delivered the opinion of the Court.

The doctrine of abuse of the writ defines the circumstances in which federal courts decline to entertain a claim presented for the first time in a second or subsequent petition for a writ of habeas corpus. Petitioner Warren McCleskey in a second federal habeas petition presented a claim under Massiah v. United States, 377 U. S. 201 (1964), that he failed to include in his first federal petition. The Court of Appeals for the Eleventh Circuit held that assertion of the Massiah claim in this manner abused the writ. Though our analysis differs from that of the Court of Appeals, we agree that the petitioner here abused the writ, and we affirm the judgment.

I — I

McCleskey and three other men, all armed, robbed a Georgia furniture store in 1978. One of the robbers shot and killed an off duty policeman who entered the store in the midst of the crime. McCleskey confessed to the police that he participated in the robbery. When on trial for both the robbery and the murder, however, McCleskey renounced his confession after taking the stand with an alibi denying all involvement. To rebut McCleskey’s testimony, the prosecution called Offie Evans, who had occupied a jail cell next to McCleskey’s. Evans testified that McCleskey admitted shooting the officer during the robbery and boasted that he would have shot his way out of the store even in the face of a dozen policemen.

Although no one witnessed the shooting, further direct and circumstantial evidence supported McCleskey’s guilt of the murder. An eyewitness testified that someone ran from the store carrying a pearl-handled pistol soon after the robbery. Other witnesses testified that McCleskey earlier had stolen a pearl-handled pistol of the same caliber as the bullet that killed the officer. Ben Wright, one of McCleskey’s accomplices, confirmed that during the crime McCleskey carried a white-handled handgun matching the caliber of the fatal bul*471let. Wright also testified that McCleskey admitted shooting the officer. Finally, the prosecutor introduced McCleskey’s confession of participation in the robbery.

In December 1978, the jury convicted McCleskey of murder and sentenced him to death. Since his conviction, Mc-Cleskey has pursued direct and collateral remedies for more than a decade. We describe this procedural history in detail, both for a proper understanding of the case and as an illustration of the context in which allegations of abuse of the writ arise.

On direct appeal to the Supreme Court of Georgia, McCles-key raised six grounds of error. A summary of McCleskey’s claims on direct appeal, as well as those he asserted in each of his four collateral proceedings, is set forth in the Appendix to this opinion, infra, at 503. The portion of the appeal relevant for our purposes involves McCleskey’s attack on Evans’ rebuttal testimony. McCleskey contended that the trial court “erred in allowing evidence of [McCleskey’s] oral statement admitting the murder made to [Evans] in the next cell, because the prosecutor had deliberately withheld such statement” in violation of Brady v. Maryland, 373 U. S. 83 (1963). McClesky v. State, 245 Ga. 108, 112, 263 S. E. 2d 146, 149 (1980). A unanimous Georgia Supreme Court acknowledged that the prosecutor did not furnish Evans’ statement to the defense, but ruled that because the undisclosed evidence was not exculpatory, McCleskey suffered no material prejudice and was not denied a fair trial under Brady. 245 Ga., at 112-113, 263 S. E. 2d, at 149. The court noted, moreover, that the evidence McCleskey wanted to inspect was “introduced to the jury in its entirety” through Evans’ testimony, and that McCleskey’s argument that “the evidence was needed in order to prepare a proper defense or impeach other witnesses ha[d] no merit because the evidence requested was statements made by [McCleskey] himself.” Ibid. The court rejected McCleskey’s other contentions and *472affirmed his conviction and sentence. Ibid. We denied cer-tiorari. McClesky v. Georgia, 449 U. S. 891 (1980).

McCleskey then initiated postconviction proceedings. In January 1981, he filed a petition for state habeas corpus relief. The amended petition raised 23 challenges to his murder conviction and death sentence. See Appendix, infra, at 503. Three of the claims concerned Evans’ testimony. First, McCleskey contended that the State violated his due process rights under Giglio v. United States, 405 U. S. 150 (1972), by its failure to disclose an agreement to drop pending escape charges against Evans in return for his cooperation and testimony. App. 20. Second, McCleskey reasserted his Brady claim that the State violated his due process rights by the deliberate withholding of the statement he made to Evans while in jail. App. 21. Third, McCleskey alleged that admission of Evans’ testimony violated the Sixth Amendment right to counsel as construed in Massiah v. United States, supra. On this theory, “[t]he introduction into evidence of [his] statements to [Evans], elicited in a situation created to induce [McCleskey] to make incriminating statements without the assistance of counsel, violated [McCleskey’s] right to counsel under the Sixth Amendment to the Constitution of the United States.” App. 22.

At the state habeas corpus hearing, Evans testified that one of the detectives investigating the murder agreed to speak a word on his behalf to the federal authorities about certain federal charges pending against him. The state ha-beas court ruled that the ex parte recommendation did not implicate Giglio, and it denied relief on all other claims. The Supreme Court of Georgia denied McCleskey’s application for a certificate of probable cause, and we denied his second petition for a writ of certiorari. McCleskey v. Zant, 454 U. S. 1093 (1981).

In December 1981, McCleskey filed his first federal habeas corpus petition in the United States District Court for the Northern District of Georgia, asserting 18 grounds for relief. *473See Appendix, infra, at 504-505. The petition failed to allege the Massiah claim, but it did reassert the Giglio and Brady claims. Following extensive hearings in August and October 1983, the District Court held that the detective’s statement to Evans was a promise of favorable treatment, and that failure to disclose the promise violated Giglio. McCleskey v. Zant, 580 F. Supp. 338, 380-384 (ND Ga. 1984). The District Court further held that Evans’ trial testimony may have affected the jury’s verdict on the charge of malice murder. On these premises it granted relief. Id., at 384.

The Court of Appeals reversed the District Court’s grant of the writ. McCleskey v. Kemp, 753 F. 2d 877 (CA11 1985). The court held that the State had not made a promise to Evans of the kind contemplated by Giglio, and that in any event the Giglio error would be harmless. 753 F. 2d, at 884-885. The court affirmed the District Court on all other grounds. We granted certiorari limited to the question whether Georgia’s capital sentencing procedures were constitutional, and denied relief. McCleskey v. Kemp, 481 U. S. 279 (1987).

McCleskey continued his postconviction attacks by filing a second state habeas corpus action in 1987 which, as amended, contained five claims for relief. See Appendix, infra, at 505. One of the claims again centered on Evans’ testimony, alleging that the State had an agreement with Evans that it had failed to disclose. The state trial court held a hearing and dismissed the petition. The Supreme Court of Georgia denied McCleskey’s application for a certificate of probable cause.

In July 1987, McCleskey filed a second federal habeas action, the one we now review. In the District Court, McCles-key asserted seven claims, including a Massiah challenge to the introduction of Evans’ testimony. See Appendix, infra, at 506. McCleskey had presented a Massiah claim, it will be recalled, in his first state habeas action when he alleged that the conversation recounted by Evans at trial had been “elic*474ited in a situation created to induce” him to make an incriminating statement without the assistance of counsel. The first federal petition did not present a Massiah claim. The proffered basis for the Massiah claim in the second federal petition was a 21-page signed statement that Evans made to the Atlanta Police Department on August 1, 1978, two weeks before the trial began. The department furnished the document to McCleskey one month before he filed his second federal petition.

The statement related pretrial jailhouse conversations that Evans had with McCleskey and that Evans overheard between McCleskey and Bernard Dupree. By the statement’s own terms, McCleskey participated in all the reported jail-cell conversations. Consistent with Evans’ testimony at trial, the statement reports McCleskey admitting and boasting about the murder. It also recounts that Evans posed as Ben Wright’s uncle and told McCleskey he had talked with Wright about the robbery and the murder.

In his second federal habeas petition, McCleskey asserted that the statement proved Evans “was acting in direct concert with State officials” during the incriminating conversations with McCleskey, and that the authorities “deliberately elicited” inculpatory admissions in violation of McCleskey’s Sixth Amendment right to counsel. Massiah v. United States, 377 U. S., at 206. 1 Tr. Exh. 1, pp. 11-12. Among other responses, the State of Georgia contended that Mc-Cleskey’s presentation of a Massiah claim for the first time in the second federal petition was an abuse of the writ. 28 U. S. C. § 2244(b); Rule 9(b) of the Rules Governing §2254 Cases.

The District Court held extensive hearings in July and August 1987 focusing on the arrangement the jailers had made for Evans’ cell assignment in 1978. Several witnesses denied that Evans had been placed next to McCleskey by design or instructed to overhear conversations or obtain statements from McCleskey. McCleskey’s key witness was Ulysses *475Worthy, a jailer at the Fulton County Jail during the summer of 1978. McCleskey’s lawyers contacted Worthy after a detective testified that the 1978 Evans statement was taken in Worthy’s office. The District Court characterized Worthy’s testimony as “often confused and self-contradictory.” McCleskey v. Kemp, No. C87-1517A (ND Ga., Dec. 23, 1987), App. 81. Worthy testified that someone at some time requested permission to move Evans near McCleskey’s cell. He contradicted himself, however, concerning when, why, and by whom Evans was moved, and about whether he overheard investigators urging Evans to engage McCleskey in conversation. Id., at 76-81.

On December 23, 1987, the District Court granted McCles-key relief based upon a violation of Massiah. Id., at 63-97. The court stated that the Evans statement “contains strong indication of an ab initio relationship between Evans and the authorities.” Id., at 84. In addition, the court credited Worthy’s testimony suggesting that the police had used Evans to obtain incriminating information from McCleskey. Based on the Evans statement and portions of Worthy’s testimony, the District Court found that the jail authorities had placed Evans in the cell adjoining McCleskey’s “for the purpose of gathering incriminating information”; that “Evans was probably coached in how to approach McCleskey and given critical facts unknown to the general public”; that Evans talked with McCleskey and eavesdropped on McCles-key’s conversations with others; and that Evans reported what he had heard to the authorities. Id.,, at 83. These findings, in the District Court’s view, established a Massiah violation.

In granting habeas relief, the District Court rejected the State’s argument that McCleskey’s assertion of the Massiah claim for the first time in the second federal petition constituted an abuse of the writ. The court ruled that McCleskey did not deliberately abandon the claim after raising it in his first state habeas petition. “This is not a case,” the District *476Court reasoned, “where petitioner has reserved his proof or deliberately withheld his claim for a second petition.” Id., at 84. The District Court also determined that when McCles-key filed his first federal petition, he did not know about either the 21-page Evans document or the identity of Worthy, and that the failure to discover the evidence for the first federal petition “was not due to [McCleskey’s] inexcusable neglect.” Id., at 85.

The Eleventh Circuit reversed, holding that the District Court abused its discretion by failing to dismiss McCleskey’s Massiah claim as an abuse of the writ. 890 F. 2d 342 (1989). The Court of Appeals agreed with the District Court that the petitioner must “show that he did not deliberately abandon the claim and that his failure to raise it [in the first federal habeas proceeding] was not due to inexcusable neglect.” Id., at 346-347. Accepting the District Court’s findings that at the first petition stage McCleskey knew neither the existence of the Evans statement nor the identity of Worthy, the court held that the District Court “miscon-stru[ed] the meaning of deliberate abandonment.” Id., at 348-349. Because McCleskey included a Massiah claim in his first state petition, dropped it in his first federal petition, and then reasserted it in his second federal petition, he “made a knowing choice not to pursue the claim after having raised it previously” that constituted a prima facie showing of “deliberate abandonment.” 890 F. 2d, at 349. The court further found the State’s alleged concealment of the Evans statement irrelevant because it “was simply the catalyst that caused counsel to pursue the Massiah claim more vigorously” and did not itself “demonstrate the existence of a Massiah violation.” Id., at 350. The court concluded that McCleskey had presented no reason why counsel could not have discovered Worthy earlier. Ibid. Finally, the court ruled that McCleskey’s claim did not fall within the ends of justice exception to the abuse-of-the-writ doctrine because any *477Massiah violation that may have been committed would have been harmless error. 890 F. 2d, at 350-351.

