announced the judgment of the Court and delivered an opinion,
in which The CHIEF Justice and Justice Scalia joined.In this case, we must determine whether the Court of Appeals for the Fourth Circuit correctly applied our decision in Jackson v. Virginia, 443 U. S. 307 (1979), in concluding that the evidence against respondent Frank West was insufficient, as a matter of due process, to support his state-court conviction for grand larceny.
I
Between December 13 and December 26, 1978, someone broke into the Westmoreland County, Virginia, home of Angelo Cardova and stole items valued at approximately $3,500. On January 10,1979, police conducted a lawful search of the Gloucester County, Virginia, home of West and his wife. They discovered several of the items stolen from the Car-dova home, including various electronic equipment (two television sets and a record player); articles of clothing (an imitation mink coat with the name “Esther” embroidered in it, a silk jacket emblazoned “Korea 1970,” and a pair of shoes); decorations (several wood earvings and a mounted lobster); and miscellaneous household objects (a mirror framed with seashells, a coffee table, a bar, a sleeping bag, and some silverware). These items were valued at approximately $800, and the police recovered other, unspecified items of Cardo-va’s property with an approximate value of $300.
West was charged with grand larceny. Testifying at trial on his own behalf, he admitted to a prior felony conviction, but denied having taken anything from Cardova’s house. *281He explained that he had bought and sold “a lot of. . . merchandise” from “several guys” at “flea bargain places” where, according to West, “a lot of times you buy things ... that are stolen” although “you néver know it.” App. 21. On cross-examination, West said that he had bought many of the stolen items from a Ronnie Elkins, whom West claimed to have known for years. West testified that he purchased one of the wood carvings, the jacket, mounted lobster, mirror, and bar from Elkins for about $500. West initially guessed, and then twice positively asserted, that this sale occurred before January 1, 1979. In addition, West claimed to have purchased the coat from Elkins for $5 around January 1, 1979. His testimony did not make clear whether he was describing one transaction or two, whether there were any other transactions between himself and Elkins, where the transactions occurred, and whether the transactions occurred at flea markets.1 West testified further that he had purchased one of *282the television sets in an entirely separate transaction in Goochland County, from an individual whose name he had forgotten. Finally, West testified that he did not remember how he had acquired the second television, the coffee table, and the silverware.
Under then-applicable Virginia law, grand larceny was defined as the wrongful and nonconsensual taking of property worth at least $100, with the intent to deprive the owner of it permanently. See Va. Code Ann. § 18.2-95 (1975); Skeeter v. Commonwealth, 217 Va. 722, 725, 232 S. E. 2d 756, 758 (1977). Virginia law permits an inference that a person who fails to explain, or falsely explains, his exclusive possession of recently stolen property is the thief. See, e. g., Moehring v. Commonwealth, 223 Va. 564, 568, 290 S. E. 2d 891, 893 (1982); Best v. Commonwealth, 222 Va. 387, 389, 282 S. E. 2d 16, 17 (1981). The trial court instructed the jurors about this permissive inference, but warned that the inference did not compromise their constitutional obligation to acquit unless they found that the State had established every element *283of the crime beyond a reasonable doubt. See In re Winship, 397 U. S. 358 (1970).2
The jury returned a guilty verdict, and West received a 10-year prison sentence. West petitioned for an appeal, contending (among other things) that the evidence was insufficient to support a finding of guilt beyond a reasonable doubt. In May 1980, the Supreme Court of Virginia refused the petition — a disposition indicating that the court found the petition without merit, see Saunders v. Reynolds, 214 Va. 697, 700, 204 S. E. 2d 421, 424 (1974). Seven years later, West filed a petition for a writ of habeas corpus in the same court, supported by an affidavit executed by Ronnie Elkins in April 1987. West renewed his claim that the original trial record contained insufficient evidence to support the conviction, and he argued in the alternative that Elkins’ affidavit, which tended to corroborate West’s trial testimony in certain respects, constituted new evidence entitling him to a new trial. The Supreme Court of Virginia again denied relief. West then filed a petition for a writ of habeas corpus in the District Court for the Eastern District of Virginia, which rejected both claims and denied relief.
