendants who are morally blameless. E. g., Mullaney v. Wilbur, 421 U.S., at 697-698, 95 S.Ct., at 1888-1889 (requirement of proof beyond a reasonable doubt is not “limit[ed] to those facts which, that if not proved, would wholly exonerate” the accused). Under our system of criminal justice even a thief is entitled to complain that he has been unconstitutionally convicted and imprisoned as a burglar.
We hold that in a challenge to a state criminal conviction brought under 28 U.S.C. ? 2254?if the settled procedural prerequisites for such a claim have otherwise been satisfied?the applicant is entitled to habeas corpus relief if it is found that upon the record evidence adduced at the trial no rational trier of fact could have found proof of guilt beyond a reasonable doubt.16
Turning finally to the specific facts of this case, we reject the petitioner’s claim that under the constitutional standard dictated by Winship his conviction of first-degree murder cannot stand. A review of the record in the light most favorable to the prosecution convinces us that a rational factfinder could readily have found the petitioner guilty beyond a reasonable doubt of first-degree murder under Virginia law.
There was no question at the trial that the petitioner had fatally shot Mary Cole. The crucial factual dispute went to the sufficiency of the evidence to support a finding that he had specifically intended to kill her. This question, as the Court of Appeals recognized, must be gauged in the light of applicable Virginia law defining the element of premeditation. Under that law it is well settled that premeditation need not exist for any particular length of time, and that an intent to kill may be formed at the moment of the commission of the unlawful act. Commonwealth v. Brown, 90 Va. 671, 19 S.E. 447. From the circumstantial evidence in the record, it is clear that the trial judge could reasonably have found beyond a reasonable doubt that the petitioner did possess the necessary intent at or before the time of the killing.
The prosecution’s uncontradicted evidence established that the petitioner shot the victim not once but twice. The petitioner himself admitted that the fatal shooting had occurred only after he had first fired several shots into the ground and then reloaded his gun. The evidence was clear that the two shots that killed the victim were fired at close, and thus predictably fatal, range by a person who was experienced in the use of the murder weapon. Immediately after the shooting, the petitioner drove without mishap from Virginia to North Carolina, a fact quite at odds with his story of extreme intoxication. Shortly before the fatal episode, he had publicly expressed an intention to have sexual relations with the victim. Her body was found partially unclothed. From these uncontradicted circumstances, a rational factfinder readily could have inferred beyond a reasonable doubt that the petitioner, notwithstanding evidence that he had been drinking on the day of the killing, did have the capacity to form and had in fact formed an intent to kill the victim.
The petitioner’s calculated behavior both before and after the killing demonstrated that he was fully capable of committing premeditated murder. His claim of self-defense would have required the trial judge to draw a series of improbable inferences from the basic facts, prime among them the inference that he was wholly uninterested in sexual activity with the victim but that she was so interested as to have willingly removed part of her clothing and then attacked him with a knife when he resisted her advances, even though he was armed with a loaded revolver that he had just demonstrated he knew how to use. It is evident from the record that the trial judge found this story, including the petitioner’s belated contention that he had been so intoxicated as to be incapable of premeditation, incredible.
Only under a theory that the prosecution was under an affirmative duty to rule out every hypothesis except that of guilt beyond a reasonable doubt could this petitioner’s challenge be sustained. That theory the Court has rejected in the past. Holland v. United States, 348 U.S. 121, 140, 75 S.Ct. 127, 137, 99 L.Ed. 150. We decline to adopt it today. Under the standard established in this opinion as necessary to preserve the due process protection recognized in Winship, a federal habeas corpus court faced with a record of historical facts that supports conflicting inferences must presume?even if it does not affirmatively appear in the record?that the trier of fact resolved any such conflicts in favor of the prosecution, and must defer to that resolution. Applying these criteria, we hold that a rational trier of fact could reasonably have found that the petitioner committed murder in the first degree under Virginia law.
For these reasons, the judgment of the Court of Appeals is affirmed.
It is so ordered.
Mr. Justice POWELL took no part in the consideration or decision of this case.
Mr. Justice STEVENS, with whom THE CHIEF JUSTICE and Mr. Justice REHNQUIST join, concurring in the judgment.
The Constitution prohibits the criminal conviction of any person except upon proof sufficient to convince the trier of fact of guilt beyond a reasonable doubt. Cf. ante, at 309. This rule has prevailed in our courts “at least from our early years as a Nation.” In re Winship, 397 U.S. 358, 361, 90 S.Ct. 1068, 1071, 25 L.Ed.2d 368.
Today the Court creates a new rule of law?one that has never prevailed in our jurisprudence. According to the Court, the Constitution now prohibits the criminal conviction of any person including, apparently, a person against whom the facts have already been found beyond a reasonable doubt by a jury, a trial judge, and one or more levels of state appellate judges?except upon proof sufficient to convince a federal judge that a “rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt.” Ante, at 319.