McCleskey petitioned this Court for a writ of certiorari, alleging numerous errors in the Eleventh Circuit’s abuse-of-the-writ analysis. In our order granting the petition, we requested the parties to address the following additional question: “Must the State demonstrate that a claim was deliberately abandoned in an earlier petition for a writ of ha-beas corpus in order to establish that inclusion of that claim in a subsequent habeas petition constitutes abuse of the writ?” 496 U. S. 904 (1990).

II

The parties agree that the government has the burden of pleading abuse of the writ, and that once the government makes a proper submission, the petitioner must show that he has not abused the writ in seeking habeas relief. See Sanders v. United States, 373 U. S. 1, 10-11 (1963); Price v. Johnston, 334 U. S. 266, 292 (1948). Much confusion exists though, on the standard for determining when a petitioner abuses the writ. Although the standard is central to the proper determination of many federal habeas corpus actions, we have had little occasion to define it. Indeed, there is truth to the observation that we have defined abuse of the writ in an oblique way, through dicta and denials of certiorari petitions or stay applications. See Witt v. Wainwright, 470 U. S. 1039, 1043 (1985) (Marshall, J., dissenting). Today we give the subject our careful consideration. We begin by tracing the historical development of some of the substantive and procedural aspects of the writ, and then consider the standard for abuse that district courts should apply in actions seeking federal habeas corpus relief.

A

The Judiciary Act of 1789, ch. 20, § 14, 1 Stat. 81-82, empowered federal courts to issue writs of habeas corpus to prisoners “in custody, under or by colour of the authority of *478the United States.” In the early decades of our new federal system, English common law defined the substantive scope of the writ. Ex parte Watkins, 3 Pet. 193, 201-203 (1830). Federal prisoners could use the writ to challenge confinement imposed by a court that lacked jurisdiction, ibid., or detention by the Executive without proper legal process, see Ex parte Wells, 18 How. 307 (1856).

The common-law limitations on the scope of the writ were subject to various expansive forces, both statutory and judicial. See generally Bator, Finality in Criminal Law and Federal Habeas Corpus for State Prisoners, 76 Harv. L. Rev. 441, 463-499 (1963). The major statutory expansion of the writ occurred in 1867, when Congress extended federal habeas corpus to prisoners held in state custody. Act of Feb. 5, 1867, ch. 28, § 1, 14 Stat. 385. For the most part, however, expansion of the writ has come through judicial de-cisionmaking. As then-JuSTiCE Rehnquist explained in Wainwright v. Sykes, 433 U. S. 72, 79 (1977), the Court began by interpreting the concept of jurisdictional defect with generosity to include sentences imposed without statutory authorization, Ex parte Lange, 18 Wall. 163, 176 (1874), and convictions obtained under an unconstitutional statute, Ex parte Siebold, 100 U. S. 371, 376-377 (1880). Later, we allowed habeas relief for confinement under a state conviction obtained without adequate procedural protections for the defendant. Frank v. Mangum, 237 U. S. 309 (1915); Moore v. Dempsey, 261 U. S. 86 (1923).

Confronting this line of precedents extending the reach of the writ, in Waley v. Johnston, 316 U. S. 101 (1942), “the Court openly discarded the concept of jurisdiction — by then more a fiction than anything else — as a touchstone of the availability of federal habeas review, and acknowledged that such review is available for claims of ‘disregard of the constitutional rights of the accused, and where the writ is the only effective means of preserving his rights.’” Wainwright v. Sykes, supra, at 79 (quoting Waley v. Johnston, supra, at *479104-105). With the exception of Fourth Amendment violations that a petitioner has been given a full and fair opportunity to litigate in state court, Stone v. Powell, 428 U. S. 465, 495 (1976), the writ today appears to extend to all dispositive constitutional claims presented in a proper procedural manner. See Brown v. Allen, 344 U. S. 443 (1953); Wainwright v. Sykes, supra, at 79.

One procedural requisite is that a petition not lead to an abuse of the writ. We must next consider the origins and meaning of that rule.

B

At common law, res judicata did not attach to a court’s denial of habeas relief. “[A] refusal to discharge on one writ [was] not a bar to the issuance of a new writ.” 1 W. Bailey, Habeas Corpus and Special Remedies 206 (1913) (citing cases). “[A] renewed application could be made to every other judge or court in the realm, and each court or judge was bound to consider the question of the prisoner’s right to a discharge independently, and not to be influenced by the previous decisions refusing discharge.” W. Church, Writ of Habeas Corpus §386, p. 570 (2d ed. 1893) (hereinafter Church). See, e. g., Ex parte Kaine, 14 F. Cas. 78, 80 (No. 7,597) (CC SDNY 1853); In re Kopel, 148 F. 505, 506 (SDNY 1906). The rule made sense because at common law an order denying habeas relief could not be reviewed. Church 570; L. Yackle, Postconviction Remedies §151, p. 551 (1981); Goddard, A Note on Habeas Corpus, 65 L. Q. Rev. 30, 32 (1949). Successive petitions served as a substitute for appeal. See W. Duker, A Constitutional History of Habeas Corpus 5-6 (1980); Church 570; Goddard, supra, at 35.

As appellate review became available from a decision in ha-beas refusing to discharge the prisoner, courts began to question the continuing validity of the common-law rule allowing endless successive petitions. Church 602. Some courts rejected the common-law rule, holding a denial of habeas relief *480res judicata. See, e. g., Perry v. McLendon, 62 Ga. 598, 603-605 (1879); McMahon v. Mead, 30 S. D. 515, 518, 139 N. W. 122, 123 (1912); Ex parte Heller, 146 Wis. 517, 524, 131 N. W. 991, 994 (1911). Others adopted a middle position between the extremes of res judicata and endless successive petitions. Justice Field’s opinion on circuit in Ex parte Cuddy, 40 F. 62 (CC SD Cal. 1889), exemplifies this balance.

“[Wjhile the doctrine of res judicata does not apply, . . . the officers before whom the second application is made may take into consideration the fact that a previous application had been made to another officer and refused; and in some instances that fact may justify a refusal of the second. The action of the court or justice on the second application will naturally be affected to some degree by the character of the court or officer to whom the first application was made, and the fullness of the consideration given to it. . . . In what I have said I refer, of course, to cases where a second application is made upon the same facts presented, or which might have been presented, on the first. The question is entirely different when subsequent occurring events have changed the situation of the petitioner so as in fact to present a new case for consideration. In the present application there are no new facts which did not exist when the first was presented. ... I am of the opinion that in such a case a second application should not be heard . . . .” Id., at 65-66.

Cf. Ex parte Moebus, 148 F. 39, 40-41 (NH 1906) (second petition disallowed “unless some substantial change in the circumstances had intervened”).

We resolved the confusion over the continuing validity of the common-law rule, at least for federal courts, in Salinger v. Loisel, 265 U. S. 224 (1924), and Wong Doo v. United States, 265 U. S. 239 (1924). These decisions reaffirmed that res judicata does not apply “to a decision on habeas corpus refusing to discharge the prisoner.” Salinger v. Loisel, *481supra, at 230; see Wong Doo v. United States, supra, at 240. They recognized, however, that the availability of appellate review required a modification of the common-law rule allowing endless applications. As we explained in Salinger:

“In early times when a refusal to discharge was not open to appellate review, courts and judges were accustomed to exercise an independent judgment on each successive application, regardless of the number. But when a right to an appellate review was given the reason for that practice ceased and the practice came to be materially changed . . . .” 265 U. S., at 230-231.

Relying on Justice Field’s opinion in Ex parte Cuddy, we announced that second and subsequent petitions should be

“disposed of in the exercise of a sound judicial discretion guided and controlled by a consideration of whatever has a rational bearing on the propriety of the discharge sought. Among the matters which may be considered, and even given controlling weight, are (a) the existence of another remedy, such as a right in ordinary course to an appellate review in the criminal case, and (b) a prior refusal to discharge on a like application.” 265 U. S., at 231.

Because the lower court in Salinger had not disposed of the subsequent application for habeas corpus by reliance on dismissal of the prior application, the decision did not present an opportunity to apply the doctrine of abuse of the writ. 265 U. S., at 232. Wong Doo did present the question. There, the District Court had dismissed on res judicata grounds a second petition containing a due process claim that was raised, but not argued, in the first federal habeas petition. The petitioner “had full opportunity to offer proof of [his due process claim] at the hearing on the first petition,” and he offered “[n]o reason for not presenting the proof at the outset.. . .” Wong Doo, 265 U. S., at 241. The record of the first petition did not contain proof of the due process claim, *482but “what [was] said of it there and in the briefs show[ed] that it was accessible all the time.” Ibid. In these circumstances, we upheld the dismissal of the second petition. We held that “according to a sound judicial discretion, controlling weight must have been given to the prior refusal.” Ibid. So while we rejected res judicata in a strict sense as a basis for dismissing a later habeas action, we made clear that the prior adjudication bore vital relevance to the exercise of the court’s discretion in determining whether to consider the petition.

Price v. Johnston, 334 U. S. 266 (1948), the next decision in this line, arose in a somewhat different context from Salinger or Wong Doo. In Price, the petitioner’s fourth habeas petition alleged a claim that, arguably at least, was neither the explicit basis of a former petition nor inferable from the facts earlier alleged. The District Court and Court of Appeals dismissed the petition without hearing on the sole ground that the claim was not raised in one of the earlier ha-beas actions. We reversed and remanded, reasoning that the dismissal “precluded a proper development of the issue of the allegedly abusive use of the habeas corpus writ.” 334 U. S., at 293. We explained that the State must plead an abuse of the writ with particularity, and that the burden then shifts to petitioner to show that presentation of the new claim does not constitute abuse. Id., at 292. The District Court erred because it dismissed the petition without affording the petitioner an opportunity to explain the basis for raising his claim late. We gave directions for the proper inquiry in the trial court. If the explanation “is inadequate, the court may dismiss the petition without further proceedings.” Ibid. But if a petitioner “presents] adequate reasons for not making the allegation earlier, reasons which make it fair and just for the trial court to overlook the delay,” he must be given the opportunity to develop these matters in a hearing. Id., at 291-292. Without considering whether the petitioner had abused the writ, we remanded the case.

*483Although Price recognized that abuse-of-the-writ principles limit a petitioner’s ability to file repetitive petitions, it also contained dicta touching on the standard for abuse that appeared to contradict this point. Price stated that “the three prior refusals to discharge petitioner can have no bearing or weight on the disposition to be made of the new matter raised in the fourth petition.” Id., at 289. This proposition ignored the significance of appellate jurisdictional changes, see supra, at 479-480, as well as the general disfavor we had expressed in Salinger and Wong Doo toward endless repetitive petitions. It did not even comport with language in Price itself which recognized that in certain circumstances new claims raised for the first time in a second or subsequent petition should not be entertained. As will become clear, the quoted portion of Price has been ignored in our later decisions.

One month after the Price decision, Congress enacted legislation, which for the first time addressed the issue of repetitive federal habeas corpus petitions:

“No circuit or district judge shall be required to entertain an application for a writ of habeas corpus to inquire into the detention of a person pursuant to a judgment of a court of the United States, or of any State, if it appears that the legality of such detention has been determined by a judge or court of the United States on a prior application for a writ of habeas corpus and the petition presents no new ground not theretofore presented and determined, and the judge or court is satisfied that the ends of justice will not be served by such inquiry.” 28 U. S. C. §2244 (1964 ed.).

