Argued March 22, 2006. Decided June 22, 2006.
The Prison Litigation Reform Act of 1995 (PLRA) requires a prisoner to exhaust any available administrative remedies before challenging prison conditions in federal court. 42 U.S.C. § 1997e(a). Respondent filed a grievance with California prison officials about his prison conditions, but it was rejected as untimely under state law. He subsequently sued petitioner officials under § 1983 in the Federal District Court, which granted petitioners' motion to dismiss on the ground that respondent had not fully exhausted his administrative remedies under § 1997e(a). Reversing, the Ninth Circuit held that respondent had exhausted those remedies because none remained available to him.
Held: The PLRA's exhaustion requirement requires proper exhaustion of administrative remedies. Pp. 87-103.
(a) Petitioners claim that a prisoner must complete the administrative review process in accordance with applicable procedural rules, including deadlines, as a precondition to bringing suit in federal court, but respondent contends that § 1997e(a) allows suit once administrative remedies are no longer available, regardless of the reason. To determine the correct interpretation, the Court looks for guidance to both administrative and habeas corpus law, where exhaustion is an important doctrine. Administrative law requires proper exhaustion of administrative remedies, which "means using all steps that the agency holds out, and doing so properly." Pozo v. McCaughtry, 286 F. 3d 1022, 1024. Habeas law has substantively similar rules, though its terminology is different. Pp. 87-93.
(b) Given this background, the Court is persuaded that the PLRA requires proper exhaustion. Pp. 93-99.
(1) By referring to "such administrative remedies as are available," § 1997e(a)'s text strongly suggests "exhausted" means what it means in administrative law. P. 93.
(2) Construing § 1997e(a) to require proper exhaustion also serves the PLRA's goals. It gives prisoners an effective incentive to make full use of the prison grievance process, thus providing prisons with a fair opportunity to correct their own errors. It reduces the quantity of prisoner suits. And it improves the quality of those suits that are filed because proper exhaustion often results in creation of an administrative record helpful to the court. In contrast, respondent's interpretation would make the PLRA's exhaustion scheme totally ineffective, since exhaustion's benefits can be realized only if the prison grievance system is given a fair opportunity to consider the grievance. That cannot happen unless the grievant complies with the system's critical procedural rules. Respondent's arguments that his interpretation would filter out frivolous claims are unpersuasive. Pp. 93-96.
(3) As interpreted by respondent, the PLRA exhaustion requirement would be unprecedented. No statute or case purports to require exhaustion while at the same time allowing a party to bypass deliberately the administrative process by flouting the agency's procedural rules. None of his models is apt. He first suggests that the PLRA requirement was patterned on habeas law as it existed between 1963 and 1977 when, under Fay v. Noia, 372 U. S. 391, 438, a federal habeas claim could be procedurally defaulted only if the prisoner deliberately bypassed state remedies. That would be fanciful, however. The PLRA was enacted contemporaneously with the Antiterrorism and Effective Death Penalty Act of 1996, which gave federal habeas review a structure markedly different from what existed before 1977. Furthermore, respondent's interpretation would not duplicate that scheme, for it would permit a prisoner to bypass deliberately administrative review with no risk of sanction. Respondent next suggests that the PLRA exhaustion requirement is patterned on § 14(b) of the Age Discrimination in Employment Act of 1967 and § 706(e) of Title VII of the Civil Rights Act of 1964, but neither provision is in any sense an exhaustion provision. Pp. 96-99.
(c) Respondent's remaining arguments regarding § 1997e(a)'s interpretation are also unconvincing. Pp. 99-103.
403 F. 3d 620, reversed and remanded.
ALITO, J., delivered the opinion of the Court, in which ROBERTS, C. J., and SCALIA, KENNEDY, and THOMAS, JJ., joined. BREYER, J., filed an opinion concurring in the judgment, post, p. 103. STEVENS, J., filed a dissenting opinion, in which SOUTER and GINSBURG, JJ., joined, post, p. 104.
Jennifer G. Perkell, Deputy Attorney General of California, argued the cause for petitioners. With her on the briefs were Bill Lockyer, Attorney General, Manuel M. Madeiros, State Solicitor General, James M. Humes, Chief Assistant Attorney General, Frances T. Grander, Senior Assistant Attorney General, and Thomas S. Patterson, Supervising Deputy Attorney General. Dan Himmelfarb argued the cause for the United States as amicus curiae urging reversal. With him on the brief were Solicitor General Clement, Assistant Attorney General Keisler, Deputy Solicitor General Garre, and Barbara L. Herwig. Meir Feder argued the cause for respondent. With him on the brief were Charles R. A. Morse and Donald B. Ayer.