The Court of Appeals for the Fourth Circuit reversed. 931 F. 2d 262 (1991). As the court correctly recognized, a *284claim that evidence is insufficient to support a conviction as a matter of due process depends on "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.” Jackson v. Virginia, 443 U. S., at 319 (emphasis in original). Five considerations led the court to conclude that this standard was not met: first, the items were recovered no sooner than two weeks after they had been stolen; second, only about a third of the items stolen from Cardova (measured by value) were recovered from West; third, the items were found in West’s house in plain view, and not hidden away as contraband; fourth, West’s explanation of his possession was not so “inherently implausible,” even if it were disbelieved, that it could “fairly be treated as positive evidence of guilt”; and fifth, there was no corroborating evidence (such as fingerprints or eyewitness testimony) beyond the fact of mere possession. See 931 F. 2d, at 268-270. The court viewed West’s testimony as “at most, a neutral factor,” id., at 270, despite noting his “confusion” about the details of his alleged purchases, id., at 269, and despite conceding that his testimony “at first blush . .. may itself seem incredible,” id., at 270, n. 7. In holding that the Jackson standard was not met, the court did not take into consideration the fact that the Supreme Court of Virginia had twice previously concluded otherwise.
After the Fourth Circuit denied rehearing en banc by an equally divided court, see App. to Pet. for Cert. 34-35, the warden and the State Attorney General sought review in this Court on, among other questions, whether the Court of Appeals had applied Jackson correctly in this case. We granted certiorari, 502 U. S. 1012 (1991), and requested additional briefing on the question whether a federal habeas court should afford deference to state-court determinations applying law to the specific facts of a case, 502 U. S. 1021 (1991). We now reverse.
*285II
The habeas corpus statute permits a federal court to entertain a petition from a state prisoner “only on the ground that he is in custody in violation of the Constitution or laws or treaties of the United States.” 28 U. S. C. § 2254(a). The court must “dispose of the matter as law and justice require.” § 2243. For much of our history, we interpreted these bare guidelines and their predecessors to reflect the common-law principle that a prisoner seeking a writ of habeas corpus could challenge only the jurisdiction of the court that had rendered the judgment under which he was in custody. See, e.g., In re Wood, 140 U. S. 278, 285-287 (1891) (Harlan, J.); Ex parte Watkins, 3 Pet. 193, 202 (1830) (Marshall, C. J.). Gradually, we began to expand the category of claims deemed to be jurisdictional for habeas purposes. See, e. g., Ex parte Siebold, 100 U. S. 371, 377 (1880) (court without jurisdiction to impose sentence under unconstitutional statute); Ex parte Lange, 18 Wall. 163, 176 (1874) (court without jurisdiction to impose sentence not authorized by statute). Next, we began to recognize federal claims by state prisoners if no state court had provided a full and fair opportunity to litigate those claims. See, e. g., Moore v. Dempsey, 261 U. S. 86, 91-92 (1923); Frank v. Mangum, 237 U. S. 309, 335-336 (1915). Before 1953, however, the inverse of this rule also remained true: Absent an alleged jurisdictional defect, “habeas corpus would not lie for a [state] prisoner ... if he had been given an adequate opportunity to obtain full and fair consideration of his federal claim in the state courts.” Fay v. Noia, 372 U. S. 391, 459-460 (1963) (Harlan, J., dissenting). See generally Bator, Finality in Criminal Law and Federal Habeas Corpus for State Prisoners, 76 Harv. L. Rev. 441, 478-499 (1963). In other words, the state-court judgment was entitled to “absolute respect,” Kuhlmann v. Wil*286son, 477 U. S. 436, 446 (1986) (opinion of Powell, J.) (emphasis added), and a federal habeas court could not review it even for reasonableness.3