Because § 2244 allowed a district court to dismiss a successive petition that “present[ed] no new ground not theretofore presented and determined,” one might have concluded, by negative implication, that Congress denied permission to dismiss any petition that alleged new grounds for relief. Such an interpretation would have superseded the judicial principles *484recognizing that claims not raised or litigated in a prior petition could, when raised in a later petition, constitute abuse. But the Reviser’s Note to the 1948 statute made clear that as a general matter Congress did not intend the new section to disrupt the judicial evolution of habeas principles, 28 U. S. C. §2244 (1964 ed.) (Reviser’s Note), and we confirmed in Sanders v. United States, 373 U. S., at 11-12, that Congress’ silence on the standard for abuse of the writ involving a new claim was “not intended to foreclose judicial application of the abuse-of-writ principle as developed in Wong Doo and Price.”

Sanders also recognized our special responsibility in the development of habeas corpus with respect to another provision of the 1948 revision of the judicial code, 28 U. S. C. §2255 (1964 ed.). The statute created a new postconviction remedy for federal prisoners with a provision for repetitive petitions different from the one found in §2244. While §2244 permitted dismissal of subsequent habeas petitions that “presented] no new ground not theretofore presented and determined,” §2255 allowed a federal district court to refuse to entertain a subsequent petition seeking “similar relief.” On its face, §2255 appeared to announce a much stricter abuse-of-the-writ standard than its counterpart in §2244. We concluded in Sanders, however, that the language in §2255 “cannot be taken literally,” and construed it to be the “material equivalent” of the abuse standard in §2244. Sanders v. United States, 373 U. S., at 13-14.

In addition to answering these questions, Sanders undertook a more general “formulation of basic rules to guide the lower federal courts” concerning the doctrine of abuse of the writ. Id., at 15. After reiterating that the government must plead abuse of the writ and the petitioner must refute a well-pleaded allegation, Sanders addressed the definition of and rationale for the doctrine. It noted that equitable principles governed abuse of the writ, including “the principle that a suitor’s conduct in relation to the matter at hand may *485disentitle him to the relief he seeks,” and that these principles must be applied within the sound discretion of district courts. Id., at 17-18. The Court furnished illustrations of writ abuse:

“Thus, for example, if a prisoner deliberately withholds one of two grounds for federal collateral relief at the time of filing his first application, in the hope of being granted two hearings rather than one or for some other such reason, he may be deemed to have waived his right to a hearing on a second application presenting the withheld ground. The same may be true if, as in Wong Doo, the prisoner deliberately abandons one of his grounds at the first hearing. Nothing in the traditions of habeas corpus requires the federal courts to tolerate needless, piecemeal litigation, or to entertain collateral proceedings whose only purpose is to vex, harass, or delay.” Id., at 18.

The Court also cited Fay v. Noia, 372 U. S. 391, 438-440 (1963), and Townsend v. Sain, 372 U. S. 293, 317 (1963), for further guidance on the doctrine of abuse of the writ, stating that the principles of those cases “govern equally here.” 373 U. S., at 18. Finally, Sanders established that federal courts must reach the merits of an abusive petition if “the ends of justice demand.” Ibid.

Three years after Sanders, Congress once more amended the habeas corpus statute. The amendment was an attempt to alleviate the increasing burden on federal courts caused by successive and abusive petitions by “introducing a greater degree of finality of judgments in habeas corpus proceedings.” S. Rep. No. 1797, 89th Cong., 2d Sess., 2 (1966); see also H. R. Rep. No. 1892, 89th Cong., 2d Sess., 5-6 (1966). The amendment recast §2244 into three subparagraphs. Subparagraph (a) deletes the reference to state prisoners in the old §2244 but left the provision otherwise intact. 28 U. S. C. § 2244(a). Subparagraph (c) states that where a state prisoner seeks relief for an alleged denial of a federal *486constitutional right before this Court, any decision rendered by the Court shall be “conclusive as to all issues of fact or law with respect to an asserted denial of a Federal right . . . 28 U. S. C. § 2244(c).

Congress added subparagraph (b) to address repetitive applications by state prisoners:

“(b) When after an evidentiary hearing on the merits of a material factual issue, or after a hearing on the merits of an issue of law, a person in custody pursuant to the judgment of a State court has been denied by a court of the United States or a justice or judge of the United States release from custody or other remedy on an application for a writ of habeas corpus, a subsequent application for a writ of habeas corpus on behalf of such person need not be entertained by a court of the United States or a justice or judge of the United States unless the application alleges and is predicated on a factual or other ground not adjudicated on the hearing of the earlier application for the writ, and unless the court, justice, or judge is satisfied that the applicant has not on the earlier application deliberately withheld the newly asserted ground or otherwise abused the writ.” 28 U. S. C. § 2244(b).

Subparagraph (b) establishes a “qualified application of the doctrine of res judicata.” S. Rep. No. 1797, supra, at 2. It states that a federal court “need not entertain” a second or subsequent habeas petition “unless” the petitioner satisfies two conditions. First, the subsequent petition must allege a new ground, factual or otherwise. Second, the applicant must satisfy the judge that he did not deliberately withhold the ground earlier or “otherwise abus[e] the writ.” See Smith v. Yeager, 393 U. S. 122, 125 (1968) (“essential question [under § 2244(b)] is whether the petitioner ‘deliberately withheld the newly asserted ground’ in the prior proceeding, or ‘otherwise abused the writ’”). If the petitioner meets these conditions, the court must consider the subsequent pe*487tition as long as other habeas errors, such as nonexhaustion, 28 U. S. C. § 2254(b), or procedural default, Wainwright v. Sykes, 433 U. S. 72 (1977), are not present.

Section 2244(b) raises, but does not answer, other questions. It does not state whether a district court may overlook a deliberately withheld or otherwise abusive claim to entertain the petition in any event. That is, it does not state the limits on the district court’s discretion to entertain abusive petitions. Nor does the statute define the term “abuse of the writ.” As was true of similar silences in the original 1948 version of § 2244, however, see supra, at 484, Congress did not intend § 2244(b) to foreclose application of the court-announced principles defining and limiting a district court’s discretion to entertain abusive petitions. See Delo v. Stokes, 495 U. S. 320, 321-322 (1990) (District Court abused discretion in entertaining a new claim in a fourth federal petition that was an abuse of the writ).

Rule 9(b) of the Rules Governing Habeas Corpus Proceedings, promulgated in 1976, also speaks to the problem of new grounds for relief raised in subsequent petitions. It provides:

“A second or successive petition may be dismissed if the judge finds that it fails to allege new or different grounds for relief and the prior determination was on the merits or, if new and different grounds are alleged, the judge finds that the failure of the petitioner to assert those grounds in a prior petition constituted an abuse of the writ.” 28 U. S. C. §2254 Rule 9(b).

Like 28 U. S. C. § 2244(b), Rule 9(b) “incorporates the judge-made principle governing the abuse of the writ set forth in Sanders.” Rose v. Lundy, 455 U. S. 509, 521 (1982) (plurality opinion); id., at 533 (Brennan, J., concurring in part and dissenting in part) (same). The Advisory Committee Notes make clear that a new claim in a subsequent petition should not be entertained if the judge finds the failure to raise it earlier “inexcusable.” Advisory Committee Notes to *488Rule 9, 28 U. S. C., pp. 426-427. The Notes also state that a retroactive change in the law and newly discovered evidence represent acceptable excuses for failing to raise the claim earlier. Id., at 427.

In recent years we have applied the abuse-of-the-writ doctrine in various contexts. In Woodard v. Hutchins, 464 U. S. 377 (1984) (per curiam), the petitioner offered no explanation for asserting three claims in a second federal ha-beas petition not raised in the first. Five Justices inferred from the lack of explanation that the three claims “could and should have been raised in” the first petition, and that the failure to do so constituted abuse of the writ. Id., at 378-379, and n. 3 (Powell, J., joined by four Justices, concurring in grant of application to vacate stay). Similarly, in Antone v. Dugger, 465 U. S. 200 (1984) (per curiam), we upheld the Court of Appeals’ judgment that claims presented for the first time in a second federal petition constituted an abuse of the writ. We rejected petitioner’s argument that he should be excused from his failure to raise the claims in the first federal petition because his counsel during first federal habeas prepared the petition in haste and did not have time to become familiar with the ease. Id., at 205-206, and n. 4. And just last Term, we held that claims raised for the first time in a fourth federal habeas petition abused the writ because they “could have been raised” or “could have been developed” in the first federal habeas petition. Delo v. Stokes, supra, at 321-322. See also Kuhlmann v. Wilson, 477 U. S. 436, 444, n. 6 (1986) (plurality opinion) (petition that raises grounds “available but not relied upon in a prior petition” is an example of abuse of the writ); Straight v. Wainwright, 476 U. S. 1132, 1133 (1986) (Powell, J., joined by three Justices, concurring in denial of stay) (new arguments in second petition that “plainly could have been raised earlier” constitute abuse of the writ); Rose v. Lundy, supra, at 521 (plurality opinion) (prisoner who proceeds with exhausted claims in first federal *489petition and deliberately sets aside his unexhausted claims risks dismissal of subsequent federal petitions).

III

Our discussion demonstrates that the doctrine of abuse of the writ refers to a complex and evolving body of equitable principles informed and controlled by historical usage, statutory developments, and judicial decisions. Because of historical changes and the complexity of the subject, the Court has not “always followed an unwavering line in its conclusions as to the availability of the Great Writ.” Fay v. Noia, 372 U. S., at 411-412. Today we attempt to define the doctrine of abuse of the writ with more precision.

Although our decisions on the subject do not all admit of ready synthesis, one point emerges with clarity: Abuse of the writ is not confined to instances of deliberate abandonment. Sanders mentioned deliberate abandonment as but one example of conduct that disentitled a petitioner to relief. Sanders cited a passage in Townsend v. Sain, 372 U. S., at 317, which applied the principle of inexcusable neglect, and noted that this principle also governs in the abuse-of-the-writ context, Sanders v. United States, 373 U. S., at 18.

As Sanders’ reference to Townsend demonstrates, as many Courts of Appeals recognize, see, e. g., 890 F. 2d, at 346-347 (case below); Hall v. Lockhart, 863 F. 2d 609, 610 (CA8 1988); Jones v. Estelle, 722 F. 2d 159, 163 (CA5 1983); Miller v. Bordenkircher, 764 F. 2d 245, 260-252 (CA4 1985), and as McCleskey concedes, Brief for Petitioner 39-40, 45-48, a petitioner may abuse the writ by failing to raise a claim through inexcusable neglect. Our recent decisions confirm that a petitioner can abuse the writ by raising a claim in a subsequent petition that he could have raised in his first, regardless of whether the failure to raise it earlier stemmed from a deliberate choice. See, e. g., Delo v. Stokes, 495 U. S., at 321-322; Antone v. Dugger, supra, at 205-206. See also 28 U. S. C. § 2244(b) (recognizing that a petitioner *490can abuse the writ in a fashion that does not constitute deliberate abandonment).

The inexcusable neglect standard demands more from a petitioner than the standard of deliberate abandonment. But we have not given the former term the content necessary to guide district courts in the ordered consideration of allegedly abusive habeas corpus petitions. For reasons we explain below, a review of our habeas corpus precedents leads us to decide that the same standard used to determine whether to excuse state procedural defaults should govern the determination of inexcusable neglect in the abuse-of-the-writ context.

The prohibition against adjudication in federal habeas corpus of claims defaulted in state court is similar in purpose and design to the abuse-of-the-writ doctrine, which in general prohibits subsequent habeas consideration of claims not raised, and thus defaulted, in the first federal habeas proceeding. The terms “abuse of the writ” and “inexcusable neglect,” on the one hand, and “procedural default,” on the other, imply a background norm of procedural regularity binding on the petitioner. This explains the presumption against habeas adjudication both of claims defaulted in state court and of claims defaulted in the first round of federal ha-beas. A federal habeas court’s power to excuse these types of defaulted claims derives from the court’s equitable discretion. See Reed v. Ross, 468 U. S. 1, 9 (1984) (procedural default); Sanders v. United States, 373 U. S., at 17-18 (abuse of the writ). In habeas, equity recognizes that “a suitor’s conduct in relation to the matter at hand may disentitle him to the relief he seeks.” Id., at 17. For these reasons, both the abuse-of-the-writ doctrine and our procedural default jurisprudence concentrate on a petitioner’s acts to determine whether he has a legitimate excuse for failing to raise a claim at the appropriate time.