The adoption of this novel constitutional rule is not necessary to the decision of this case. Moreover, I believe it is an unwise act of lawmaking. Despite its chimerical appeal as a new counterpart to the venerable principle recognized in Winship, I am persuaded that its precipitous adoption will adversely affect the quality of justice administered by federal judges. For that reason I shall analyze this new brainchild with some care.
I shall begin by explaining why neither the record in this case, nor general experience with challenges to the sufficiency of the evidence supporting criminal convictions, supports, much less compels, the conclusion that there is any need for this new constitutional precept. I shall next show that it is not logically compelled by either the holding or the analysis in In re Winship, supra. Finally, I shall try to demonstrate why the Court’s new rule?if it is not just a meaningless shibboleth threatens serious harm to the quality of our judicial system.
* It is, of course, part of this Court’s tradition that new rules of law emerge from the process of case-by-case adjudication of constitutional issues. Widespread concern that existing constitutional doctrine is unjust often provides the occasion, and is sometimes even relied upon as a justification, for the exercise of such lawmaking authority by the Court. Without entering the debate over the legitimacy of this justification for judicial action, it is at least certain that it should not be the basis for dramatic?indeed, for any ?constitutional lawmaking efforts unless (1) those efforts are necessary to the decision of the case at hand and (2) powerful reasons favor a change in the law. See Ashwander v. TVA, 297 U.S. 288, 345-348, 56 S.Ct. 466, 482-483, 80 L.Ed. 688 (Brandeis, J., concurring).
In this case, the Court’s analysis fails on both counts. It has accordingly formulated a new constitutional principle under the most dangerous possible circumstances?i. e., where the exercise of judicial authority is neither necessitated nor capable of being limited by “the precise facts to which [the rule is originally] to be applied,” Liverpool, N. Y. & P. S. S. Co. v. Emigration Comm’rs, 113 U.S. 33, 39, 5 S.Ct. 352, 355, 28 L.Ed. 899, nor even by some broader set of identifiable experiences with the evil supposedly involved.
Most significantly, the Court has announced its new constitutional edict in a case in which it has absolutely no bearing on the outcome. The only factual issue at stake is whether petitioner intended to kill his victim. If the evidence is viewed “in the light most favorable to the prosecution,” ante, at 319?and, indeed, we may view it through the eyes of the actual factfinder, whose observations about the evidence are recorded in the trial transcript?there can be only one answer to that question no matter what standard of appellate review is applied. In Part IV of its opinion, the Court accepts this conclusion. There is, therefore, no need to fashion a broad new rule of constitutional law to dispose of this squalid but rather routine murder case. Under any view, the evidence is sufficient.
The Court’s new rule is adopted simply to forestall some hypothetical evil that has not been demonstrated, and in my view is not fairly demonstrable. Although the Judiciary has received its share of criticism?principally because of the delays and costs associated with litigation?I am aware of no general dissatisfaction with the accuracy of the factfinding process or the adequacy of the rules applied by state appellate courts when reviewing claims of insufficiency.
What little evidence the Court marshals in favor of a contrary conclusion is unconvincing. See ante, at 317-318 n. 10. The Court is simply incorrect in implying that there are a significant number of occasions when federal convictions are overturned on appeal because no rational trier of fact could have found guilt beyond a reasonable doubt. The two opinions of this Court cited ante, at 317, stand for no such proposition. In neither was a conviction reversed for insufficiency. See Glasser v. United States, 315 U.S. 60, 62 S.Ct. 457, 86 L.Ed. 680; Bronston v. United States, 409 U.S. 352, 93 S.Ct. 595, 34 L.Ed.2d 568.
Moreover, a study of the 127 federal criminal convictions that were reviewed by the various Courts of Appeals and reported in the most recent hardbound volume of the Federal Reporter, Second Series, Volume 589, reveals that only 3 were overturned on sufficiency grounds. And of those, one was overturned under a “no evidence” standard, while the other two, in which a total of only 3 out of 36 counts were actually reversed, arguably involved legal issues masquerading as sufficiency questions.1 It is difficult to believe that the federal courts will turn up more sufficiency problems than this on habeas review when, instead of acting as the first level of review, as in the cases studied, they will be acting as the second, third, or even fourth level of appellate review. In short, there is simply no reason to tinker with an elaborate mechanism that is now functioning well.
There is nothing in the facts of this case or, so far as the Court has demonstrated, in those of cases like it to warrant today’s excursion into constitutional rulemaking. The Court instead portrays its rule as the logical corollary of the principle recognized in Winship regarding the subjective state of mind that persons charged with the responsibility of evaluating the credibility of evidence must possess before they find the defendant guilty in a criminal case. But an examination of Winship reveals that it has nothing to do with appellate, much less habeas corpus, review standards; that the reasoning used in that case to reach its conclusion with respect to the trier of fact does not support, and indeed counsels against, the Court’s conclusion with respect to federal habeas judges; and that there is no necessary connection between the rule recognized in Winship and the rule invented by the Court today.