*287We rejected the principle of absolute deference in our landmark decision in Brown v. Allen, 344 U. S. 443 (1953). There, we held that a state-court judgment of conviction “is not res judicata” on federal habeas with respect to federal constitutional claims, id., at 458, even if the state court has rejected all such claims after a full and fair hearing. Instead, we held, a district court must determine whether the state-court adjudication “has resulted in a satisfactory conclusion.” Id., at 463. We had no occasion to explore in detail the question whether a “satisfactory” conclusion was one that the habeas court considered correct, as opposed to merely reasonable, because we concluded that the constitutional claims advanced in Brown itself would fail even if the state courts’ rejection of them were reconsidered de novo. See id., at 465-476. Nonetheless, we indicated that the federal courts enjoy at least the discretion to take into consideration the fact that a state court has previously rejected the federal claims asserted on habeas. See id., at 465 (“As the state and federal courts have the same responsibilities to protect persons from violation of their constitutional rights, we conclude that a federal district court may decline, without a rehearing of the facts, to award a writ of habeas corpus to a state prisoner where the legality of such detention has been *288determined, on the facts presented, by the highest state court with jurisdiction”).4
In an influential separate opinion endorsed by a majority of the Court, Justice Frankfurter also rejected the principle of absolute deference to fairly litigated state-court judgments. He emphasized that a state-court determination of federal constitutional law is not “binding” on federal habeas, id., at 506, regardless of whether the determination involves a pure question of law, ibid., or a “so-called mixed questio[n]” requiring the application of law to fact, id., at 507. Nonetheless, he stated quite explicitly that a “prior State determination may guide [the] discretion [of the district court] in deciding upon the appropriate course to be followed in disposing of the application.” Id., at 500. Discussing mixed questions specifically, he noted further that “there is no need for the federal judge, if he could, to shut his eyes to the State consideration.” Id., at 508.5
*289In subsequent eases, we repeatedly reaffirmed Brown’s teaching that mixed constitutional questions are “open to review on collateral attack,” Cuyler v. Sullivan, 446 U. S. 335, 342 (1980), without ever explicitly considering whether that “review” should be de novo or deferential. In some of these eases, we would have denied habeas relief even under de novo review, see, e. g., Strickland v. Washington, 466 U. S. 668, 698 (1984) (facts make it “clear” that habeas petitioner did not receive ineffective assistance of counsel); Neil v. Biggers, 409 U. S. 188, 201 (1972) (facts disclose “no substantial likelihood” that habeas petitioner was subjected to unreliable pretrial lineup); in others, we would have awarded habeas relief even under deferential review, see, e. g., Brewer v. Williams, 430 U. S. 387, 405 (1977) (facts provide “no reasonable basis” for finding valid waiver of right to counsel); Irvin v. Dowd, 366 U. S. 717, 725 (1961) (facts show “clear and convincing” evidence of biased jury); and in yet others, we remanded for application of a proper legal rule without addressing that standard of review question, see, e, g., Cuyler, supra, at 342, 350. Nonetheless, because these cases never qualified our early citation of Brown for the proposition that a federal habeas court must reexamine mixed constitutional questions “independently,” Townsend v. Sain, 372 U. S. 293, 318 (1963) (dictum), we have gradually come to treat as settled the rule that mixed constitutional questions are “subject to plenary federal review” on habeas, Miller v. Fenton, 474 U. S. 104, 112 (1985).6 .