A brief of amici curiae urging reversal was filed for the State of New York et al. by Eliot Spitzer, Attorney General of New York, Caitlin J. Halligan, Solicitor General, Robert H. Easton, Deputy Solicitor General, and Richard Dearing, Assistant Solicitor General, and by the Attorneys General for their respective jurisdictions as follows: Troy King of Alabama, Terry Goddard of Arizona, Mike Beebe of Arkansas, John W. Slithers of Colorado, Carl C. Danberg of Delaware, Robert J. Spagnoletti of the District of Columbia, Thurbert E. Baker of Georgia, Mark J. Bennett of Hawaii, Lawrence G. Wasden of Idaho, Lisa Madigan of Illinois, Thomas J. Miller of Iowa, Phill Kline of Kansas, Thomas F. Reilly of Massachusetts, Michael A Cox of Michigan, Jeremiah W. (Jay) Nixon of Missouri, Mike McGrath of Montana, Jon Bruning of Nebraska, George J. Chanos of Nevada, Wayne Stenehjem of North Dakota, Jim Petro of Ohio, W. A. Drew Edmondson of Oklahoma, Hardy Myers of Oregon, Thomas W. Corbett, Jr., of Pennsylvania, Henry McMaster of South Carolina, Greg Abbott of Texas, Mark L. Shurtleff of Utah, William H. Sorrell of Vermont, Judith Williams Jagdmann of Virginia, Kerry E. Drue of the Virgin Islands, and Rob McKenna of Washington.
Briefs of amici curiae urging affirmance were filed for the American Bar Association by Michael S. Greco; for the American Civil Liberties Union et al. by Margo Schlanger, David C. Fathi, Elizabeth Alexander, Steven R. Shapiro, Steven Banks, and John Boston; for the Jerome N. Frank Legal Services Organization of the Yale Law School by Giovanna Shay; for Law Professors by Kermit Roosevelt III, Erwin Chemerinsky, David L. Franklin, Amanda Frost, Seth Kreimer, Daniel Manville, John Oakley, Malla Pollack, and David Rudovsky, all pro se.
This case presents the question whether a prisoner can satisfy the Prison Litigation Reform Act's exhaustion requirement, 42 U.S.C. § 1997e(a), by filing an untimely or otherwise procedurally defective administrative grievance or appeal. We hold that proper exhaustion of administrative remedies is necessary.
Congress enacted the Prison Litigation Reform Act of 1995 (PLRA), 110 Stat. 1321-71, as amended, 42 U.S.C. § 1997e et seq., in 1996 in the wake of a sharp rise in prisoner litigation in the federal courts, see, e. g., Alexander v. Hawk, 159 F. 3d 1321, 1324-1325 (CA11 1998) (citing statistics). The PLRA contains a variety of provisions designed to bring this litigation under control. See, e. g., § 1997e(c) (requiring district courts to weed out prisoner claims that clearly lack merit); § 1997e(e) (prohibiting claims for emotional injury without prior showing of physical injury); § 1997e(d) (restricting attorney's fees).
A centerpiece of the PLRA's effort "to reduce the quantity . . . of prisoner suits" is an "invigorated" exhaustion provision, § 1997e(a). Porter v. Nussle, 534 U. S. 516, 524 (2002). Before 1980, prisoners asserting constitutional claims had no obligation to exhaust administrative remedies. See Wilwording v. Swenson, 404 U. S. 249, 251 (1971) ( per curiam). In the Civil Rights of Institutionalized Persons Act, § 7, 94 Stat. 352-353, Congress enacted a weak exhaustion provision, which authorized district courts to stay actions under Rev. Stat. § 1979, 42 U.S.C. § 1983, for a limited time while a prisoner exhausted "such plain, speedy, and effective administrative remedies as are available." § 1997e(a)(1) (1994 ed.). "Exhaustion under the 1980 prescription was in large part discretionary; it could be ordered only if the State's prison grievance system met specified federal standards, and even then, only if, in the particular case, the court believed the requirement `appropriate and in the interests of justice.'" Nussle, supra, at 523 (quoting § 1997e). In addition, this provision did not require exhaustion if the prisoner sought only money damages and such relief was not available under the relevant administrative scheme. See McCarthy v. Madigan, 503 U. S. 140, 150-151 (1992).