The doctrines of procedural default and abuse of the writ implicate nearly identical concerns flowing from the signifi*491cant costs of federal habeas corpus review. To begin with, the writ strikes at finality. One of the law’s very objects is the finality of its judgments. Neither innocence nor just punishment can be vindicated until the final judgment is known. “Without finality, the criminal law is deprived of much of its deterrent effect.” Teague v. Lane, 489 U. S. 288, 809 (1989). And when a habeas petitioner succeeds in obtaining a new trial, the “ ‘erosion of memory’ and ‘dispersion of witnesses’ that occur with the passage of time,” Kuhlmann v. Wilson, supra, at 453, prejudice the government and diminish the chances of a reliable criminal adjudication. Though Fay v. Noia, supra, may have cast doubt upon these propositions, since Fay we have taken care in our habeas corpus decisions to reconfirm the importance of finality. See, e. g., Teague v. Lane, supra, at 308-309; Murray v. Carrier, 477 U. S. 478, 487 (1986); Reed v. Ross, supra, at 10; Engle v. Isaac, 456 U. S. 107, 127 (1982).

Finality has special importance in the context of a federal attack on a state conviction. Murray v. Carrier, supra, at 487; Engle v. Isaac, supra, at 128. Reexamination of state convictions on federal habeas “frustrate[s] . . . ‘both the States’ sovereign power to punish offenders and their good-faith attempts to honor constitutional rights.’” Murray v. Carrier, supra, at 487 (quoting Engle, supra, at 128). Our federal system recognizes the independent power of a State to articulate societal norms through criminal law; but the power of a State to pass laws means little if the State cannot enforce them.

Habeas review extracts further costs. Federal collateral litigation places a heavy burden on scarce federal judicial resources, and threatens the capacity of the system to resolve primary disputes. Schneckloth v. Bustamonte, 412 U. S. 218, 260 (1973) (Powell, J., concurring). Finally, habeas corpus review may give litigants incentives to withhold claims for manipulative purposes and may establish disincentives to *492present claims when evidence is fresh. Reed v. Ross, supra, at 13; Wainwright v. Sykes, 433 U. S., at 89.

Far more severe are the disruptions when a claim is presented for the first time in a second or subsequent federal ha-beas petition. If “[collateral review of a conviction extends the ordeal of trial for both society and the accused,” Engle v. Isaac, supra, at 126-127, the ordeal worsens during subsequent collateral proceedings. Perpetual disrespect for the finality of convictions disparages the entire criminal justice system.

“A procedural system which permits an endless repetition of inquiry into facts and law in a vain search for ultimate certitude implies a lack of confidence about the possibilities of justice that cannot but war with the effectiveness of underlying substantive commands. . . . There comes a point where a procedural system which leaves matters perpetually open no longer reflects humane concern but merely anxiety and a desire for immobility.” Bator, 76 Harv. L. Rev., at 452-453 (footnotes omitted).

If reexamination of a conviction in the first round of federal habeas stretches resources, examination of new claims raised in a second or subsequent petition spreads them thinner still. These later petitions deplete the resources needed for federal litigants in the first instance, including litigants commencing their first federal habeas action. The phenomenon calls to mind Justice Jackson’s admonition that “[i]t must prejudice the occasional meritorious application to be buried in a flood of worthless ones.” Brown v. Allen, 344 U. S., at 537 (opinion concurring in result). And if reexamination of convictions in the first round of habeas offends federalism and comity, the offense increases when a State must defend its conviction in a second or subsequent habeas proceeding on grounds not even raised in the first petition.

The federal writ of habeas corpus overrides all these considerations, essential as they are to the rule of law, when a petitioner raises a meritorious constitutional claim in a *493proper manner in a habeas petition. Our procedural default jurisprudence and abuse-of-the-writ jurisprudence help define this dimension of procedural regularity. Both doctrines impose on petitioners a burden of reasonable compliance with procedures designed to discourage baseless claims and to keep the system open for valid ones; both recognize the law’s interest in finality; and both invoke equitable principles to define the court’s discretion to excuse pleading and procedural requirements for petitioners who could not comply with them in the exercise of reasonable care and diligence. It is true that a habeas court’s concern to honor state procedural default rules rests in part on respect for the integrity of procedures “employed by a coordinate jurisdiction within the federal system,” Wainwright v. Sykes, supra, at 88, and that such respect is not implicated when a petitioner defaults a claim by failing to raise it in the first round of federal habeas review. Nonetheless, the doctrines of procedural default and abuse of the writ are both designed to lessen the injury to a State that results through reexamination of a state conviction on a ground that the State did not have the opportunity to address at a prior, appropriate time; and both doctrines seek to vindicate the State’s interest in the finality of its criminal judgments.

We conclude from the unity of structure and purpose in the jurisprudence of state procedural defaults and abuse of the writ that the standard for excusing a failure to raise a claim at the appropriate time should be the same in both contexts. We have held that a procedural default will be excused upon a showing of cause and prejudice. Wainwright v. Sykes, supra. We now hold that the same standard applies to determine if there has been an abuse of the writ through inexcusable neglect.

In procedural default cases, the cause standard requires the petitioner to show that “some objective factor external to the defense impeded counsel’s efforts” to raise the claim in state court. Murray v. Carrier, 477 U. S., at 488. Objec*494tive factors that constitute cause include “ ‘interference by officials’” that makes compliance with the State’s procedural rule impracticable, and “a showing that the factual or legal basis for a claim was not reasonably available to counsel.” Ibid. In addition, constitutionally “[ineffective assistance of counsel... is cause.” Ibid. Attorney error short of ineffective assistance of counsel, however, does not constitute cause and will not excuse a procedural default. Id., at 486-488. Once the petitioner has established cause, he must show “ ‘actual prejudice’ resulting from the errors of which he complains.” United States v. Frady, 456 U. S. 152, 168 (1982).

Federal courts retain the authority to issue the writ of habeas corpus in a further, narrow class of cases despite a petitioner’s failure to show cause for a procedural default. These are extraordinary instances when a constitutional violation probably has caused the conviction of one innocent of the crime. We have described this class of cases as implicating a fundamental miscarriage of justice. Murray v. Carrier, supra, at 485.

The cause and prejudice analysis we have adopted for cases of procedural default applies to an abuse-of-the-writ inquiry in the following manner. When a prisoner files a second or subsequent application, the government bears the burden of pleading abuse of the writ. The government satisfies this burden if, with clarity and particularity, it notes petitioner’s prior writ history, identifies the claims that appear for the first time, and alleges that petitioner has abused the writ. The burden to disprove abuse then becomes petitioner’s. To excuse his failure to raise the claim earlier, he must show cause for failing to raise it and prejudice therefrom as those concepts have been defined in our procedural default decisions. The petitioner’s opportunity to meet the burden of cause and prejudice will not include an evidentiary hearing if the district court determines as a matter of law that petitioner cannot satisfy the standard. If petitioner cannot show cause, the failure to raise the claim in an earlier petition may *495nonetheless be excused if he or she can show that a fundamental miscarriage of justice would result from a failure to entertain the claim. Application of the cause and prejudice standard in the abuse-of-the-writ context does not mitigate the force of Teague v. Lane, 489 U. S. 288 (1989), which prohibits, with certain exceptions, the retroactive application of new law to claims raised in federal habeas. Nor does it imply that there is a constitutional right to counsel in federal habeas corpus. See Pennsylvania v. Finley, 481 U. S. 551, 555 (1987) (“[T]he right to appointed counsel extends to the first appeal of right, and no further”).

Although the cause and prejudice standard differs from some of the language in Price v. Johnston, 334 U. S. 266 (1948), it is consistent with Cuddy, Salinger, Wong Doo, and Sanders, as well as our modern abuse-of-the-writ decisions, including Antone, Woodard, and Delo. In addition, the exception to cause for fundamental miscarriages of justice gives meaningful content to the otherwise unexplained “ends of justice” inquiry mandated by Sanders. Sanders drew the phrase “ends of justice” from the 1948 version of § 2244. 28 U. S. C. §2244 (1964 ed.) (judge need not entertain subsequent application if he is satisfied that “the ends of justice will not be served by such inquiry”). Sanders v. United States, 373 U. S., at 15-17. Although the 1966 revision to the habeas statute eliminated any reference to an “ends of justice” inquiry, a plurality of the Court in Kuhlmann v. Wilson, 477 U. S., at 454, held that this inquiry remained appropriate, and required federal courts to entertain successive petitions when a petitioner supplements a constitutional claim with a “colorable showing of factual innocence.” The miscarriage of justice exception to cause serves as “an additional safeguard against compelling an innocent man to suffer an unconstitutional loss of liberty,” Stone v. Powell, 428 U. S., at 492-493, n. 31, guaranteeing that the ends of justice will be served in full.

*496Considerations of certainty and stability in our discharge of the judicial function support adoption of the cause and prejudice standard in the abuse-of-the-writ context. Well defined in the case law, the standard will be familiar to federal courts. Its application clarifies the imprecise contours of the term “inexcusable neglect.” The standard is an objective one, and can be applied in a manner that comports with the threshold nature of the abuse-of-the-writ inquiry. See Price v. Johnston, 334 U. S., at 287 (abuse of the writ is “preliminary as well as collateral to a decision as to the sufficiency or merits of the allegation itself”). Finally, the standard provides “a sound and workable means of channeling the discretion of federal habeas courts.” Murray v. Carrier, 477 U. S., at 497. “[I]t is important, in order to preclude individualized enforcement of the Constitution in different parts of the Nation, to lay down as specifically as the nature of the problem permits the standards or directions that should govern the District Judges in the disposition of applications for habeas corpus by prisoners under sentence of State Courts.” Brown v. Allen, 344 U. S., at 501-502 (opinion of Frankfurter, J.).

The cause and prejudice standard should curtail the abusive petitions that in recent years have threatened to undermine the integrity of the habeas corpus process. “Federal courts should not continue to tolerate — even in capital cases — this type of abuse of the writ of habeas corpus.” Woodard v. Hutchins, 464 U. S., at 380. The writ of habeas corpus is one of the centerpieces of our liberties. “But the writ has potentialities for evil as well as for good. Abuse of the writ may undermine the orderly administration of justice and therefore weaken the forces of authority that are essential for civilization.” Brown v. Allen, supra, at 512 (opinion of Frankfurter, J.). Adoption of the cause and prejudice standard acknowledges the historic purpose and function of the writ in our constitutional system, and, by preventing its abuse, assures its continued efficacy.

*497We now apply these principles to the case before us.

>

McCleskey based the Massiah claim in his second federal petition on the 21-page Evans document alone. Worthy’s identity did not come to light until the hearing. The District Court found, based on the document’s revelation of the tactics used by Evans in engaging McCleskey in conversation (such as his pretending to be Ben Wright’s uncle and his claim that he was supposed to participate in the robbery), that the document established an ab initio relationship between Evans and the authorities. It relied on the finding and on Worthy’s later testimony to conclude that the State committed a Massiah violation.

This ruling on the merits cannot come before us or any federal court if it is premised on a claim that constitutes an abuse of the writ. We must consider, therefore, the preliminary question whether McCleskey had cause for failing to raise the Massiah claim in his first federal petition. The District Court found that neither the 21-page document nor Worthy were known or discoverable before filing the first federal petition. Relying on these findings, McCleskey argues that his failure to raise the Massiah claim in the first petition should be excused. For reasons set forth below, we disagree.