*290Jackson itself contributed to this trend. There, we held that a conviction violates due process if supported only by evidence from which “no rational trier of fact could find guilt beyond a reasonable doubt.” 443 U. S., at 317. We stated explicitly that a state-court judgment applying the Jackson rule in a particular case “is of course entitled to deference” on federal habeas. Id., at 323; see also id., at 336, n. 9 (Stevens, J., concurring in judgment) (“State judges are more familiar with the elements of state offenses than are federal judges and should be better able to evaluate sufficiency claims”). Notwithstanding these principles, however, we then indicated that the habeas court itself should apply the Jackson rule, see id., at 324, rather than merely reviewing the state courts’ application of it for reasonableness. Ultimately, though, we hád no occasion to resolve our conflicting statements on the standard of review question, because we concluded that the habeas petitioner was not entitled, to relief even under our own de novo application of Jackson. See id., at 324-326.7
*291Despite our apparent adherence to a standard of de novo habeas review with respect to mixed constitutional questions, we have implicitly questioned that standard, at least with respect to pure legal questions, in our recent retroactivity precedents. In Penry v. Lynaugh, 492 U. S. 302, 313-314 (1989), a majority of this Court endorsed the retroactivity analysis advanced by Justice O’Connor for a plurality in Teague v. Lane, 489 U. S. 288 (1989). Under Teague, a ha-beas petitioner generally cannot benefit from a new rule of criminal procedure announced after his conviction has become final on direct appeal. See id., at 305-310 (opinion of O’Connor, J.). Teague defined a “new” rule as one that was “not dictated by precedent existing at the time the defendant’s conviction became final.” Id., at 301 (emphasis in origi- , nal). In Butler v. McKellar, 494 U. S. 407, 415 (1990), we explained that the definition includes all rules “susceptible to debate among reasonable minds.” Thus, if a state court has reasonably rejected the legal claim asserted by a habeas petitioner under existing law, then the claim seeks the benefit of a “new” rule under Butler, and is therefore not cognizable on habeas under Teague. In other words, a federal ha-beas court “must defer to the state court’s decision rejecting the claim unless that decision is patently unreasonable.” Butler, supra, at 422 (Brennan, J., dissenting).8
*292Teague was premised on the view that retroactivity questions in habeas corpus proceedings must take account of the nature and function of the writ, which we describéd as “ ‘a collateral remedy... not designed as a substitute for direct review/” 489 U. S., at 306 (opinion of O’Connor, J.) (quoting Mackey v. United States, 401 U. S. 667, 682-683 (1971) (Harlan, J., concurring in judgments in part and dissenting in part)) (emphasis in Mackey). Justice Stevens reasoned similarly in Jackson, where he stressed 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,” but only “to guard against extreme malfunctions in the state criminal justice systems.” 443 U. S., at 332, n. 5 (opinion concurring in judgment); see also Greer v. Miller, 483 U. S. 756, 768-769 (1987) (Stevens, J., concurring in judgment). Indeed, the notion that different standards should apply on direct and collateral review runs throughout our recent habeas jurisprudence. We have said, for example, that new rules always have retroactive application to *293criminal cases pending on direct review, see Griffith v. Kentucky, 479 U. S. 314, 320-328 (1987), but that they generally do not have retroactive application to criminal eases pending on habeas, see Teague, supra, at 305-310 (opinion of O’Con-nor, J.)- We have held that the Constitution guarantees the right to counsel on a first direct appeal, see, e. g., Douglas v. California, 372 U. S. 353, 355-358 (1963), but that it guarantees no right to counsel on habeas, see, e. g., Pennsylvania v. Finley, 481 U. S. 551, 555 (1987). On direct review, we have announced and enforced the rule that state courts must exclude evidence obtained in violation of the Fourth Amendment. See, e. g., Mapp v. Ohio, 367 U. S. 643, 654-660 (1961). We have also held, however, that claims under Mapp are not cognizable on habeas as long as the state courts have provided a full and fair opportunity to litigate them at trial or on direct review. See Stone v. Powell, 428 U. S. 465, 489-496 (1976).
These differences simply reflect the fact that habeas review “entails significant costs.” Engle v. Isaac, 456 U. S. 107, 126 (1982). Among other things, “‘[i]t disturbs the State’s significant interest in repose for concluded litigation, denies society the right to punish some admitted offenders, and intrudes on state sovereignty to a degree matched by few exercises of federal judicial authority.’ ” Duckworth v. Eagan, 492 U. S. 195, 210 (1989) (O’Connor, J., concurring) (quoting Harris v. Reed, 489 U. S. 255, 282 (1989) (Kennedy, J., dissenting)). In various contexts, we have emphasized that these costs, as well as the countervailing benefits, must be taken into consideration in defining the scope of the writ. See, e. g., Coleman v. Thompson, 501 U.S. 722, 738-739 (1991) (procedural default); McCleskey v. Zant, 499 U. S. 467, 490-493 (1991) (abuse of the writ); Teague, supra, at 308-310 (opinion of O’Connor, J.) (retroactivity); Kuhlmann v. Wilson, 477 U. S., at 444-455 (opinion of Powell, J.) (successive petitions); Stone v. Powell, supra, at 491-492, n. 31 (cogniza-bility of particular claims).