The PLRA strengthened this exhaustion provision in several ways. Exhaustion is no longer left to the discretion of the district court, but is mandatory. See Booth v. Churner, 532 U. S. 731, 739 (2001). Prisoners must now exhaust all "available" remedies, not just those that meet federal standards. Indeed, as we held in Booth, a prisoner must now exhaust administrative remedies even where the relief sought — monetary damages — cannot be granted by the administrative process. Id., at 734. Finally, exhaustion of available administrative remedies is required for any suit challenging prison conditions, not just for suits under § 1983. Nussle, supra, at 524.
California has a grievance system for prisoners who seek to challenge their conditions of confinement. To initiate the process, an inmate must fill out a simple form, Dept. of Corrections, Inmate/Parolee Appeal Form, CDC 602 (12/87) (hereinafter Form 602), that is made "readily available to all inmates." Cal. Code Regs., tit. 15, § 3084.1(c) (2004). The inmate must fill out two parts of the form: part A, which is labeled "Describe Problem," and part B, which is labeled "Action Requested." Then, as explained on Form 602 itself, the prisoner "must first informally seek relief through discussion with the appropriate staff member." App. 40-41. The staff member fills in part C of Form 602 under the heading "Staff Response" and then returns the form to the inmate.
If the prisoner is dissatisfied with the result of the informal review, or if informal review is waived by the State, the inmate may pursue a three-step review process. See §§ 3084.5(b)-(d). Although California labels this "formal" review (apparently to distinguish this process from the prior step), the three-step process is relatively simple. At the first level, the prisoner must fill in part D of Form 602, which states: "If you are dissatisfied, explain below." Id., at 40. The inmate then must submit the form, together with a few other documents, to the appeals coordinator within 15 working days — three weeks — of the action taken. § 3084.6(c). This level may be bypassed by the appeals coordinator in certain circumstances. § 3084.5(b). Within 15 working days after an inmate submits an appeal, the reviewer must inform the inmate of the outcome by completing part E of Form 602 and returning the form to the inmate.
If the prisoner receives an adverse determination at this first level, or if this level is bypassed, the inmate may proceed to the second level of review conducted by the warden. §§ 3084.5(c), (e)(1). The inmate does this by filling in part F of Form 602 and submitting the form within 15 working days of the prior decision. Within 10 working days thereafter, the reviewer provides a decision on a letter that is attached to the form. If the prisoner's claim is again denied or the prisoner otherwise is dissatisfied with the result, the prisoner must explain the basis for his or her dissatisfaction on part H of the form and mail the form to the Director of the California Department of Corrections and Rehabilitation within 15 working days. § 3084.5(e)(2). An inmate's appeal may be rejected where "[t]ime limits for submitting the appeal are exceeded and the appellant had the opportunity to file within the prescribed time constraints." § 3084.3(c)(6).
Respondent is a prisoner who was convicted for murder and is serving a life sentence in the California prison system. In October 2000, respondent was placed in administrative segregation for allegedly engaging in "inappropriate activity" in the prison chapel. Two months later, respondent was returned to the general population, but respondent claims that he was prohibited from participating in "special programs," including a variety of religious activities. Approximately six months after that restriction was imposed, respondent filed a grievance with prison officials challenging that action. That grievance was rejected as untimely because it was not filed within 15 working days of the action being challenged. See §§ 3084.3(c)(6), 3084.6(c).
Respondent appealed that decision internally without success, and subsequently sued petitioners — California correctional officials — under 42 U.S.C. § 1983 in Federal District Court. The District Court granted petitioners' motion to dismiss because respondent had not fully exhausted his administrative remedies as required by § 1997e(a). See App. to Pet. for Cert. 24-25.
The Court of Appeals for the Ninth Circuit reversed and held that respondent had exhausted administrative remedies simply because no such remedies remained available to him. 403 F. 3d 620, 629-630 (2005). The Ninth Circuit's decision, while consistent with the decision of a divided panel of the Sixth Circuit in Thomas v. Woolum, 337 F. 3d 720 (2003), conflicts with decisions of four other Courts of Appeals. See Pozo v. McCaughtry, 286 F. 3d 1022, 1025 (CA7) ("To exhaust remedies, a prisoner must file complaints and appeals in the place, and at the time, the prison's administrative rules require"), cert. denied, 537 U. S. 949 (2002); Ross v. County of Bernalillo, 365 F. 3d 1181, 1185-1186 (CA10 2004) (same); Spruill v. Gillis, 372 F. 3d 218, 230 (CA3 2004) (same); Johnson v. Meadows, 418 F. 3d 1152, 1159 (CA11 2005) (same). We granted certiorari to address this conflict, 546 U. S. 1015 (2005), and we now reverse.