In distinct contrast to the circumstances of this case, the facts of Winship presented “a case where the choice of the standard of proof has made a difference: the [trial] judge below forthrightly acknowledged that he believed by a preponderance of the evidence [in], but was not convinced beyond a reasonable doubt” of, the juvenile’s guilt. 397 U.S., at 369, 90 S.Ct., at 1075 (Harlan, J., concurring). Because the trier of fact entertained such a doubt, this Court held that the juvenile was constitutionally entitled to the same verdict that an adult defendant in a criminal case would receive. In so holding, the Court merely extended to juveniles a protection that had traditionally been available to defendants in criminal trials in this Nation. Id., at 361, 90 S.Ct., at 1071.
But nothing in the Winship opinion suggests that it also bore on appellate or habeas corpus procedures. Although it repeatedly emphasized the function of the reasonable-doubt standard as describing the requisite “subjective state of certitude” of the “factfinder,”2 it never mentioned the question of how appellate judges are to know whether the trier of fact really was convinced beyond a reasonable doubt, or indeed, whether the factfinder was a “rational” person or group of persons.
Moreover, the mode of analysis employed in Winship finds no counterpart in the Court’s opinion in this case. For example, in Winship, the Court pointed out the breadth of both the historical and the current acceptance of the reasonable-doubt trial standard.3 In this case, by contrast, the Court candidly recognizes that the Federal Courts of Appeals have “generally”rejected the habeas standard that it adopts today. Ante, at 316.4
The Winship court relied on nine prior opinions of this Court that bore directly on the issue presented. 397 U.S., at 362, 90 S.Ct., at 1971. Here, the Court purportedly relies on two prior decisions, but as is pointed out, supra, at 2794, neither of these cases itself applied a “reasonable doubt” appellate standard to overturn a conviction, neither purported to be interpreting the Constitution, and neither expressed any view whatsoever on the appropriate standard in collateral proceedings such as are involved in this case.5 As the Court itself notes, we have instead repeatedly endorsed the “no evidence” test, and have continued to do so after Winship was decided. Vachon v. New Hampshire, 414 U.S. 478, 94 S.Ct. 664, 38 L.Ed.2d 666; Douglas v. Buder, 412 U.S. 430, 93 S.Ct. 2199, 37 L.Ed.2d 52; Gregory v. Chicago, 394 U.S. 111, 89 S.Ct. 946, 22 L.Ed.2d 134; Adderley v. Florida, 385 U.S. 39, 87 S.Ct. 242, 17 L.Ed.2d 149; Thompson v. Louisville, 362 U.S. 199, 80 S.Ct. 624, 4 L.Ed.2d 654. See also Clyatt v. United States, 197 U.S. 207, 222, 25 S.Ct. 429, 432, 49 L.Ed. 726.
The primary reasoning of the Court in Winship is also inapplicable here. The Court noted in that case that the reasonable-doubt standard has the desirable effect of significantly reducing the risk of an inaccurate factfinding and thus of erroneous convictions, as well as of instilling confidence in the criminal justice system. 397 U.S., at 363-364, 90 S.Ct., at 1072-1073. See also id., at 370-372, 90 S.Ct., at 1075-1077 (Harlan, J., concurring). In this case, however, it would be impossible (and the Court does not even try) to demonstrate that there is an appreciable risk that a factfinding made by a jury beyond a reasonable doubt, and twice reviewed by a trial judge in ruling on directed verdict and post-trial acquittal motions and by one or more levels of appellate courts on direct appeal, as well as by two federal habeas courts under the Thompson “no evidence” rule, is likely to be erroneous.6 Indeed, the very premise of Winship is that properly selected judges and properly instructed juries act rationally, that the former will tell the truth when they declare that they are convinced beyond a reasonable doubt and the latter will conscientiously obey and understand the reasonable-doubt instructions they receive before retiring to reach a verdict, and therefore that either factfinder will itself provide the necessary bulwark against erroneous factual determinations. To presume otherwise is to make light of Winship.7
Having failed to identify the evil against which the rule is directed, and having failed to demonstrate how it follows from the analysis typically used in due process cases of this character, the Court places all of its reliance on a dry, and in my view incorrect syllogism: If Winship requires the factfinder to apply a reasonable-doubt standard, then logic requires a reviewing judge to apply a like standard.
But, taken to its ultimate conclusion, this “logic” would require the reviewing court to “ask itself whether it believes that the evidence at the trial established guilt beyond a reasonable doubt.” Woodby v. INS, 385 U.S. 276, 282, 87 S.Ct. 483, 486, 17 L.Ed.2d 362 (emphasis added). The Court, however, rejects this standard, as well as others that might be considered consistent with Winship. For example, it does not require the reviewing court to view just the evidence most favorable to the prosecution and then to decide whether that evidence convinced it beyond a reasonable doubt, nor whether, based on the entire record, rational triers of fact could be convinced of guilt beyond a reasonable doubt. Instead, and without explanation, it chooses a still narrower standard that merely asks whether, “after viewing the evidence in the light most favorable to the prosecution, any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt.” Ante, at 319.8 It seems to me that if “logic” allows this choice after Winship it should also allow the presumption that the Court has rejected?that trial judges and juries will act rationally and honestly in applying the reasonable-doubt standard, at least so long as the trial is free of procedural error and the record contains evidence tending to prove each of the elements of the offense.