*294In light of these principles, petitioners ask that we reconsider our statement in Miller v. Fenton that mixed constitutional questions are “subject to plenary federal review” on habeas, 474 U. S., at 112. By its terms, Teague itself is not directly controlling, because West sought federal habeas relief under Jackson, which was decided a year before his conviction became final on direct review. Nonetheless, petitioners contend, the logic of Teague makes our statement in Miller untenable. Petitioners argue that if deferential review for reasonableness strikes an appropriate balance with respect to purely legal claims, then it must strike an appropriate balance with respect to mixed questions as well. Moreover, they note that under the habeas statute itself, a state-court determination of a purely factual question must be “presumed correct,” and can be overcome only by “convincing evidence,” unless one of eight statutorily enumerated exceptions is present. 28 U. S. C. § 2254(d). It makes no sense, petitioners assert, for a habeas court generally to review factual determinations and legal determinations deferentially, but to review applications of law to fact de novo. Finally, petitioners find the prospect of deferential review for mixed questions at least implicit in our recent statement that Teague concerns are fully implicated “by the application of an old rule in a manner that was not dictated by precedent.” Stringer v. Black, 503 U. S. 222, 228 (1992) (emphasis added). For these reasons, petitioners invite us to reaffirm that a habeas judge need not — and indeed may not — “shut his eyes” entirely to state-court applications of law to fact. Brown v. Allen, 344 U. S., at 508 (opinion of Frankfurter, J.). West develops two principal counterarguments: first, that Congress implicitly codified a de novo standard with respect to mixed constitutional questions when it amended the habeas statute in 1966; and second, that *295de novo federal review is necessary to vindicate federal constitutional rights.9
We need not decide such far-reaching issues in this case. As in both Brown and Jackson, the claim advanced by the habeas petitioner must fail even assuming that the state court's rejection of it should be reconsidered de novo. Whatever the appropriate standard of review, we conclude that there was more than enough evidence to support West’s conviction.
The ease against West was strong. Two to four weeks after the Cardova home had been burglarized, over 15 of the items stolen were recovered from West's home. On direct examination at trial, West said nothing more than that he frequently bought and sold items at different flea markets. He failed to offer specific information about how he had come ' to acquire any of the stolen items, and he did not even mention Ronnie Elkins. by name. When pressed on cross-examination about the details of his purchases, West contradicted himself repeatedly about where he supposedly had bought the stolen goods, and he gave vague, seemingly eva*296sive answers to various other questions. See n. 1, supra. He said further that he could not remember how he had acquired such major household items as a television set and a coffee table, and -he failed to offer any explanation whatsoever about how he had acquired Cardova’s record player, among other things. Moreover, he testified that he had acquired Cardova’s second television set from a seller other than Elkins (who remained unidentified) in an entirely unrelated (but roughly contemporaneous) transaction. Finally, he failed to produce any other supporting evidence, such as testimony from Elkins, whom he claimed to have known for years and done business with on a regular basis.
As the trier of fact, the jury was entitled to disbelieve West’s uncorroborated and eonfüsed testimony. In evaluating that testimony, moreover, the jury was entitled to discount West’s credibility on account of his prior felony conviction, see Va. Code Ann. §19.2-269 (1990); Sadoski v. Commonwealth, 219 Va. 1069, 254 S. E. 2d 100 (1979), and to take into account West’s demeanor when testifying, which neither the Court of Appeals nor we may review. And if the jury did disbelieve West, it was further entitled to consider whatever it concluded to be perjured testimony as affirmative evidence of guilt, see, e. g., Wilson v. United States, 162 U. S. 613, 620-621 (1896); United States v. Zafiro, 945 F. 2d 881, 888 (CA7 1991) (Posner, J.), cert. granted on other grounds, 503 U. S. 935 (1992); Dyer v. MacDougall, 201 F. 2d 265, 269 (CA2 1952) (L. Hand, J.).