A The PLRA provides as follows:
"No action shall be brought with respect to prison conditions under section 1983 of this title, or any other Federal law, by a prisoner confined in any jail, prison, or other correctional facility until such administrative remedies as are available are exhausted." § 1997e(a) (2000 ed.) (emphasis added).
There is no dispute that this language requires a prisoner to "exhaust" administrative remedies, but the parties differ sharply in their understanding of the meaning of this requirement. Petitioners argue that this provision requires proper exhaustion. This means, according to petitioners, that a prisoner must complete the administrative review process in accordance with the applicable procedural rules, including deadlines, as a precondition to bringing suit in federal court. Respondent, on the other hand, argues that this provision demands what he terms "exhaustion simpliciter." Brief for Respondent 7. In his view, § 1997e(a) simply means that a prisoner may not bring suit in federal court until administrative remedies are no longer available. Under this interpretation, the reason why administrative remedies are no longer available is irrelevant. Bare unavailability suffices even if this results from a prisoner's deliberate strategy of refraining from filing a timely grievance so that the litigation of the prisoner's claim can begin in federal court.
The key for determining which of these interpretations of § 1997e(a) is correct lies in the term of art "exhausted." Exhaustion is an important doctrine in both administrative and habeas law, and we therefore look to those bodies of law for guidance.
"The doctrine of exhaustion of administrative remedies is well established in the jurisprudence of administrative law." McKart v. United States, 395 U. S. 185, 193 (1969). "The doctrine provides `that no one is entitled to judicial relief for a supposed or threatened injury until the prescribed administrative remedy has been exhausted.'" Ibid. (quoting Myers v. Bethlehem Shipbuilding Corp., 303 U. S. 41, 50-51 (1938)). Exhaustion of administrative remedies serves two main purposes. See McCarthy, 503 U. S., at 145.
First, exhaustion protects "administrative agency authority." Ibid. Exhaustion gives an agency "an opportunity to correct its own mistakes with respect to the programs it administers before it is haled into federal court," and it discourages "disregard of [the agency's] procedures." Ibid.
Second, exhaustion promotes efficiency. Ibid. Claims generally can be resolved much more quickly and economically in proceedings before an agency than in litigation in federal court. In some cases, claims are settled at the administrative level, and in others, the proceedings before the agency convince the losing party not to pursue the matter in federal court. See ibid.; Parisi v. Davidson, 405 U. S. 34, 37 (1972); McKart, supra, at 195. "And even where a controversy survives administrative review, exhaustion of the administrative procedure may produce a useful record for subsequent judicial consideration." McCarthy, supra, at 145.
Because of the advantages of administrative review, some aggrieved parties will voluntarily exhaust all avenues of administrative review before resorting to federal court, and for these parties an exhaustion requirement is obviously unnecessary. Statutes requiring exhaustion serve a purpose when a significant number of aggrieved parties, if given the choice, would not voluntarily exhaust. Aggrieved parties may prefer not to exhaust administrative remedies for a variety of reasons. Although exhaustion promotes overall efficiency, a party may conclude — correctly or incorrectly — that exhaustion is not efficient in that party's particular case. In addition, some aggrieved parties may prefer to proceed directly to federal court for other reasons, including bad faith. See Thomas, 337 F.3d, at 752-753 (Rosen, J., dissenting in part and concurring in judgment).
One can conceive of an inmate's seeking to avoid creating an administrative record with someone that he or she views as a hostile factfinder, filing a lawsuit primarily as a method of making some corrections official's life difficult, or perhaps even speculating that a suit will mean a welcome — if temporary — respite from his or her cell.
Because exhaustion requirements are designed to deal with parties who do not want to exhaust, administrative law creates an incentive for these parties to do what they would otherwise prefer not to do, namely, to give the agency a fair and full opportunity to adjudicate their claims. Administrative law does this by requiring proper exhaustion of administrative remedies, which "means using all steps that the agency holds out, and doing so properly (so that the agency addresses the issues on the merits)." Pozo, 286 F.3d, at 1024 (emphasis in original). This Court has described the doctrine as follows: "[A]s a general rule . . . courts should not topple over administrative decisions unless the administrative body not only has erred, but has erred against objection made at the time appropriate under its practice." United States v. L. A. Tucker Truck Lines, Inc., 344 U. S. 33, 37 (1952) (emphasis added). See also Sims v. Apfel, 530 U. S. 103, 108 (2000); id., at 112 (O'Connor, J., concurring in part and concurring in judgment) ("On this underlying principle of administrative law, the Court is unanimous"); id., at 114-115 (BREYER, J., dissenting); Unemployment Compensation Comm'n of Alaska v. Aragon, 329 U. S. 143, 155 (1946); Hormel v. Helvering, 312 U. S. 552, 556-557 (1941); 2 K. Davis R. Pierce, Administrative Law Treatise § 15:8, pp. 341-344 (3d ed. 1994). Proper exhaustion demands compliance with an agency's deadlines and other critical procedural rules because no adjudicative system can function effectively without imposing some orderly structure on the course of its proceedings.