Time may prove that the rule the Court has adopted today is the wisest compromise between one extreme that maximizes the protection against the risk that innocent persons will be erroneously convicted and the other extreme that places the greatest faith in the ability of fair procedures to produce just verdicts. But the Court’s opinion should not obscure the fact that its new rule is not logically compelled by the analysis or the holding in Winship or in any other precedent, or the fact that the rule reflects a new policy choice rather than the application of a pre-existing rule of law.
The Court cautions against exaggerating the significance of its new rule. Ante, at 321. It is true that in practice there may be little or no difference between a record that does not contain at least some evidence tending to prove every element of an offense and a record containing so little evidence that no rational factfinder could be persuaded of guilt beyond a reasonable doubt. Moreover, I think the Court is quite correct when it acknowledges that “most meritorious challenges to constitutional sufficiency of the evidence undoubtedly will be recognized in the state courts.” Ibid. But this only means that the new rule will seldom, if ever, provide a convicted state prisoner with any tangible benefits. It does not mean that the rule will have no impact on the administration of justice. On the contrary, I am persuaded that it will be seriously harmful both to the state and federal judiciaries.
The Court indicates that the new standard to be applied by federal judges in habeas corpus proceedings may be substantially the same as the standard most state reviewing courts are already applying. Ante, at 322. The federal district courts are therefore being directed simply to duplicate the reviewing function that is now being performed adequately by state appellate courts. In my view, this task may well be inconsistent with the prohibition?added by Congress to the federal habeas statute in order to forestall undue federal interference with state proceedings, see Wainwright v. Sykes, 433 U.S. 72, 80, 97 S.Ct. 2497, 2503, 53 L.Ed.2d 594?against overturning “a determination after a hearing on the merits of a factual issue, made by a State court of competent jurisdiction.” 28 U.S.C. ? 2254(d). See La Vallee v. Delle Rose, 410 U.S. 690, 93 S.Ct. 1203, 35 L.Ed.2d 637. In any case, to assign a single federal district judge the responsibility of directly reviewing, and inevitably supervising, the most routine work of the highest courts of a State can only undermine the morale and the esteem of the state judiciary particularly when the stated purpose of the additional layer of review is to determine whether the State’s factfinder is “rational.”9 Such consequences are intangible but nonetheless significant.
The potential effect on federal judges is even more serious. Their burdens are already so heavy that they are delegating to staff assistants more and more work that we once expected judges to perform.10 The new standard will invite an unknown number of state prisoners to make sufficiency challenges that they would not have made under the old rule. Moreover, because the “rational trier of fact” must certainly base its decisions on all of the evidence, the Court’s broader standard may well require that the entire transcript of the state trial be read whenever the factfinders’ rationality is challenged under the Court’s rule.11 Because this task will confront the courts of appeals as well as district courts, it will surely impose countless additional hours of unproductive labor on federal judges and their assistants.12 The increasing volume of work of this character has already led some of our most distinguished lawyers to discontinue or reject service on the federal bench.13 The addition of a significant volume of pointless labor can only impair the quality of justice administered by federal judges and thereby undermine “the respect and confidence of the community in applications of the . . . law.” In re Winship, 397 U.S., at 364, 90 S.Ct., at 1072.
For these reasons, I am unable to join the Court’s gratuitous directive to our colleagues on the federal bench.
The degrees of murder in Virginia are specified in Va.Code ? 18.2-32 (1975) as follows:
“Murder, other than capital murder, by poison, lying in wait, imprisonment, starving, or by any willful, deliberate, and premeditated killing, or in the commission of, or attempt to commit, arson, rape, robbery, burglary or abduction . . . is murder of the first degree, punishable as a Class 2 felony.
“All murder other than capital murder and murder in the first degree is murder of the second degree and is punishable as a Class 3 felony.”
Class 2 felonies carry a term of 20 years to life. ? 18.2-10(b) (1975). The sentence for Class 3 felonies can range from 5 to 20 years, ? 18.2-10(c). Murder itself takes its definition in Virginia from the common law. Stapleton v. Commonwealth, 123 Va. 825, 96 S.E. 801.
Under Virginia law, voluntary intoxication?although not an affirmative defense to second-degree murder?is material to the element of premeditation and may be found to have negated it. Hatcher v. Commonwealth, 218 Va. 811, 241 S.E.2d 756.