In Jackson, we emphasized repeatedly the deference owed to the trier of fact and, correspondingly, the sharply limited nature of constitutional sufficiency review. We said that “all of the evidence is to be considered in the light most favorable to the prosecution,” 443 U. S., at 319 (emphasis in original); that the prosecution need not affirmatively “rule out every hypothesis except that of guilt,” id., at 326; and that a reviewing court “faced with a record of historical facts that supports conflicting inferences must presume — even if *297it 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,” ibid. Under these standards, we think it clear that the trial record contained sufficient evidence to support West’s conviction.
Having granted relief on West’s Jackson claim, the Court of Appeals declined to address West’s additional claim that he was entitled to a new trial, as a matter of due process, on the basis of newly discovered evidence. See 931 F. 2d, at 271, n. 9. As that claim is not properly before us, we decline to address it here. The judgment of the Court of Appeals is reversed, and the case is remanded for further proceedings consistent with this opinion.
It is so ordered.
The quality of West’s testimony on these matters can best be appreciated by example:
“Q Axe those items that you bought at a flea market?
“A Well, I didn’t buy these items at a flea market, no sir.
“Q Whose items are they?
“A They are some items that I got from a Ronnie Elkins.
“Q All of the items you bought from him?
“A I can’t say all.
"Q Which ones did you buy from him?
“A I can’t say, because I don't have an inventory.
“Q Can you tell me the ones you bought from Ronnie Elkins?
“A Yes, I am sure I can.
“Q Which ones?
“A I would say the platter.
“Q How about the sea shell mirror?
“A Yes, sir, I think so.
“Q Where did you buy that?
“A In Newport News at a flea market.” App. 21-22.
“Q I want to know about your business transactions with Ronnie Elkins..
*282“A I buy and sell different items from different individuals at flea markets.
“Q Tell us where that market is.
“A In Richmond. You have them in Gloucester.
“Q Where is Ronnie Elkins’ flea market?
“A He does not have one.
“Q Didn’t you say you bought some items from Ronnie Elkins?
“A At a flea market.
“Q Tell the jury where that is at [sic].
"A In Gloucester.
“Q Tell the jury about this flea market and Ronnie Elkins, some time around January 1, and these items, not the other items.
"A Ronnie Elkins does not own a flea market.
“Q Tell the jury, if you will, where Ronnie Elkins was on the day that you bought the items?
"A I don’t remember. It was before January 1.
“Q Where was it?
“A I bought stuff from him in Richmond, Gloucester, and Newport News.” Id., at 26-27.
The instruction on the permissive inference read:
“If you belie[ve] from the evidence beyond a reasonable doubt that property of a value of $100.00 or more was stolen from Angelo F. C[a]rdova, and that it was recently thereafter found in the exclusive and personal possession of the defendant, and that such possession has been unexplained or falsely denied by the defendant, then such possession is sufficient to raise an inference that the defendant was the thief; and if such inference, taking into consideration the whole evidence, leads you to believe beyond a reasonable doubt that the defendant committed the theft, then you shall find the defendant guilty.” App. 34.
Several other instructions emphasized that despite the permissive inference, “[t]he burden is upon the Commonwealth to prove by the evidence beyond a reasonable doubt every material and necessary element of the offense charged against the defendant.” Ibid.
Justice O’Connor offers three criticisms of our summary of the history of habeas corpus before 1963, none of which we find convincing. First, she contends that the full-and-fair litigation standard in Frank v. Mangum, 237 U. S. 309 (1915), and Moore v. Dempsey, 261 U. S. 86 (1923), served no purpose other than to define the scope of the underlying alleged constitutional violation. See post, at 297-299. Frank and Moore involved claims, rejected by the state appellate courts, that a trial had been so dominated by a mob as to violate due process. In Frank, we denied relief not because the state appellate court had decided the federal claim correctly (the relevant question on direct review), and not even because the state appellate court had decided the federal claim reasonably, but only “because Frank’s federal claims had been considered by a competent and unbiased state tribunal,” Stone v. Powell, 428 U. S. 465, 476 (1976). In Moore, which reaffirmed Frank expressly, see 261 U. S., at 90-91, we ordered the District Court to consider the mob domination claim on the merits because the state appellate court's “perfunctory treatment” of it “was not in fact acceptable corrective process.” Noia, 372 U. S., at 458 (Harlan, J., dissenting); see also Bator, 76 Harv. L. Rev., at 488-489. In both cases, a claim that the habeas petitioner had been denied due process at trial was not cognizable on habeas unless the petitioner also had been denied a full and fair opportunity to raise that claim on appeal.