That McCleskey did not possess, or could not reasonably have obtained, certain evidence fails to establish cause if other known or discoverable evidence could have supported the claim in any event. “[C]ause . . . requires a showing of some external impediment preventing counsel from constructing or raising the claim.” Murray v. Carrier, 477 U. S., at 492 (emphasis added). For cause to exist, the external impediment, whether it be government interference or the reasonable unavailability of the factual basis for the claim, must have prevented petitioner from raising the claim. See id., at 488 (cause if “interference by officials . . . made compliance *498impracticable”); Amadeo v. Zant, 486 U. S. 214, 222 (1988) (cause if unavailable evidence “was the reason” for default). Abuse-of-the-writ doctrine examines petitioner's conduct: The question is whether petitioner possessed, or by reasonable means could have obtained, a sufficient basis to allege a claim in the first petition and pursue the matter through the habeas process, see 28 U. S. C. §2254 Rule 6 (Discovery); Rule 7 (Expansion of Record); Rule 8 (Evidentiary Hearing). The requirement of cause in the abuse-of-the-writ context is based on the principle that petitioner must conduct a reasonable and diligent investigation aimed at including all relevant claims and grounds for relief in the first federal habeas petition. If what petitioner knows or could discover upon reasonable investigation supports a claim for relief in a federal habeas petition, what he does not know is irrelevant. Omission of the claim will not be excused merely because evidence discovered later might also have supported or strengthened the claim.

In applying these principles, we turn first to the 21-page signed statement. It is essential at the outset to distinguish between two issues: (1) Whether petitioner knew about or could have discovered the 21-page document; and (2) whether he knew about or could have discovered the evidence the document recounted, namely, the jail-cell conversations. The District Court’s error lies in its conflation of the two inquiries, an error petitioner would have us perpetuate here.

The 21-page document unavailable to McCleskey at the time of the first petition does not establish that McCleskey had cause for failing to raise the Massiah claim at the outset.* Based on testimony and questioning at trial, McCles-*499key knew that he had confessed the murder during jail-cell conversations with Evans, knew that Evans claimed to be a relative of Ben Wright during the conversations, and knew that Evans told the police about the conversations. Knowledge of these facts alone would put McCleskey on notice to pursue the Massiah claim in his first federal habeas petition as he had done in the first state habeas petition.

But there was more. The District Court’s finding that the 21-page document established an ab initio relationship between Evans and the authorities rested in its entirety on conversations in which McCleskey himself participated. *500Though at trial MeCleskey denied the inculpatory conversations, his current arguments presuppose them. Quite apart from the inequity in McCleskey’s reliance on that which he earlier denied under oath, the more fundamental point remains that because MeCleskey participated in the conversations reported by Evans, he knew everything in the document that the District Court relied upon to establish the ab initio connection between Evans and the police. MeCleskey has had at least constructive knowledge all along of the facts he now claims to have learned only from the 21-page document. The unavailability of the document did not prevent MeCleskey from raising the Massiah claim in the first federal petition and is not cause for his failure to do so. And of course, MeCleskey cannot contend that his false representations at trial constitute cause for the omission of a claim from the first federal petition.

The District Court’s determination that jailer Worthy’s identity and testimony could not have been known prior to the first federal petition does not alter our conclusion. It must be remembered that the 21-page statement was the only new evidence MeCleskey had when he filed the Massiah claim in the second federal petition in 1987. Under McCles-key’s own theory, nothing was known about Worthy even then. If MeCleskey did not need to know about Worthy and his testimony to press the Massiah claim in the second petition, neither did he need to know about him to assert it in the first. Ignorance about Worthy did not prevent MeCleskey from raising the Massiah claim in the first federal petition and will not excuse his failure to do so.

Though this reasoning suffices to show the irrelevance of the District Court’s finding concerning Worthy, the whole question illustrates the rationale for requiring a prompt investigation and the full pursuit of habeas claims in the first petition. At the time of the first federal petition, written logs and records with prison staff names and assignments existed. By the time of the second federal petition officials had *501destroyed the records pursuant to normal retention schedules. Worthy’s inconsistent and confused testimony in this case demonstrates the obvious proposition that factfinding processes are impaired when delayed. Had McCleskey presented this claim in the first federal habeas proceeding when official records were available, he could have identified the relevant officers and cell assignment sheets. The critical facts for the Massiah claim, including the reason for Evans’ placement in the cell adjacent to McCleskey’s and the precise conversation that each officer had with Evans before he was put there, likely would have been reconstructed with greater precision than now can be achieved. By failing to raise the Massiah claim in 1981, McCleskey foreclosed the procedures best suited for disclosure of the facts needed for a reliable determination.

McCleskey nonetheless seeks to hold the State responsible for his omission of the Massiah claim in the first petition. His current strategy is to allege that the State engaged in wrongful conduct in withholding the 21-page document. This argument need not detain us long. When all is said and done, the issue is not presented in the case, despite all the emphasis upon it in McCleskey’s brief and oral argument. The Atlanta police turned over the 21-page document upon request in 1987. The District Court found no misrepresentation or wrongful conduct by the State in failing to hand over the document earlier, and our discussion of the evidence in the record concerning the existence of the statement, see n., swpra, as well as the fact that at least four courts have considered and rejected petitioner’s Brady claim, belies McCles-key’s characterization of the case. And as we have taken care to explain, the document is not critical to McCleskey’s notice of a Massiah claim anyway.

Petitioner’s reliance on the procedural default discussion in Amadeo v. Zant, 486 U. S. 214 (1988), is misplaced. In Amadeo the Court mentioned that government concealment of evidence could be cause for a procedural default if it “was *502the reason for the failure of a petitioner’s lawyers to raise the jury challenge in the trial court.” Id., at 222. This case differs from Amadeo in two crucial respects. First, there is no finding that the State concealed evidence. And second, even if the State intentionally concealed the 21-page document, the concealment would not establish cause here because, in light of McCleskey’s knowledge of the information in the document, any initial concealment would not have prevented him from raising the claim in the first federal petition.

As McCleskey lacks cause for failing to raise the Massiah claim in the first federal petition, we need not consider whether he would be prejudiced by his inability to raise the alleged Massiah violation at this late date. See Murray v. Carrier, 477 U. S., at 494 (rejecting proposition that showing of prejudice permits relief in the absence of cause).

We do address whether the Court should nonetheless exercise its equitable discretion to correct a miscarriage of justice. That narrow exception is of no avail to McCleskey. The Massiah violation, if it be one, resulted in the admission at trial of truthful inculpatory evidence which did not affect the reliability of the guilt determination. The very statement McCleskey now seeks to embrace confirms his guilt. As the District Court observed:

“After having read [the Evans statement], the court has concluded that nobody short of William Faulkner could have contrived that statement, and as a consequence finds the testimony of Offie Evans absolutely to be true, and the court states on the record that it entertains absolutely no doubt as to the guilt of Mr. McCleskey.” 4 Tr. 4.

We agree with this conclusion. McCleskey cannot demonstrate that the alleged Massiah violation caused the conviction of an innocent person. Murray v. Carrier, supra, at 496.

The history of the proceedings in this case, and the burden upon the State in defending against allegations made for the *503first time in federal court some nine years after the trial, reveal the necessity for the abuse-of-the-writ doctrine. The cause and prejudice standard we adopt today leaves ample room for consideration of constitutional errors in a first federal habeas petition and in a later petition under appropriate circumstances. Petitioner has not satisfied this standard for excusing the omission of the Massiah claim from his first petition. The judgment of the Court of Appeals is

Affirmed.

APPENDIX TO OPINION OF THE COURT

Petitioner’s Claims for Relief at Various Stages of the Litigation

1. Direct Appeal. On direct appeal, McCleskey raised the following claims: (1) the death penalty was administered in a discriminatory fashion because of prosecutorial discretion; (2) the prosecutor conducted an illegal postindictment lineup; (3) the trial court erred in admitting at trial the statement McCleskey made to the police; (4) the trial court erred in allowing Evans to testify about McCleskey’s jail-house confession; (5) the prosecutor failed to disclose certain impeachment evidence; and (6) the trial court erred in admitting evidence of McCleskey’s prior criminal acts. McClesky v. State, 245 Ga. 108, 112-114, 263 S. E. 2d 146, 149-151 (1980).

2. First State Habeas Corpus Petition. McCleskey’s first state habeas petition alleged the following constitutional violations: (1) the Georgia death penalty is administered arbitrarily, capriciously, and whimsically; (2) Georgia officials imposed McCleskey’s capital sentence pursuant to a pattern and practice of discrimination on the basis of race, sex, and poverty; (3) the death penalty lacks theoretical or factual justification and fails to serve any rational interest; (4) McCleskey’s death sentence is cruel and unusual punishment in light of all mitigating factors; (5) McCleskey received inadequate notice and opportunity to be heard; (6) the jury did not constitute a fair cross section of the community; (7) the jury was biased *504in favor of the prosecution; (8) the trial court improperly excused two jurors who were opposed to the death penalty; (9) McCleskey’s postarrest statement should have been excluded because it was obtained after an allegedly illegal arrest; (10) the postarrest statement was extracted involuntarily; (11) the State failed to disclose an “arrangement” with one of its key witnesses, Evans; (12) the State deliberately withheld a statement made by McCleskey to Evans; (13) the trial court erred in failing to grant McCleskey funds to employ experts in aid of his defense; (14) three witnesses for the State witnessed a highly suggestive lineup involving McCleskey prior to trial; (15) the trial court’s jury instructions concerning intent impermissibly shifted the burden of persuasion to McCleskey; (16) the prosecution impermissibly referred to the appellate process during the sentencing phase; (17) the trial court improperly admitted evidence of other crimes for which McCleskey had not been convicted; (18) the trial court’s instructions concerning evidence of McCleskey’s other bad acts was overbroad; (19) the appellate review procedures of Georgia denied McCleskey effective assistance of counsel, a fair hearing, and the basic tools of an adequate defense; (20) the means by which the death penalty is administered inflicts wanton and unnecessary torture; (21) McCleskey was denied effective assistance of counsel in numerous contexts; (22) introduction of statements petitioner made to Evans were elicited in a situation created to induce McCleskey to make incriminating statements; and (23) the evidence was insufficient to convict McCleskey of capital murder. Petition, HC No. 4909, 2 Tr., Exh. H.

3. First Federal Habeas Corpus Petition. McCleskey raised the following claims in his first federal habeas petition: (1) the Georgia death penalty discriminated on the basis of race; (2) the State failed to disclose an “understanding” with Evans; (3) the trial court’s instructions to the jury impermis-sibly shifted the burden to McCleskey; (4) the prosecutor improperly referred to the appellate process at the sentencing *505phase; (5) the trial court impermissibly refused to grant Mc-Cleskey funds to employ experts in aid of his defense; (6) the trial court’s instructions concerning evidence of McCleskey’s other bad acts was overbroad; (7) the trial court’s instructions gave the jury too much discretion to consider nonstatu-tory aggravating circumstances; (8) the trial court improperly admitted evidence of other crimes for which McCleskey had not been convicted; (9) three witnesses for the State witnessed a highly suggestive lineup involving McCleskey prior to trial; (10) McCleskey’s postarrest statement should have been excluded because it was extracted involuntarily; (11) the trial court impermissibly excluded two jurors who were opposed to the death penalty; (12) the death penalty lacks theoretical or factual justification and fails to serve any rational interest; (13) the State deliberately withheld a statement made by McCleskey to Evans; (14) the evidence was insufficient to convict McCleskey of capital murder; (15) McCles-key’s counsel failed to investigate the State’s evidence adequately; (16) McCleskey’s counsel failed to raise certain objections or make certain motions at trial; (17) McCleskey’s counsel failed to undertake an independent investigation of possible mitigating circumstances prior to trial; and (18) after trial, McCleskey’s counsel failed to review and correct the judge’s sentence report. McCleskey v. Zant, 580 F. Supp. 338 (ND Ga. 1984).