When trial without a jury is had on a not guilty plea in Virginia, the court is to “have and exercise all the powers, privileges and duties given to juries . . . .” Va.Code ? 19.2-257 (1975).
There is no appeal as of right from a criminal conviction in Virginia. Saunders v. Reynolds, 214 Va. 697, 204 S.E.2d 421. Each petition for writ of error under Va.Code ? 19.2-317 (1975) is reviewed on the merits, however, and the effect of a denial is to affirm the judgment of conviction on the merits. Saunders v. Reynolds, supra.
The petition for writ of error alleged that “the trial Court erred in finding the Petitioner guilty of first degree murder in light of the evidence introduced on behalf of the Commonwealth, and on unwarranted inferences drawn from this evidence.” The petitioner contended that an affirmance would violate the Due Process Clause of the Fourteenth Amendment. In its order denying Jackson’s petition, the Virginia Supreme Court stated it was “of [the] opinion that there is no reversible error in the judgment complained of . . . .” Virginia law requires sufficiency claims to be raised on direct appeal; such a claim may not be raised in a state habeas corpus proceeding. Pettus v. Peyton, 207 Va. 906, 153 S.E.2d 278.
The District Court correctly found that the petitioner had exhausted his state remedies on this issue. See n. 4, supra.
The opinions of the District Court and the Court of Appeals are not reported. The Court of Appeals’ judgment order is reported at 580 F.2d 1048.
The Court of Appeals in the present case, of course, recognized that Winship may have changed the constitutional standard in federal habeas corpus. And the Court of Appeals for the Sixth Circuit recently recognized the possible impact of Winship on federal habeas corpus in a case in which it held that “a rational trier of fact could have found the defendant . . . guilty beyond a reasonable doubt.” Spruytte v. Koehler, affirmance order, 590 F.2d 335. An even more recent case in that court provoked a lively debate among three of its members regarding the effect of Winship upon federal habeas corpus. The writ was granted in that case, even though the trial record concededly contained “some evidence” of the applicant’s guilt. See Speigner v. Jago, 603 F.2d 1208.
The trier of fact in this case was a judge and not a jury. But this is of no constitutional significance. The record makes clear that the judge deemed himself “properly instructed.”
A “reasonable doubt” has often been described as one “based on reason which arises from the evidence or lack of evidence.” Johnson v. Louisiana, 406 U.S. 356, 360, 92 S.Ct. 1620, 1624, 32 L.Ed.2d 152 (citing cases). For a discussion of variations in the definition used in jury instructions, see Holland v. United States, 348 U.S. 121, 140, 75 S.Ct. 127, 137, 99 L.Ed. 150 (rejecting contention that circumstantial evidence must exclude every hypothesis but that of guilt).
This, of course, does not mean that convictions are frequently reversed upon this ground. The practice in the federal courts of entertaining properly preserved challenges to evidentiary sufficiency, see Fed.Rule Crim.Proc. 29, serves only to highlight the traditional understanding in our system that the application of the beyond-a-reasonable-doubt standard to the evidence is not irretrievably committed to jury discretion. To be sure, the factfinder in a criminal case has traditionally been permitted to enter an unassailable but unreasonable verdict of “not guilty.” This is the logical corollary of the rule that there can be no appeal from a judgment of acquittal, even if the evidence of guilt is overwhelming. The power of the factfinder to err upon the side of mercy, however, has never been thought to include a power to enter an unreasonable verdict of guilty. Carpenters & Joiners v. United States, 330 U.S. 395, 408, 67 S.Ct. 775, 782, 91 L.Ed. 973. Cf. Capital Traction Co. v. Hof, 174 U.S. 1, 13-14, 19 S.Ct. 580, 585-586, 43 L.Ed. 873. Any such premise is wholly belied by the settled practice of testing evidentiary sufficiency through a motion for judgment of acquittal and a postverdict appeal from the denial of such a motion. See generally 4 L. Orfield, Criminal Procedure Under the Federal Rules ?? 29:1-29:29 (1967 and Supp.1978).
Until 1972, the Court of Appeals for the Second Circuit took the position advanced today by the opinion concurring in the judgment that the beyond-a-reasonable-doubt standard is merely descriptive of the state of mind required of the factfinder in a criminal case and not of the actual quantum and quality of proof necessary to support a criminal conviction. Thus, that court held that in a jury trial the judge need not distinguish between criminal and civil cases for the purpose of ruling on a motion for judgment of acquittal. United States v. Feinberg, 140 F.2d 592, 594. In United States v. Taylor, 464 F.2d 40 (CA2), Feinberg was overruled, partly on the strength of Winship. The Taylor court adopted the directed-verdict criterion articulated in Curley v. United States, 81 U.S.App.D.C. 389, 392-393, 160 F.2d 229, 232-233 (If “reasonable” jurors “must necessarily have . . . a reasonable doubt” as to guilt, the judge “must require acquittal, because no other result is permissible within the fixed bounds of jury consideration”). This is now the prevailing criterion for judging motions for acquittal in federal criminal trials. See generally 2 C. Wright, Federal Practice and Procedure ? 467 (1969 and Supp.1978).