Second, Justice O’Connor states that we mischaracterize the views of Justice Powell about the history of habeas law between 1915 and 1953. See post, at 299. In fact, however, Justice Powell has often recounted exactly the same familiar history that we summarize above. In Rose v. Mitchell, 443 U. S. 545 (1979), for example, he described Frank as having “modestly expanded” the “scope of the writ” in order to “encompass those cases where the defendant’s federal constitutional claims had not been considered in the state-court proceeding.” 443 U. S., at 580 (opinion concurring in judgment). Similarly, in Schneckloth v. Bustamante, 412 U. S. 218 (1973), he described Frank as having extended “[t]he scope of federal ha-beas corpus” to permit consideration of “whether the applicant had been given an adequate opportunity in state court to raise his constitutional claims.” 412 U. S., at 255-256 (concurring opinion). In neither case, nor in Kuhlmann, did Justice Powell even suggest that federal habeas was available before 1963 to a prisoner who had received a full and fair opportunity to litigate his federal claim in state court.
Third, Justice O’Connor criticizes our failure to acknowledge Salinger v. Loisel, 265 U. S. 224 (1924), which she describes as the first case ex*287plicitly to hold that “res judicata is not strictly followed on federal ha-beas.” Post, at 299. Salinger, however, involved the degree of preclu-sive effect of a habeas judgment upon subsequent habeas petitions filed by a federal prisoner. This case, of course, involves the degree of preclusive effect of a criminal conviction upon an initial habeas petition filed by a state prisoner. We cannot fault ourselves for limiting our focus to the latter context. But even assuming its relevance, Salinger hardly advances the position advocated by Justice O’Connor that a habeas court must exercise de novo review with respect to mixed questions of law and fact. Despite acknowledging that a prior habeas judgment is not entitled to absolute preclusive effect under the doctrine of res judicata, Salinger also indicated that the prior habeas judgment “may be considered, and even given controlling weight.” 265 U. S., at 231 (emphasis added).
Justice O'Connor contends that the inclusion of this passage in a section of our opinion entitled “Right to a Plenary Hearing” makes clear that we were discussing only the resolution of factual questions. See post, at 300-301. In our introduction to that section, however, we indicated that both factual and legal questions were at issue. See 344 U. S., at 460 (noting contentions “that the District Court committed error when it took no evidence and heard no argument on the federal constitutional issues” (emphasis added)). Indeed, if only factual questions were at issue, we would have authorized a denial of the writ not whenever the state-court proceeding “has resulted in a satisfactory conclusion” (as we did), id., at 463 (emphasis added), but only whenever the state-court proceeding has resulted in satisfactory factfinding.
Justice O’Connor quotes Justice Frankfurter for the proposition that a district judge on habeas “ ‘must exercise his ownjudgment’ ” with respect to mixed questions. Post, at 300 (quoting 344 U. S., at 507). Although we agree with Justice O’Connor that this passage by itself suggests a de novo standard, it is not easily reconciled with Justice Frankfurter’s later statement that “there is no need for the federal judge, if he could, to shut his eyes to the State consideration” of the mixed question, id., at 508. These statements can be reconciled, of course, on the assumption that the habeas judge must review the state-court determination for reasonableness. But we need not attempt to defend that conclusion in detail, for *289we conclude not that Brown v. Allen establishes deferential review for reasonableness, but only that Brown does not squarely foreclose it.
We have no disagreement -with Justice O’Connok that Brown v. Allen quickly came to be cited for the proposition that a habeas court should review mixed questions “independently”; that several of our cases since Brown have applied a de novo standard with respect to pure and mixed legal questions; and that the de novo standard thus appeared well settled with respect to both categories by the time the Court decided Miller v. Fenton in 1985. See post, at 301-302. Despite her extended discussion of the leading cases from Brown through Miller, however, Justice O’Con-*290nor offers nothing to refute those of our limited observations with which she evidently disagrees — that an unadorned citation to Brown should not have been enough, at least as an original matter, to establish de novo review with respect to mixed questions; and that in none of our leading cases was the choice between a de novo and a deferential standard outcome determinative.