4. Second State Habeas Petition. In his second state ha-beas petition, McCleskey alleged the following claims: (1) the prosecutor systematically excluded blacks from the jury; (2) the State of Georgia imposed the death penalty against McCleskey in a racially discriminatory manner; (3) the State failed to disclose its agreement with Evans; (4) the trial court impermissibly refused to grant McCleskey funds to employ experts in aid of his defense; and (5) the prosecutor improperly referred to the appellate process at the sentencing phase. Petition, 2 Tr., Exh. G.

*5065. Second Federal Habeas Corpus Petition. In his second federal habeas petition, McCleskey alleged the following claims: (1) Evans’ testimony concerning his conversation with McCleskey was inadmissible because Evans acted as a state informant in a situation created to induce McCleskey to make incriminating statements; (2) the State failed to correct the misleading testimony of Evans; (3) the State failed to disclose “an arrangement” with Evans; (4) the prosecutor improperly referred to the appellate process at the sentencing phase; (5) the State systematically excluded blacks from McCleskey’s jury; (6) the death penalty was imposed on McCleskey pursuant to a pattern and practice of racial discrimination by Georgia officials against black defendants; and (7) the trial court impermissibly refused to grant McCleskey funds to employ experts in aid of his defense. Federal Habeas Petition, 1 Tr., Exh. 1.

with whom Justice Blackmun and Justice Stevens join, dissenting.

Today’s decision departs drastically from the norms that inform the proper judicial function. Without even the most casual admission that it is discarding longstanding legal principles, the Court radically redefines the content of the “abuse of the writ” doctrine, substituting the strict-liability “cause and prejudice” standard of Wainwright v. Sykes, 433 U. S. 72 (1977), for the good-faith “deliberate abandonment” standard of Sanders v. United States, 373 U. S. 1 (1963). This doctrinal innovation, which repudiates a line of judicial decisions codified by Congress in the governing statute and procedural rules, was by no means foreseeable when the petitioner in this case filed his first federal habeas application. Indeed, the new rule announced and applied today was not even requested by respondent at any point in this litigation. Finally, rather than remand this case for reconsideration in light of its new standard, the majority performs an independent reconstruction of the record, disregarding the factual findings of the District Court and applying its new rule in a *507manner that encourages state officials to conceal evidence that would likely prompt a petitioner to raise a particular claim on habeas. Because I cannot acquiesce in this unjustifiable assault on the Great Writ, I dissent.

hH

Disclaiming innovation, the majority depicts the “cause and prejudice” test as merely a clarification of existing law. Our decisions, the majority explains, have left “[m]uch confusion ... on the standard for determining when a petitioner abuses the writ.” Ante, at 477. But amidst this “confusion,” the majority purports to discern a trend toward the cause-and-prejudice standard and concludes that this is the rule that best comports with “our habeas corpus precedents,” ante, at 490; see ante, at 495, and with the “complex and evolving body of equitable principles” that have traditionally defined the abuse-of-the-writ doctrine, id., at 489. This attempt to gloss over the break between today’s decision and established precedents is completely unconvincing.

Drawing on the practice at common law in England, this Court long ago established that the power of a federal court to entertain a second or successive petition should turn not on “the inflexible doctrine of res judicata” but rather on the exercise of “sound judicial discretion guided and controlled by a consideration of whatever has a rational bearing on the subject.” Wong Doo v. United States, 265 U. S. 239, 240-241 (1924); accord, Salinger v. Loisel, 265 U. S. 224, 230-232 (1924). Thus, in Wong Doo, the Court held that the District Court acted within its discretion in dismissing a petition premised on a ground that was raised but expressly abandoned in an earlier petition. “The petitioner had full opportunity,” the Court explained, “to offer proof [of the abandoned ground] at the hearing on the first petition; and, if he was intending to rely on that ground, good faith required that he produce the proof then.” 265 U. S., at 241. Noting that the evidence supporting the abandoned ground had been “ac*508cessible all the time,” the Court inferred that petitioner, an alien seeking to forestall his imminent deportation, had split his claims in order to “postpone the execution of the [deportation] order.” Ibid.

In Price v. Johnston, 334 U. S. 266 (1948), in contrast, the Court held that the District Court abused its discretion by summarily dismissing a petition that raised a claim not asserted in any of three previous petitions filed by the same prisoner. Whereas it had been clear from the record that the petitioner in Wong Doo had possessed access to the facts supporting his abandoned claim, the District Court in Price had no basis for assuming that the prisoner had “acquired no new or additional information since” the disposition of his earlier petitions. Id., at 290. “[E]ven if it [had been] found that petitioner did have prior knowledge of all the facts concerning the allegation in question,” the Court added, the District Court should not have dismissed the petition before affording the prisoner an opportunity to articulate “some justifiable reason [why] he was previously unable to assert his rights or was unaware of the significance of relevant facts.” Id., at 291.

In Sanders v. United States, 373 U. S. 1 (1963), the Court crystallized the various factors bearing on a district court’s discretion to entertain a successive petition.1 The Court in Sanders distinguished successive petitions raising previously asserted grounds from those raising previously unasserted grounds. With regard to the former class of petitions, the Court explained, the district court may give “[c]ontrolling weight... to [the] denial of a prior application” unless “the ends of justice would ... be served by reaching the merits of the subsequent application.” Id., at 16. With regard to the *509latter, however, the district court must reach the merits of the petition unless “there has been an abuse of the writ . . . Id., at 17. In determining whether the omission of the claim from the previous petition constitutes an abuse of the writ, the judgment of the district court is to be guided chiefly by the “ ‘[equitable] principle that a suitor’s conduct in relation to the matter at hand may disentitle him to the relief he seeks.’” Ibid., quoting Fay v. Noia, 372 U. S. 391, 438 (1963).

“Thus, for example, if a prisoner deliberately withholds one of two grounds for federal collateral relief at the time of filing his first application, in the hope of being granted two hearings rather than one or for some other such reason, he may be deemed to have waived his right to a hearing on a second application presenting the withheld ground. The same may be true if, as in Wong Doo, the prisoner deliberately abandons one of his grounds at the first hearing. Nothing in the traditions of habeas corpus requires the federal courts to tolerate needless piecemeal litigation, or to entertain collateral proceedings whose only purpose is to vex, harass, or delay.” 373 U. S., at 18.

What emerges from Sanders and its predecessors is essentially a good-faith standard. As illustrated by Wong Doo, the principal form of bad faith that the “abuse of the writ” doctrine is intended to deter is the deliberate abandonment of a claim the factual and legal basis of which are known to the petitioner (or his counsel) when he files his first petition. The Court in Sanders stressed this point by equating its analysis with that of Fay v. Noia, supra, which established the then-prevailing “deliberate bypass” test for the cognizability of claims on which a petitioner procedurally defaulted in state proceedings. See 373 U. S., at 18. A petitioner also abuses the writ under Sanders when he uses the writ to achieve some end other than expeditious relief from unlawful confinement-such as “to vex, harass, or delay.” However, so long *510as the petitioner’s previous application was based on a good-faith assessment of the claims available to him, see Price v. Johnston, supra, at 289; Wong Doo, supra, at 241; the denial of the application does not bar the petitioner from availing himself of “new or additional information,” Price v. Johnston, supra, at 290, in support of a claim not previously raised. Accord, Advisory Committee’s Note to Habeas Corpus Rule 9, 28 U. S. C., p. 427.

“Cause and prejudice” — the standard currently applicable to procedural defaults in state proceedings, see Wainwright v. Sykes, 433 U. S. 72 (1977) — imposes a much stricter test. As this Court’s precedents make clear, a petitioner has cause for failing effectively to present his federal claim in state proceedings only when “some objective factor external to the defense impeded counsel’s efforts to comply with the State’s procedural rule . . . .” Murray v. Carrier, 477 U. S. 478, 488 (1986). Under this test, the state of mind of counsel is largely irrelevant. Indeed, this Court has held that even counsel’s reasonable perception that a particular claim is without factual or legal foundation does not excuse the failure to raise that claim in the absence of an objective, external impediment to counsel’s efforts. See Smith v. Murray, 477 U. S. 527, 535-536 (1986). In this sense, the cause component of the Wainwright v. Sykes test establishes a strict-liability standard.2

*511Equally foreign to our abuse-of-the-writ jurisprudence is the requirement that a petitioner show “prejudice.” Under Sanders, a petitioner who articulates a justifiable reason for failing to present a claim in a previous habeas application is not required in addition to demonstrate any particular degree of prejudice before the habeas court must consider his claim. If the petitioner demonstrates that his claim has merit, it is the State that must show that the resulting constitutional error was harmless beyond a reasonable doubt. See L. Yackle, Postconviction Remedies § 133, p. 503 (1981).3

*512HH HH

The real question posed by the majority s analysis is not whether the cause-and-prejudice test departs from the principles of Sanders — for it clearly does — but whether the majority has succeeded in justifying this departure as an exercise of this Court’s common-lawmaking discretion. In my view, the majority does not come close to justifying its new standard.

A

Incorporation of the cause-and-prejudice test into the abuse-of-the-writ doctrine cannot be justified as an exercise of this Court’s common-lawmaking discretion, because this Court has no discretion to exercise in this area. Congress has affirmatively ratified the Sanders good-faith standard in the governing statute and procedural rules, thereby insulating that standard from judicial repeal.

The abuse-of-the-writ doctrine is embodied in 28 U. S. C. § 2244(b) and in Habeas Corpus Rule 9(b). Enacted three years after Sanders, § 2244(b) recodified the statutory authority of a district court to dismiss a second or successive petition, amending the statutory language to incorporate the Sanders criteria:

“[A] subsequent application for a writ of habeas corpus . . . need not be entertained by a court . . . unless the application alleges and is predicated on a factual or other ground not adjudicated on the hearing of the earlier application for the writ, and unless the court... is satisfied that the applicant has not on the earlier application deliberately withheld the newly asserted ground or otherwise abused the writ.” 28 U. S. C. § 2244(b).

Consistent with Sanders, the purpose of the recodification was to spare a district court the obligation to entertain a petition “containing allegations identical to those asserted in a previous application that has been denied, or predicated upon grounds obviously well known to [the petitioner] when [he] *513filed the preceding application.” S. Rep. No. 1797, 89th Cong., 2d Sess., 2 (1966) (emphasis added). Rule 9(b) likewise adopts Sanders’ terminology:

“A second or successive petition may be dismissed if the judge finds that it fails to allege new or different grounds for relief and the prior determination was on the merits or, if new and different grounds are alleged, the judge finds that the failure of the petitioner to assert those grounds in a prior petition constituted an abuse of the writ.”

There can be no question that § 2244(b) and Rule 9(b) codify Sanders. The legislative history of, and Advisory Committee’s Notes to, Rule 9(b) expressly so indicate, see 28 U. S. C., pp. 426-427; H. R. Rep. No. 94-1471, pp. 5-6 (1976), and such has been the universal understanding of this Court, see Rose v. Lundy, 455 U. S. 509, 521 (1982), of the lower courts, see, e. g., Williams v. Lockhart, 862 F. 2d 155, 157 (CA8 1988); Neuschafer v. Whitley, 860 F. 2d 1470, 1474 (CA9 1988), cert. denied, sub nom. Demosthenes v. Neuschafer, 493 U. S. 906 (1989); 860 F. 2d, at 1479 (Alarcon, J., con curring in result); Davis v. Dugger, 829 F. 2d 1513, 1518, n. 13 (CA11 1987); Passman v. Blackburn, 797 F. 2d 1335, 1341 (CA5 1986), cert. denied, 480 U. S. 948 (1987); United States v. Talk, 597 F. 2d 249, 250-251 (CA10 1979); United States ex rel. Fletcher v. Brierley, 460 F. 2d 444, 446, n. 4A (CA3), cert. denied, 409 U. S. 1044 (1972), and of commentators, see, e. g., 17A C. Wright, A. Miller & E. Cooper, Federal Practice and Procedure §4267, pp. 477-478 (2d ed. 1988); L. Yackle, supra, § 154.4

*514The majority concedes that § 2244(b) and Rule 9(b) codify Sanders, see ante, at 487, but concludes nonetheless that Congress did “not answer” all of the “questions” concerning the abuse-of-the-writ doctrine, ibid. The majority emphasizes that § 2244(b) refers to second or successive petitions from petitioners who have “deliberately withheld the newly asserted ground ... or otherwise abused the writ” without exhaustively cataloging the ways in which the writ may “otherwise” be “abused.” See ante, at 486, 489-490. From this “silenc[e],” the majority infers a congressional delegation of lawmaking power broad enough to encompass the engrafting of the cause-and-prejudice test onto the abuse-of-the-writ doctrine. Ante, at 487.