Contrary to the suggestion in the opinion concurring in the judgment, the criterion announced today as the constitutional minimum required to enforce the due process right established in Winship is not novel. See, e. g., United States v. Amato, 495 F.2d 545, 549 (CA5) (“whether, taking the view [of the evidence] most favorable to the Government, a reasonably-minded jury could accept the relevant evidence as adequate and sufficient to support the conclusion of the defendant’s guilt beyond a reasonable doubt”) (emphasis added); United States v. Jorgenson, 451 F.2d 516, 521 (CA10) (whether, “considering the evidence in the light most favorable to the government, there is substantial evidence from which a jury might reasonably find that an accused is guilty beyond a reasonable doubt”) (emphasis added). Glasser v. United States, 315 U.S. 60, 80, 62 S.Ct. 457, 469, 86 L.Ed. 680, has universally been understood as a case applying this criterion. See, e. g., Harding v. United States, 337 F.2d 254, 256 (CA8). See generally, 4 Orfield, supra n.10, ? 29.28.
The question whether the evidence is constitutionally sufficient is of course wholly unrelated to the question of how rationally the verdict was actually reached. Just as the standard announced today does not permit a court to make its own subjective determination of guilt or innocence, it does not require scrutiny of the reasoning process actually used by the factfinder?if known. See generally 3 F. Wharton, Criminal Procedure ? 520 (12th ed. 1975 and Supp.1978).
Application of the Thompson standard to assess the validity of a criminal conviction after Winship could lead to absurdly unjust results. Our cases have indicated that failure to instruct a jury on the necessity of proof of guilt beyond a reasonable doubt can never be harmless error. See Cool v. United States, 409 U.S. 100, 93 S.Ct. 354, 34 L.Ed.2d 335. Cf. Taylor v. Kentucky, 436 U.S. 478, 98 S.Ct. 1930, 56 L.Ed.2d 468. Thus, a defendant whose guilt was actually proved by overwhelming evidence would be denied due process if the jury was instructed that he could be found guilty on a mere preponderance of the evidence. Yet a defendant against whom there was but one slender bit of evidence would not be denied due process so long as the jury has been properly instructed on the prosecution’s burden of proof beyond a reasonable doubt. Such results would be wholly faithless to the constitutional rationale of Winship.
The Virginia Supreme Court’s order denying Jackson’s petition for writ of error does not make clear what criterion was applied to the petitioner’s claim that the evidence in support of his first-degree murder conviction was insufficient. See n. 4, supra. At oral argument, counsel for the petitioner contended that the Virginia sufficiency standard is not keyed to Winship. Counsel for the State disagreed. Under these circumstances, we decline to speculate as to the criterion that the state court applied. The fact that a state appellate court invoked the proper standard, however, although entitled to great weight, does not totally bar a properly presented claim of this type under ? 2254.
The respondents have suggested that this constitutional standard will invite intrusions upon the power of the States to define criminal offenses. Quite to the contrary, the standard must be applied with explicit reference to the substantive elements of the criminal offense as defined by state law. Whether the State could constitutionally make the conduct at issue criminal at all is, of course, a distinct question. See Papachristou v. Jacksonville, 405 U.S. 156, 92 S.Ct. 839, 31 L.Ed.2d 110; Robinson v. California, 370 U.S. 660, 82 S.Ct. 1417, 8 L.Ed.2d 758.
In United States v. Tarr, 589 F.2d 55 (CA1 1978), the court overturned one of two counts of which appellant was convicted because there was insufficient evidence to prove that he had the intent to aid and abet the unauthorized transfer of a machinegun in violation of 26 U.S.C. ? 5861(e) and 18 U.S.C. ? 2. The court found “no evidence” that appellant had the requisite knowledge. 589 F.2d, at 60.
In United States v. Whetzel, 191 U.S.App.D.C. 184, 589 F.2d 707 (1978), the court overturned 2 of the 35 counts of appellant’s conviction because “the Government failed to offer proof that would permit a jury to reasonably infer that the merchandise [appellant] transported had a value of $5,000.” Id., 191 U.S.App.D.C., at 188, 589 F.2d, at 711. However, the basis for this determination was that the Government’s valuation method, which the trial court allowed the jury to consider, was legally erroneous. Similarly, in United States v. Fearn, 589 F.2d 1316 (CA7 1978), the court overturned the conviction based on a federal nonconstitutional rule, which surely would not apply in habeas review of state convictions, “that a conviction must rest upon firmer ground than the uncorroborated admission or confession of the accused.” Id., at 1321. The court did not independently analyze whether the uncorroborated confession involved in that case could itself have allowed a rational trier of fact to find guilt beyond a reasonable doubt.