Justice O’Connor asserts that Jackson “expressly rejected” a “deferential standard of review” that she characterizes as “very much like the one” urged on us by petitioners. Post, at 303 (citing 443 U. S., at 323). What Jackson expressly rejected, however, was a proposal that habeas review “should be foreclosed” if the state courts provide “appellate review of the sufficiency of the evidence.” Ibid. That rule, of course, would permit no habeas review of a state-court sufficiency determination. As we understand it, however, petitioners’ proposal would permit limited review for reasonableness, a standard surely consistent with our own statement that that state-court determination “is of course entitled to deference.” Ibid. We agree with Justice O’Connor that Jackson itself applied a de novo standard. See post, at 303. Nonetheless, given our statement *291expressly endorsing a notion of at least limited deference, and given that the Jackson petitioner would have lost under either a de novo standard or a reasonableness standard, we cannot agree that the case “expressly rejected” the latter. Post, at 303.
Justice O’Connor suggests that Teague and its progeny “did not establish a standard of review at all.” Post, at 303-304. Instead, she contends, these cases merely prohibit the retroactive application of new rules on habeas, ibid., and establish the criterion for distinguishing new rules from old ones, ibid. We have no difficulty with describing Teague as a case about retroactivity, rather than standards of review, although we do not dispute Justice O'Connor’s suggestion that the difference, at least in practice, might well be “only ‘a matter of phrasing.’ ” Post, at 304 (cita*292tion omitted). We do disagree, however, with Justice O'Connor’s definition of what constitutes a “new rule” for Teague purposes. A rule is new, she contends, if it “can be meaningfully distinguished from that established by binding precedent at the time [the] state court conviction became final.” Post, at 304. This definition leads her to suggest that a habeas court must determine whether the state courts have interpreted old precedents “properly.” Post, at 805. Our precedents, however, require a different standard. We have held that a rule is “new” for Teague purposes whenever its validity under existing precedents is subject to debate among “reasonable minds,” Butler, 494 U. S., at 415, or among “reasonable jurists,” Sawyer v. Smith, 497 U. S. 227, 234 (1990). Indeed, each of our last four relevant precedents has indicated that Teague insulates on habeas review the state courts’ “‘reasonable, good-faith interpretations of existing precedents.’” Ibid, (quoting Butler, supra, at 414); Saffle v. Parks, 494 U. S. 484, 488 (1990) (citing Butler); see Stringer v. Black, 503 U. S. 222, 237 (1992) (“The purpose of the new rule doctrine is to validate reasonable interpretations of existing precedents”). Thus, Teague bars habeas relief whenever the state courts have interpreted old precedents reasonably, not only when they have done so “properly.” Post, at 305.
Justice O’Connor criticizes our failure to highlight in text the fact that Congress has considered, but failed to enact, several bills introduced during the last 25 years to prohibit de novo review explicitly. See post, at 305; see also Brief for Senator Biden et al. as Amici Curiae 10-16 (discussing various proposals). Our task, however, is not to construe bills that Congress has failed to enact, but to construe statutes that Congress has enacted. The habeas corpus statute was last amended in 1966. See 80 Stat. 1104-1105. We have grave doubts that post-1966 legislative history is of any value in construing its provisions, for we have often observed that “‘the views of a subsequent Congress form a hazardous basis for inferring the intent of an earlier one.”’ Consumer Product Safety Comm’n v. GTE Sylvania, Inc., 447 U. S. 102, 117 (1980), quoting United States v. Price, 361 U. S. 304, 313 (1960). Compare also Sullivan v. Fink-elstein, 496 U. S. 617, 628, n. 8 (1990) (acknowledging “all the usual difficulties inherent in relying on subsequent legislative history”), with id., at 632 (Scalia, J., concurring in part) (“Arguments based on subsequent legislative history, like arguments based on antecedent futurity, should not be taken seriously”).