It is difficult to take this reasoning seriously. Because “cause” under Sykes makes the mental state of the petitioner (or his counsel) irrelevant, “cause” completely subsumes “deliberate abandonment.” See Engle v. Isaac, 456 U. S. 107, 130, n. 36 (1982); see also Wainwright v. Sykes, 433 U. S., at 87. Thus, if merely failing to raise a claim without “cause”— that is, without some external impediment to raising it — necessarily constitutes an abuse of the writ, the statutory reference to deliberate withholding of a claim would be rendered superfluous. Insofar as Sanders was primarily concerned with limiting dismissal of a second or subsequent petition to instances in which the petitioner had deliberately abandoned the new claim, see 373 U. S., at 18, the suggestion that Congress invested courts with the discretion to read this language out of the statute is completely irreconcilable with the proposition that § 2244(b) and Rule 9(b) codify Sanders.

To give content to “otherwise abus[e] the writ” as used in § 2244(b), we must look to Sanders. As I have explained, *515the Court in Sanders identified two broad classes of bad-faith conduct that bar adjudication of a claim not raised in a previous habeas application: the deliberate abandonment or withholding of that claim from the first petition; and the filing of a petition aimed at some purpose other than expeditious relief from unlawful confinement, such as “to vex, harass, or delay.” See ibid. By referring to second or successive applications from habeas petitioners who have “deliberately withheld the newly asserted ground or otherwise abused the writ,” § 2244(b) tracks this division. Congress may well have selected the phrase “otherwise abused the writ” with the expectation that courts would continue to elaborate upon the types of dilatory tactics that, in addition to deliberate abandonment of a known claim, constitute an abuse of the writ. But consistent with Congress’ intent to codify Sanders’ good-faith test, such elaborations must be confined to circumstances in which a petitioner’s omission of an unknown claim is conjoined with his intentional filing of a petition for an improper purpose, such as “to vex, harass or delay.”

The majority tacitly acknowledges this constraint on the Court’s interpretive discretion by suggesting that “cause” is tantamount to “inexcusable neglect.” This claim, too, is untenable. The majority exaggerates when it claims that the “inexcusable neglect” formulation — which this Court has never applied in an abuse-of-the-writ decision — functions as an independent standard for evaluating a petitioner’s failure to raise a claim in a previous habeas application. It is true that Sanders compared its own analysis to the analysis in Townsend v. Sain, 372 U. S. 293 (1963), which established that a district court should deny an evidentiary hearing if the habeas petitioner inexcusably neglected to develop factual evidence in state proceedings. See id., at 317. Townsend, however, expressly equated “inexcusable neglect” with the “deliberate bypass” test of Fay v. Noia. See 372 U. S., at *516317.5 But even if “inexcusable neglect” does usefully describe a class of abuses separate from deliberate abandonment, the melding of “cause and prejudice” into the abuse-of-the-writ doctrine cannot be defended as a means of “giving content” to “inexcusable neglect.” Ante, at 490. For under Sykes’ strict-liability standard, mere attorney negligence is never excusable. See Murray v. Carrier, 477 U. S., at 488 (“So long as a defendant is represented by counsel whose performance is not constitutionally ineffective ... , we discern no inequity in requiring him to bear the risk of attorney error that results in a procedural default”).

Confirmation that the majority today exercises legislative power not properly belonging to this Court is supplied by Congress’ own recent consideration and rejection of an amendment to § 2244(b). It is axiomatic that this Court does not function as a backup legislature for the reconsideration of failed attempts to amend existing statutes. See Bowsher v. Merck & Co., 460 U. S. 824, 837, n. 12 (1983); FTC v. Ruberoid Co., 343 U. S. 470, 478-479 (1952); see also North Haven Bd. of Ed. v. Bell, 456 U. S. 512, 534-535 (1982). Yet that is exactly the effect of today’s decision. As reported out of the House Committee on the Judiciary, § 1303 of H. R. 5269, 101st Cong., 2d Sess. (1990), would have required dismissal of any second or subsequent application by a habeas petitioner under sentence of death unless the peti*517tioner raised a new claim “the factual basis of [which] could not have been discovered by the exercise of reasonable diligence,” H. R. Rep. No. 101-681, pt. 1, p. 29 (1990) (emphasis added).6 The Committee Report accompanying this legislation explained that “courts have properly construed section 2244(b) and Rule 9(b) as codifications of the guidelines the [Supreme] Court itself prescribed in Sanders” Id., at 119 (citation omitted). The Report justified adoption of the tougher “reasonable diligence” standard on the ground that “[t]he Sanders guidelines have not. . . satisfactorily met concerns that death row prisoners may file second or successive habeas corpus applications as a means of extending litigation.” Ibid. Unfazed by Congress’ rejection of this legislation, the majority arrogates to itself the power to repeal Sanders and to replace it with a tougher standard.7

B

Even if the fusion of cause-and-prejudice into the abuse-of-the-writ doctrine were not foreclosed by the will of Congress, the majority fails to demonstrate that such a rule would be a wise or just exercise of the Court’s common-lawmaking discretion. In fact, the majority’s abrupt change in law subverts the policies underlying § 2244(b) and unfairly prejudices the petitioner in this case.

The majority premises adoption of the cause-and-prejudice test almost entirely on the importance of “finality.” See ante, at 490-493. At best, this is an insufficiently developed justification for cause-and-prejudice or any other possible conception of the abuse-of-the-writ doctrine. For the very *518essence of the Great Writ is our criminal justice system’s commitment to suspending “[cjonventional notions of finality of litigation . . . where life or liberty is at stake and infringement of constitutional rights is alleged.” Sanders, 373 U. S., at 8. To recognize this principle is not to make the straw-man claim that the writ must be accompanied by “‘[a] procedural system which permits an endless repetition of inquiry into facts and law in a vain search for ultimate certitude.’” Ante, at 492, quoting Bator, Finality in Criminal Law and Federal Habeas Corpus for State Prisoners, 76 Harv. L. Rev. 441, 452 (1963). Rather, it is only to point out the plain fact that we may not, “[ujnder the guise of fashioning a procedural rule, . . . wip[e] out the practical efficacy of a jurisdiction conferred by Congress on the District Courts.” Brown v. Allen, 344 U. S. 443, 498-499 (1953) (opinion of Frankfurter, J.).

The majority seeks to demonstrate that cause-and-prejudice strikes an acceptable balance between the State’s interest in finality and the purposes of habeas corpus by analogizing the abuse-of-the-writ doctrine to the procedural-default doctrine. According to the majority, these two doctrines “implicate nearly identical concerns flowing from the significant costs of federal habeas corpus review.” Ante, at 490-491. And because this Court has already deemed cause- and-prejudice to be an appropriate standard for assessing procedural defaults, the majority reasons, the same standard should be used for assessing the failure to raise a claim in a previous habeas petition. See ante, at 490-493.

This analysis does not withstand scrutiny. This Court’s precedents on the procedural-default doctrine identify two purposes served by the cause-and-prejudice test. The first purpose is to promote respect for a State’s legitimate procedural rules. See, e. g., Reed v. Ross, 468 U. S. 1, 14 (1984); Sykes, 433 U. S., at 87-90. As the Court has explained, the willingness of a habeas court to entertain a claim that a state court has deemed to be procedurally barred “undercuts] the *519State’s ability to enforce its procedural rules,” Engle v. Isaac, 456 U. S., at 129, and may cause “state courts themselves [to be] less stringent in their enforcement,” Sykes, swpra, at 89. See generally Meltzer, State Court Forfeitures of Federal Rights, 99 Harv. L. Rev. 1128, 1150-1158 (1986). The second purpose of the cause-and-prejudice test is to preserve the connection between federal collateral review and the general “deterrent” function served by the Great Writ. “ ‘[T]he threat of habeas serves as a necessary additional incentive for trial and appellate courts throughout the land to conduct their proceedings in a manner consistent with established constitutional standards.’” Teague v. Lane, 489 U. S. 288, 306 (1989) (plurality opinion), quoting Desist v. United States, 394 U. S. 244, 262-263 (1969) (Harlan, J., dissenting); see Rose v. Mitchell, 443 U. S. 545, 563 (1979). Obviously, this understanding of the disciplining effect of federal habeas corpus presupposes that a criminal defendant has given the state trial and appellate courts a fair opportunity to pass on his constitutional claims. See Murray v. Carrier, 477 U. S., at 487; Engle v. Isaac, supra, at 128-129. With regard to both of these purposes, the strictness of the cause-and-prejudice test has been justified on the ground that the defendant’s procedural default is akin to an independent and adequate state-law ground for the judgment of conviction. See Sykes, supra, at 81-83.

Neither of these concerns is even remotely implicated in the abuse-of-the-writ setting. The abuse-of-the-writ doctrine clearly contemplates a situation in which a petitioner (as in this case) has complied with applicable state-procedural rules and effectively raised his constitutional claim in state proceedings; were it otherwise, the abuse-of-the-writ doctrine would not perform a screening function independent from that performed by the procedural-default doctrine and by the requirement that a habeas petitioner exhaust his state remedies, see 28 U. S. C. §§ 2254(b), (c). Cf. ante, at 486-487. Because the abuse-of-the-writ doctrine presupposes that the *520petitioner has effectively raised his claim in state proceedings, a decision by the habeas court to entertain the claim notwithstanding its omission from an earlier habeas petition will neither breed disrespect for state-procedural rules nor unfairly subject state courts to federal collateral review in the absence of a state-court disposition of a federal claim.8

Because the abuse-of-the-writ doctrine addresses the situation in which a federal habeas court must determine whether to hear a claim withheld from another federal habeas court, the test for identifying an abuse must strike an appropriate balance between finality and review in that setting. Only when informed by Sanders does § 2244(b) strike an efficient balance. A habeas petitioner’s own interest in liberty furnishes a powerful incentive to assert in his first petition all claims that the petitioner (or his counsel) believes have a reasonable prospect for success. See Note, 83 Harv. L. Rev. 1038, 1153-1154 (1970); see also Rose v. Lundy, 455 U. S., at 520 (“The prisoner’s principal interest, of course, is in obtaining speedy federal relief on his claims”). Sanders’ bar on the later assertion of claims omitted in bad faith adequately fortifies this natural incentive. At the same time, however, the petitioner faces an effective disincentive to asserting any claim that he believes does not have a reasonable prospect for *521success: the adverse adjudication of such a claim will bar its reassertion under the successive-petition doctrine, see 28 U. S. C. § 2244(b); Sanders, supra, at 17, whereas omission of the claim will not prevent the petitioner from asserting the claim for the first time in a later petition should the discovery of new evidence or the advent of intervening changes in law invest the claim with merit, S. Rep. No. 1797, at 2; Advisory Committee’s Note to Habeas Corpus Rule 9, 28 U. S. C., p. 427.