In In re Winship, 397 U.S., at 364, 90 S.Ct., at 1072, the Court stated: “As we said in Speiser v. Randall, [357 U.S. 513,] 525-526, 78 S.Ct. , at 1342 [2 L.Ed.2d 1460]: ‘There is always in litigation a margin of error, representing error in factfinding, which both parties must take into account. Where one party has at stake an interest of transcending value?as a criminal defendant his liberty?this margin of error is reduced as to him by the process of placing on the other party the burden of . . . persuading the factfinder at the conclusion of the trial of his guilt beyond a reasonable doubt. Due process commands that no man shall lose his liberty unless the Government has borne the burden of . . . convincing the factfinder of his guilt.’ To this end, the reasonable-doubt standard is indispensable, for it ‘impresses on the trier of fact the necessity of reaching a subjective state of certitude of the facts in issue.’ Dorsen & Rezneck, In Re Gault and the Future of Juvenile Law, 1 Family Law Quarterly, No. 4, pp. 1, 26 (1967).” (Emphasis added.)
Later on the same page, 90 S.Ct., at 1073, the Court added:
“It is also important in our free society that every individual going about his ordinary affairs have confidence that his government cannot adjudge him guilty of a criminal offense without convincing a proper factfinder of his guilt with utmost certainty.” Ibid. (emphasis added).
See also id., at 370, 90 S.Ct., at 1076 (Harlan, J., concurring) (“[A] standard of proof represents an attempt to instruct the factfinder concerning the degree of confidence our society thinks he should have in the correctness of factual conclusions for a particular type of adjudication”). (emphasis added).
The Court, relying on treatises that analyzed the law in all 50 States as well as in the federal system, determined both that the reasonable-doubt standard has prevailed at the trial level “at least from our early years as a Nation” and that it “is now accepted in common law jurisdictions as the measure of persuasion by which the prosecution must convince the trier of all the essential elements of guilt.” Id., at 361, 90 S.Ct., at 1071 (emphasis added). See also id., at 372, 90 S.Ct., at 1077 (Harlan, J., concurring). (“It is only because of the nearly complete and long-standing acceptance of the reasonable-doubt standard by the States in criminal trials that the Court has not before today had to hold explicitly that due process, as an expression of fundamental procedural fairness, requires a more stringent standard for criminal trials than for ordinary civil litigation”) (emphasis added).
The Court has undertaken no systematic analysis of the standards for reviewing the sufficiency of the evidence that prevail either in state habeas corpus and other collateral proceedings or in state appellate courts. What sources I have discovered suggest that “varied standards” are in use and that each is “subject to shifting and elastic definitions.” Winningham, The Dilemma of the Directed Acquittal, 15 Vand.L.Rev. 699, 705-706 (1962). See ALI Code of Criminal Procedure, Commentary on ? 321, pp. 961-962 (1930); Rules of Criminal Procedure 481(c), 522(a) and commentary, 10 U.L.A. (1974).
It hardly bears repeating that habeas corpus is not intended as a substitute for appeal, nor as a device for reviewing the merits of guilt determinations at criminal trials. See generally Stone v. Powell, 428 U.S. 465, 96 S.Ct. 3037, 49 L.Ed.2d 1067. Instead, it is designed to guard against extreme malfunctions in the state criminal justice systems.
As I discussed earlier, see supra, at 329, the incidence of factual error at the trial level in federal courts appears to be exceedingly low, even when measured by the relatively strict appellate standard used by the Federal Courts of Appeals. Presumably the incidence of errors that survive that first level of review is even smaller.
Indeed, the Court makes light of Winship by suggesting that, in the absence of its new habeas procedure, the result of that case is simply “a trial ritual.” Ante, at 316-317. Far more likely in my view is that the Court’s difficult-to-apply but largely unnecessary rule will itself result in a “collateral-attack ritual” that will undermine the integrity of both the state and federal judiciaries. See infra, at 336-339.
So far as I can determine, this standard first appeared in our jurisprudence in Mr. Justice STEWART’s opinion dissenting from the Court’s denial of certiorari in Freeman v. Zahradnick, 429 U.S. 1111, 1112, 1113, 1114, 1116, 97 S.Ct. 1150, 1151, 1152, 1153, 51 L.Ed.2d 566. At that time, it gave the impression of being somewhat narrower than?if only because it was stated quite differently from?the test used by the Courts of Appeals in reviewing federal convictions on direct appeal. See Curley v. United States, 81 U.S.App.D.C. 389, 392-393, 160 F.2d 229, 232-233 (1947). Although the Court twice repeats the Freeman test, see ante, at 313, 319, it now appears either to equate that standard with the?in my view?broader federal direct-review standard, or to endorse both standards despite their differences. See ante, at 318, and nn. 11-12.