The dissent makes two chief arguments regarding the doctrine of exhaustion in administrative law. Neither is sound.
First, the dissent contends that, "in the absence of explicit statutory directive," proper exhaustion is required only in proceedings that are in the nature of "appellate review proceedings." Post, at 112 (opinion of STEVENS, J.). The only authorities cited in support of this proposition are Sims v. Apfel, 530 U. S. 103, 108-109 (2000) — which concerns different questions, i. e., issue exhaustion and the distinction between adversarial and nonadversarial proceedings — and an amici brief, which in turns cites no supporting authority. See post, at 112 (citing Brief for Law Professors 1). The amici brief argues that "[t]he conceptual key to this case is [the] distinction" between an "original proceeding," in which "the court is simply determining the legality of out-of-court action," and a "review proceeding," in which the court must "review the decision of some other adjudicator." Id., at 2-3. According to the amici brief, habeas petitions are prime examples of "review proceeding[s]" because they "ask federal courts to review the decisions of state courts." Id., at 3. This argument is deeply flawed.
"[H]abeas corpus [is] an original . . . civil remedy for the enforcement of the right to personal liberty, rather than . . . a stage of the state criminal proceedings . . . or as an appeal therefrom." Fay v. Noia, 372 U. S. 391, 423-424 (1963) (footnote omitted). And habeas law includes the "judge-made doctrine of procedural default." Post, at 108, n. 5. This shows that the dissent and the amici brief are incorrect in contending that a proper exhaustion requirement is incompatible with an original proceeding.
Second, the dissent argues that, even if administrative law generally requires proper exhaustion, respondent falls within an exception to that rule. Post, at 114. As the dissent puts it, "[b]ecause respondent has raised constitutional claims, . . . the Court may not, as a matter of federal common law, apply an extrastatutory waiver requirement against him." Ibid. But we are not applying an "extrastatutory" requirement "as a matter of federal common law." Ibid. We are interpreting and applying the statutory requirement set out in the PLRA exhaustion provision. We interpret the PLRA exhaustion provision to require proper exhaustion, not the unprecedented scheme of exhaustion simpliciter that respondent advocates. As for the suggestion that the PLRA might be meant to require proper exhaustion of nonconstitutional claims but not constitutional claims, we fail to see how such a carve-out would serve Congress' purpose of addressing a flood of prisoner litigation in the federal courts, see supra, at 84, when the overwhelming majority of prisoner civil rights and prison condition suits are based on the Constitution.
The law of habeas corpus has rules that are substantively similar to those described above. The habeas statute generally requires a state prisoner to exhaust state remedies before filing a habeas petition in federal court. See 28 U.S.C. §§ 2254(b)(1), (c). "This rule of comity reduces friction between the state and federal court systems by avoiding the `unseem[liness]' of a federal district court's overturning a state-court conviction without the state courts having had an opportunity to correct the constitutional violation in the first instance." O'Sullivan v. Boerckel, 526 U. S. 838, 845 (1999) (alteration in original). A state prisoner is generally barred from obtaining federal habeas relief unless the prisoner has properly presented his or her claims through one "complete round of the State's established appellate review process." Ibid. In practical terms, the law of habeas, like administrative law, requires proper exhaustion, and we have described this feature of habeas law as follows: "To . . . `protect the integrity' of the federal exhaustion rule, we ask not only whether a prisoner has exhausted his state remedies, but also whether he has properly exhausted those remedies. . . ." Id., at 848 (citation omitted; emphasis in original).
The law of habeas, however, uses terminology that differs from that of administrative law. In habeas, the sanction for failing to exhaust properly (preclusion of review in federal court) is given the separate name of procedural default, although the habeas doctrines of exhaustion and procedural default "are similar in purpose and design and implicate similar concerns," Keeney v. Tamayo-Reyes, 504 U. S. 1, 7 (1992). See also Coleman v. Thompson, 501 U. S. 722, 731-732 (1991). In habeas, state-court remedies are described as having been "exhausted" when they are no longer available, regardless of the reason for their unavailability. See Gray v. Netherland, 518 U. S. 152, 161 (1996). Thus, if state-court remedies are no longer available because the prisoner failed to comply with the deadline for seeking state-court review or for taking an appeal, those remedies are technically exhausted, ibid., but exhaustion in this sense does not automatically entitle the habeas petitioner to litigate his or her claims in federal court. Instead, if the petitioner procedurally defaulted those claims, the prisoner generally is barred from asserting those claims in a federal habeas proceeding. Id., at 162; Coleman, supra, at 744-751.