The cause-and-prejudice test destroys this balance. By design, the cause-and-prejudice standard creates a near-irrebuttable presumption that omitted claims are permanently barred. This outcome not only conflicts with Congress’ intent that a petitioner be free to avail himself of newly discovered evidence or intervening changes in law, S. Rep. No. 1797, at 2; Advisory Committee’s Note to Habeas Corpus Rule 9, 28 U. S. C., p. 427, but also subverts the statutory disincentive to the assertion of frivolous claims. Rather than face the cause-and-prejudice bar, a petitioner will assert all conceivable claims, whether or not these claims reasonably appear to have merit. The possibility that these claims will be adversely adjudicated and thereafter be barred from relitigation under the successive-petition doctrine will not effectively discourage the petitioner from asserting them, for the petitioner will have virtually no expectation that any withheld claim could be revived should his assessment of its merit later prove mistaken. Far from promoting efficiency, the majority’s rule thus invites the very type of “baseless claims,” ante, at 493, that the majority seeks to avert.

The majority’s adoption of the cause-and-prejudice test is not only unwise, but also manifestly unfair. The proclaimed purpose of the majority’s new strict-liability standard is to increase to the maximum extent a petitioner’s incentive to investigate all conceivable claims before filing his first petition. See ante, at 498. Whatever its merits, this was not the rule when the petitioner in this case filed his first pe*522tition. From the legislative history of § 2244(b) and Rule 9(b) and from the universal agreement of courts and commentators, see supra, at 513, McCleskey’s counsel could have reached no other conclusion but that his investigatory efforts in preparing his client’s petition would be measured against the Sanders good-faith standard. There can be little question that his efforts satisfied that test; indeed, the District Court expressly concluded that McCleskey’s counsel on his first habeas conducted a reasonable and competent investigation before concluding that a claim based on Massiah v. United States, 377 U. S. 201 (1964), would be without factual foundation. See App. 84-85; see also infra, at 526. Before today, that would have been enough. The Court’s utter indifference to the injustice of retroactively applying its new, strict-liability standard to this habeas petitioner stands in marked contrast to this Court’s eagerness to protect States from the unfair surprise of “new rules” that enforce the constitutional rights of citizens charged with criminal wrongdoing. See Butler v. McKellar, 494 U. S. 407, 412-414 (1990); Saffle v. Parks, 494 U. S. 484, 488 (1990); Teague v. Lane, 489 U. S., at 299-310 (plurality opinion).

This injustice is compounded by the Court’s activism in fashioning its new rule. The applicability of Sykes’ cause- and-prejudice test was not litigated in either the District Court or the Court of Appeals. The additional question that we requested the parties to address reasonably could have been read to relate merely to the burden of proof under the abuse-of-the-writ doctrine;9 it evidently did not put the parties on notice that this Court was contemplating a change in the governing legal standard, since respondent did not even mention Sykes or cause-and-prejudice in his brief or at oral *523argument, much less request the Court to adopt this standard.10 In this respect, too, today’s decision departs from norms that inform the proper judicial function. See Heckler v. Campbell, 461 U. S. 458, 468, n. 12 (1983) (Court will consider ground in support of judgment not raised below only in extraordinary case); accord, Granfinanciera, S. A. v. Nordberg, 492 U. S. 33, 39 (1989). It cannot be said that Mc-Cleskey had a fair opportunity to challenge the reasoning that the majority today invokes to strip him of his Massiah claim.

I-H I — I I — I

The manner m which the majority applies its new rule is as objectionable as the manner in which the majority creates that rule. As even the majority acknowledges, see ante, at 470, the standard that it announces today is not the one employed by the Court of Appeals, which purported to rely on Sanders, see 890 F. 2d 342, 347 (CA11 1989). See ante, at 470. Where, as here, application of a different standard from the one applied by the lower court requires an in-depth review of the record, the ordinary course is to remand so that the parties have a fair opportunity to address, and the lower court to consider, all of the relevant issues. See, e. g., Anderson v. Liberty Lobby, Inc., 477 U. S. 242, 257 (1986); Mandel v. Bradley, 432 U. S. 173, 179 (1977) (per curiam); see also United States v. Hasting, 461 U. S. 499, 515-518 (1983) (Stevens, J., concurring in judgment) (Court should not undertake record-review “function that can better be performed by other judges”).

*524A remand would have been particularly appropriate in this case in view of the patent deficiencies in the reasoning of the Court of Appeals. The Court of Appeals concluded that McCleskey deliberately abandoned his Massiah claim because his counsel “made a knowing choice not to pursue the claim after having raised it” unsuccessfully on state collateral review. 890 F. 2d, at 349. This reasoning, which the majority declines to endorse, is obviously faulty. As I have explained, the abuse-of-the-writ doctrine is independent from the procedural-default and exhaustion doctrines; § 2244(b) and Rule 9(b) contemplate a habeas petitioner who has effectively presented his claim in state proceedings but withheld that claim from a previous habeas application. Because § 2244(b) and Rule 9(b) authorize the district court to consider such a claim under appropriate circumstances, it cannot be the case that a petitioner invariably abuses the writ by consciously failing to include in his first habeas petition a claim raised in state proceedings. Insofar as Congress intended that the district court excuse the withholding of a claim when the petitioner produces newly discovered evidence or intervening changes in law, S. Rep. No. 1797, at 2; Advisory Committee’s Note to Habeas Corpus Rule 9, 28 U. S. C., p. 427, a petitioner cannot be deemed to have deliberately abandoned the claim in an earlier habeas proceeding unless the petitioner was aware then of the evidence and law that support the claim. See, e. g., Wong Doo, 265 U. S., at 241. If the Court of Appeals had properly applied Sanders, it would almost certainly have agreed with the District Court’s conclusion that McCleskey was not aware of the evidence that supported his Massiah claim when he filed his first petition. In any case, because the Court of Appeals’ reversal was based on an erroneous application of Sanders, the majority’s decision not to remand cannot be justified on the ground that the Court of Appeals would necessarily have decided the case the same way under the cause-and-prejudice standard.

*525Undaunted by the difficulty of applying its new rule without the benefit of any lower court’s preliminary consideration, the majority forges ahead to perform its own independent review of the record. The majority concludes that McCleskey had no cause to withhold his Massiah claim because all of the evidence supporting that claim was available before he filed his first habeas petition. The majority purports to accept the District Court’s finding that Offie Evans’ 21-page statement was, at that point, being held beyond McCleskey’s reach. See ante, at 498, and n.11 But the State’s failure to produce this document, the majority explains, furnished no excuse for McCleskey’s failure to assert his Massiah claim “because McCleskey participated in the conversations reported by Evans,” and therefore “knew everything in the document that the District Court relied upon to establish the ab initio connection between Evans and the police.” Ante, at 500. The majority also points out that no *526external force impeded McCleskey’s discovery of the testimony of jailer Worthy. See ibid.

To appreciate the hollowness — and the dangerousness — of this reasoning, it is necessary to recall the District Court’s central finding: that the State did covertly plant Evans in an adjoining cell for the purpose of eliciting incriminating statements that could be used against McCleskey at trial. See App. 83. Once this finding is credited, it follows that the State affirmatively misled McCleskey and his counsel throughout their unsuccessful pursuit of the Massiah claim in state collateral proceedings and their investigation of that claim in preparing for McCleskey’s first federal habeas proceeding. McCleskey’s counsel deposed or interviewed the assistant district attorney, various jailers, and other government officials responsible for Evans’ confinement, all of whom denied any knowledge of an agreement between Evans and the State. See App. 25-28, 44-47, 79, 85.

Against this background of deceit, the State’s withholding of Evans’ 21-page statement assumes critical importance. The majority overstates McCleskey’s and his counsel’s awareness of the statement’s contents. For example, the statement relates that state officials were present when Evans made a phone call at McCleskey’s request to McCleskey’s girlfriend, Plaintiff’s Exh. 8, p. 14, a fact that McCleskey and his counsel had no reason to know and that strongly supports the District Court’s finding of an ab initio relationship between Evans and the State. But in any event, the importance of the statement lay much less in what the statement said than in its simple existence. Without the statement, McCleskey’s counsel had nothing more than his client’s testimony to back up counsel’s own suspicion of a possible Massiah violation; given the state officials’ adamant denials of any arrangement with Evans, and given the state ha-beas court’s rejection of the Massiah claim, counsel quite reasonably concluded that raising this claim in McCleskey’s first habeas petition would be futile. All this changed once *527counsel finally obtained the statement, for at that point, there was credible, independent corroboration of counsel’s suspicion. This additional evidence not only gave counsel the reasonable expectation of success that had previously been lacking, but also gave him a basis for conducting further investigation into the underlying claim. Indeed, it was by piecing together the circumstances under which the statement had been transcribed that McCleskey’s counsel was able to find Worthy, a state official who was finally willing to admit that Evans had been planted in the cell adjoining McCleskey’s.12

The majority’s analysis of this case is dangerous precisely because it treats as irrelevant the effect that the State’s disinformation strategy had on counsel’s assessment of the reasonableness of pursing the Massiah claim. For the majority, all that matters is that no external obstacle barred McCleskey from finding Worthy. But obviously, counsel’s decision even to look for evidence in support of a particular claim has to be informed by what counsel reasonably perceives to be the prospect that the claim may have merit; in this case, by withholding the 21-page statement and by affirmatively misleading counsel as to the State’s involvement with Evans, state officials created a climate in which McCleskey’s first habeas counsel was perfectly justified in focusing his attentions elsewhere. The sum and substance of the majority’s analysis is that McCleskey had no “cause” for failing to assert the Massiah claim because he did not try *528hard enough to pierce the State’s veil of deception. Because the majority excludes from its conception of cause any recognition of how state officials can distort a petitioner’s reasonable perception of whether pursuit of a particular claim is worthwhile, the majority’s conception of “cause” creates an incentive for state officials to engage in this very type of misconduct.

Although the majority finds it unnecessary to reach the question whether McCleskey was “prejudiced” by the Massiah violation in this case, I have no doubt that the admission of Evans’ testimony at trial satisfies any fair conception of this prong of the Sykes test. No witness from the furniture store was able to identify which of the four robbers shot the off-duty police officer. The State did put on evidence that McCleskey had earlier stolen the pearl-handled pistol that was determined to be the likely murder weapon, but the significance of this testimony was clouded by a co-defendant’s admission that he had been carrying this weapon for weeks at a time, App. 16, and by a prosecution witness’ own prior statement that she had seen only the codefendant carry the pistol, id., at 11-14. See also id., at 89 (District Court finding that “the evidence on [McCleskey’s] possession of the gun in question was conflicting”). Outside of the self-serving and easily impeachable testimony of the codefendant, the only evidence that directly supported the State’s identification of McCleskey as the triggerman was the testimony of Evans. As the District Court found, “Evans’ testimony about the petitioner’s incriminating statements was critical to the state’s case.” Id., at 89. Without it, the jury might very well have reached a different verdict.

Thus, as I read the record, McCleskey should be entitled to the consideration of his petition for habeas corpus even under the cause-and-prejudice test. The case is certainly close enough to warrant a remand so that the issues can be fully and fairly briefed.

*529> H — H

Ironically, the majority seeks to defend its doctrinal innovation on the ground that it will promote respect for the “rule of law.” Ante, at 492. Obviously, respect for the rule of law must start with those who are responsible for pronouncing the law. The majority’s invocation of “‘the orderly administration of justice,’” ante, at 496, rings hollow when the majority itself tosses aside established precedents without explanation, disregards the will of Congress, fashions rules that defy the reasonable expectations of the persons who must conform their conduct to the law’s dictates, and applies those rules in a way that rewards state misconduct and deceit. Whatever “abuse of the writ” today’s decision is designed to avert pales in comparison with the majority’s own abuse of the norms that inform the proper judicial function.

I dissent.