In the past, collateral review of state proceedings has been justified largely on the grounds (1) that federal judges have special expertise in the federal issues that regularly arise in habeas corpus proceeding, and (2) that they are less susceptible than state judges to political pressures against applying constitutional rules to overturn convictions. See, e. g., Bartels, Avoiding a Comity of Errors, 29 Stan.L.Rev. 27, 30 n. 9 (1976). Cf. Steffel v. Thompson, 415 U.S. 452, 464, 94 S.Ct. 1209, 1218, 39 L.Ed.2d 505; Mitchum v. Foster, 407 U.S. 225, 242, 92 S.Ct. 2151, 2162, 32 L.Ed.2d 705. But neither of these justifications has any force in the present context. State judges are more familiar with the elements of state offenses than are federal judges and should be better able to evaluate sufficiency claims. Moreover, of all decisions overturning convictions, the least likely to be unpopular and thus to distort state decisionmaking processes are ones based on the inadequacy of the evidence. Indeed, once federal courts were divested of authority to second-guess state courts on Fourth Amendment issues, which are far more likely to generate politically motivated state-court decisions, see Stone v. Powell, 428 U.S. 465, 96 S.Ct. 3037, 49 L.Ed.2d 1067, a like result in this case would seem to be a fortiori.
For example, the heavy federal workload has required the 13 regular and 7 senior judges on the Ninth Circuit to hire 30 staff attorneys and 33 law clerks to assist them in their labors.
Additional burdens will also be imposed if the Court’s rule is extended to federal habeas proceedings reviewing federal criminal trials, as well as to ones reviewing state civil commitment proceedings in which we have recently required at least the “clear and convincing” test to be applied as a matter of federal constitutional law. Addington v. Texas, 441 U.S. 418, 99 S.Ct. 1804, 60 L.Ed.2d 323.
This Court’s workload will also increase, of course, when its certiorari docket expands to accommodate the challenges generated by the Court’s new rule. The effect will be even greater if the Court’s opinion is read to require state appellate courts to apply the reasonable-doubt test on direct review and to require this Court to apply it when reviewing the decisions of those courts on certiorari.
Professor Bator has persuasively explained how the law of diminishing returns inevitably makes it unwise to have duplicative review processes on the “merits” in criminal cases:
“[I]f a criminal judgment is ever to be final, the notion of legality must at some point include the assignment of final competences to determine legality. But, it may be asked, why should we seek a point at which such a judgment becomes final? Conceding that no process can assure ultimate truth, will not repetition of inquiry stand a better chance of approximat-
ing it? In view of the awesomeness of the consequences of conviction, shouldn’t we allow redetermination of the merits in an attempt to make sure that no error has occurred?
“Surely the answer runs, in the first place, in terms of conservation of resources?and I mean not only simple economic resources, but all of the intellectual, moral, and political resources involved in the legal system. The presumption must be, it seems to me, that if a job can be well done once, it should not be done twice. If one set of institutions is as capable of performing the task at hand as another, we should not ask both to do it. The challenge really runs the other way: if a proceeding is held to determine the facts and law in a case, and the processes used in that proceeding are fitted to the task in a manner not inferior to those which would be used in a second proceeding, so that one cannot demonstrate that relitigation would not merely consist of repetition and second-guessing, why should not the first proceeding ‘count’? Why should we duplicate effort? After all, it is the very purpose of the first go-around to decide the case. Neither it nor any subsequent go-around can assure ultimate truth. If, then, the previous determination is to be ignored, we must have some reasoned institutional justification why this should be so.
“Mere iteration of process can do other kinds of damage. I could imagine nothing more subversive of a judge’s sense of responsibility, of the inner subjective conscientiousness which is so essential a part of the difficult and subtle art of judging well, than an indiscriminate acceptance of the notion that all the shots will always be called by someone else. Of course this does not mean that we should not have appeals. As we shall see, important functional and ethical purposes are served by allowing recourse to an appellate court in a unitary system, and to a federal supreme court in a federal system. The acute question is the effect it will have on a trial judge if we then allow still further recourse where these purposes may no longer be relevant. What seems so objectionable is second-guessing merely for the sake of second-guessing, in the service of the illusory notion that if we only try hard enough we will find the ‘truth.’ ” Bator, Finality in Criminal Law and Federal Habeas Corpus for State Prisoners, 76 Harv.L.Rev. 441, 450-451 (1963).
See also F. James, Civil Procedure 518 (1965).
The testimony of Griffin Bell at his confirmation hearings for Attorney General is particularly relevant. When asked by Senator Scott of Virginia why he had earlier resigned from his seat on the Court of Appeals for the Fifth Circuit, Judge Bell responded:
“I found it not to be a rewarding experience any longer. Whether it was because there was no more excitement after the 1960’s, or whether it was because the case load changed, but the work load was oppressive. I would not have minded the work load, but the character of the cases changed. It was almost like serving on a criminal court. I did not want to do that any longer.” Hearings on the Prospective Nomination of Griffin B. Bell, of Georgia, to be Attorney General, before the Senate Committee on the Judiciary, 95th Cong., 1st Sess., 27 (1977).