With this background in mind, we are persuaded that the PLRA exhaustion requirement requires proper exhaustion.
The text of 42 U.S.C. § 1997e(a) strongly suggests that the PLRA uses the term "exhausted" to mean what the term means in administrative law, where exhaustion means proper exhaustion. Section 1997e(a) refers to "such administrative remedies as are available," and thus points to the doctrine of exhaustion in administrative law.
Construing § 1997e(a) to require proper exhaustion also fits with the general scheme of the PLRA, whereas respondent's interpretation would turn that provision into a largely useless appendage. The PLRA attempts to eliminate unwarranted federal-court interference with the administration of prisons, and thus seeks to "affor[d] corrections officials time and opportunity to address complaints internally before allowing the initiation of a federal case." Nussle, 534 U. S., at 525. See also Booth, 532 U. S., at 739. The PLRA also was intended to "reduce the quantity and improve the quality of prisoner suits." Nussle, supra, at 524.
See, e. g., 18 U.S.C. § 3626(b)(2) (termination of prison conditions consent decrees).
Requiring proper exhaustion serves all of these goals. It gives prisoners an effective incentive to make full use of the prison grievance process and accordingly provides prisons with a fair opportunity to correct their own errors. This is particularly important in relation to state corrections systems because it is "difficult to imagine an activity in which a State has a stronger interest, or one that is more intricately bound up with state laws, regulations, and procedures, than the administration of its prisons." Preiser v. Rodriguez, 411 U. S. 475, 491-492 (1973).
Proper exhaustion reduces the quantity of prisoner suits because some prisoners are successful in the administrative process, and others are persuaded by the proceedings not to file an action in federal court. Finally, proper exhaustion improves the quality of those prisoner suits that are eventually filed because proper exhaustion often results in the creation of an administrative record that is helpful to the court. When a grievance is filed shortly after the event giving rise to the grievance, witnesses can be identified and questioned while memories are still fresh, and evidence can be gathered and preserved.
The dissent's objection, post, at 115-116, that exhaustion simpliciter is enough to reduce frivolous prisoner suits is not well taken. First, what matters is not whether proper exhaustion was necessary to reach that goal, but whether proper exhaustion was mandated by Congress. Second, the empirical support for the dissent's conclusion is weak. The dissent points to a drop in volume of prisoner litigation between 1995 and 2000 and concludes that exhaustion simpliciter "was sufficient to reduce the quantity of prisoner suits without any procedural default requirement." Post, at 116. But this mistakes correlation for causation: A requirement of exhaustion simpliciter will not, absent a mollified prisoner, prevent a case from being docketed — and thus appearing in the filing statistics the dissent cites. The credit for reduced filings more likely belongs to the PLRA's enactment of 28 U.S.C. § 1915A (requiring district courts to screen "before docketing, if feasible," prisoner civil complaints), and its amendments to § 1915 (forbidding frequent-filer prisoners from proceeding in forma pauperis). Finally, prisoner civil rights and prison conditions cases still account for an outsized share of filings: From 2000 through 2005, such cases represented between 8.3% and 9.8% of the new filings in the federal district courts, or on average about one new prisoner case every other week for each of the nearly 1,000 active and senior district judges across the country. See Administrative Office of the United States Courts, Judicial Facts and Figures, tbls. 1.1, 4.4, 4.6, http://www.uscourts.gov/judicialfactsfigures/contents.html (as visited June 19, 2006, and available in Clerk of Court's case file).
While requiring proper exhaustion serves the purposes of the PLRA, respondent's interpretation of § 1997e(a) would make the PLRA exhaustion scheme wholly ineffective. The benefits of exhaustion can be realized only if the prison grievance system is given a fair opportunity to consider the grievance. The prison grievance system will not have such an opportunity unless the grievant complies with the system's critical procedural rules. A prisoner who does not want to participate in the prison grievance system will have little incentive to comply with the system's procedural rules unless noncompliance carries a sanction, and under respondent's interpretation of the PLRA noncompliance carries no significant sanction. For example, a prisoner wishing to by-pass available administrative remedies could simply file a late grievance without providing any reason for failing to file on time. If the prison then rejects the grievance as untimely, the prisoner could proceed directly to federal court. And acceptance of the late grievance would not thwart the prisoner's wish to bypass the administrative process; the prisoner could easily achieve this by violating other procedural rules until the prison administration has no alternative but to dismiss the grievance on procedural grounds. We are confident that the PLRA did not create such a toothless scheme.
Respondent argues that his interpretation of the PLRA's exhaustion provision would filter out frivolous claims because, by the time the deadline for filing a grievance has passed, the inmate may no longer wish to file suit. Brief for Respondent 43. But since the deadline for filing an administrative grievance is generally not very long — 14 to 30 days according to the United States, see Brief for United States as Amicus Curiae 29, and even less according to respondent, see Brief for Respondent 30, n. 17 — it is doubtful that Congress thought requiring a prisoner to wait this long would provide much of a deterrent. Indeed, many prisoners would probably find it difficult to prepare, file, and serve a civil complaint before the expiration of the deadline for filing a grievance in many correctional systems.
Respondent also contends that his interpretation of the PLRA exhaustion requirement would filter out frivolous claims because prisoners could not simply wait until the deadline for filing an administrative grievance had passed. According to respondent, "most grievance systems give administrators the discretion to hear untimely grievances," and therefore a prisoner "will be required to file an untimely grievance, and thereby give the grievance system" the opportunity to address the complaint. Id., at 43. But assuming for the sake of argument that the premise of this argument is correct, i. e., that a court could never conclude that administrative remedies were unavailable unless an administrative decision had so held, but see Coleman, 501 U. S., at 735, n., a prisoner who does not want to participate in the prison grievance process would have little difficulty in forcing the prison to dismiss his administrative case on procedural grounds. Under the California system, for example, a prisoner has numerous opportunities to miss deadlines. Therefore, the task of engineering such a dismissal of a grievance on procedural grounds is unlikely to be sufficient to alter the conduct of a prisoner whose objective is to by-pass the administrative process.
Finally, as interpreted by respondent, the PLRA exhaustion requirement would be unprecedented. Respondent has not pointed to any statute or case that purports to require exhaustion while at the same time allowing a party to by-pass deliberately the administrative process by flouting the agency's procedural rules. It is most unlikely that the PLRA, which was intended to deal with what was perceived as a disruptive tide of frivolous prisoner litigation, adopted an exhaustion requirement that goes further than any other model that has been called to our attention in permitting the wholesale bypassing of administrative remedies. Respondent identifies three models for the scheme of "exhaustion simpliciter" that he believes is set out in the PLRA, but none of these examples is apt.
Respondent first looks to habeas law as it existed prior to Wainwright v. Sykes, 433 U. S. 72 (1977). Before then, a federal habeas claim could be procedurally defaulted only if the prisoner deliberately bypassed state remedies. See Fay v. Noia, 372 U. S. 391, 438 (1963). It would be fanciful, however, to suggest that the PLRA exhaustion requirement was patterned on habeas law as it existed in the years between Fay and Wainwright. As respondent stresses, the PLRA was enacted contemporaneously with the Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA), 110 Stat. 1214, which gave federal habeas review a structure markedly different from that which existed in the period between Fay and Wainwright.
Furthermore, respondent's interpretation of § 1997e(a) would not duplicate the scheme that existed in habeas during that interval. As interpreted by respondent, § 1997e(a) would permit a prisoner to bypass deliberately and flagrantly administrative review without any risk of sanction. Because it is unlikely that the PLRA was intended to permit this, the two Courts of Appeals that have held that § 1997e(a) does not require proper exhaustion both pointedly stated that their decisions did not allow a prisoner to bypass deliberately administrative remedies. See 403 F. 3d, at 629; Thomas, 337 F. 3d, at 732, and n. 4. Neither of these courts, however, explained how § 1997e(a) can be interpreted in this way — that is, so that it does not require proper exhaustion but somehow proscribes deliberate bypass.
Apparently recognizing that such an interpretation neither has a statutory basis nor refers to a concept of exhaustion from an existing body of law, respondent does not contend that § 1997e(a) prohibits deliberate bypass; in his view, all that § 1997e(a) demands is that a prisoner wait until any opportunity for administrative review has evaporated. But in making this argument, respondent asks us to hold that the PLRA was meant to adopt an exhaustion scheme that stands in sharp contrast to both current and past habeas law and is unlike any other exhaustion scheme that has been called to our attention.
Respondent next suggests that the PLRA exhaustion requirement was patterned on § 14(b) of the Age Discrimination in Employment Act of 1967 (ADEA), 81 Stat. 607, codified at 29 U.S.C. § 633(b), and § 706(e) of Title VII of the Civil Rights Act of 1964, 78 Stat. 260, as redesignated and amended, 42 U.S.C. § 2000e-5(e), but these are implausible models. Neither of these provisions makes reference to the concept of exhaustion, and neither is in any sense an exhaustion provision.
In Oscar Mayer Co. v. Evans, 441 U. S. 750 (1979), we considered § 14(b) of the ADEA, which provides that, if a State has an agency to redress state-law age-related employment-discrimination claims, an ADEA claim may not be brought in federal court "before the expiration of sixty days after proceedings have been commenced under the State law." 29 U.S.C. § 633(b) (emphasis added). This provision makes no reference to the exhaustion of state remedies, only to the "commence[ment]" of state proceedings, and this provision leaves no doubt that proper commencement of those proceedings is not required. As we noted, see Oscar Mayer, 441 U. S., at 759, § 14(b) of the ADEA states that the requirement of commencement is satisfied merely by sending the state agency a signed statement of the pertinent facts, and § 14(b) explicitly provides that the commencement requirement does not entail compliance with any other state procedural rule, including a deadline for initiating the state proceeding, id., at 760. We see little similarity between § 14(b), which merely requires the commencement of state proceedings and explicitly does not require timely commencement, and 42 U.S.C. § 1997e(a), which expressly requires exhaustion of available administrative remedies with no reference to a federally based limiting principle.
Section 706(e) of Title VII is also fundamentally different from the PLRA exhaustion provision. As interpreted by this Court, § 706(e) means that a complainant who "initially institutes proceedings with a state or local agency with authority to grant or seek relief from the practice charged" must "file a charge" with that agency, or "have the EEOC refer the charge to that agency, within 240 days of the alleged discriminatory event. . . ." EEOC v. Commercial Office Products Co., 486 U. S. 107, 110-111 (1988). Following the reasoning of Oscar Mayer, we held that this filing requirement did not demand that the charge submitted to the state or local authority be filed in compliance with the authority's time limit. 486 U.S., at 123-125. Because § 706(e) of Title VII refers only to the filing of a charge with a state or local agency and not to the exhaustion of remedies, § 706(e) cannot be viewed as a model for the PLRA exhaustion provision.
Respondent's remaining arguments regarding the interpretation of 42 U.S.C. § 1997e(a) are unconvincing. Relying on the use of the term "until" in the phrase "until such administrative remedies as are available are exhausted," respondent contends that "[t]he use of the temporal word `until' . . . conveys a timing requirement: it assumes that the question to be answered is simply whether the prisoner can file suit now or must wait until later." Brief for Respondent 11. Likewise, according to respondent, the use of the present tense ("such administrative remedies as are available," § 1997e(a) (emphasis added)) requires "a focus on whether any administrative remedies are presently available." Id., at 12. But saying that a party may not sue in federal court until the party first pursues all available avenues of administrative review necessarily means that, if the party never pursues all available avenues of administrative review, the person will never be able to sue in federal court. Thus, § 1997e(a)'s use of the term "until" and the present tense does not support respondent's position.
Respondent attaches significance to the fact that the PLRA exhaustion provision does not expressly state that a prisoner must have "properly exhausted" available administrative remedies, whereas a tolling provision of the AEDPA provides that the time for filing a federal habeas petition is tolled during the period when "a properly filed application for State post-conviction or other collateral review . . . is pending." 28 U.S.C. § 2244(d)(2) (emphasis added). In our view, respondent draws an unreasonable inference from the difference in the wording of these two provisions. Although the AEDPA and the PLRA were enacted at roughly the same time, they are separate and detailed pieces of legislation. Moreover, the AEDPA and PLRA provisions deal with separate issues: tolling in the case of the AEDPA and exhaustion in the case of the PLRA.
Respondent maintains that his interpretation of the PLRA exhaustion provision is bolstered by another PLRA provision, 42 U.S.C. § 1997e(c)(2), that permits a district court to dismiss certain prisoner claims "without first requiring the exhaustion of administrative remedies." According to respondent, this provision shows that Congress thought that, at the point when a district court might make such a ruling (which would typically be well after the filing of the complaint), a prisoner might still have the opportunity to exhaust administrative remedies. Because short administrative filing deadlines would make this impossible, respondent contends, Congress cannot have thought that a prisoner's failure to comply with those deadlines would preclude litigation in federal court.
Respondent's argument is unconvincing for at least two reasons. First, respondent has not shown that Congress had reason to believe that every prison system would have relatively short and categorical filing deadlines. Indeed, respondent asserts that most grievance systems give administrators the discretion to hear untimely grievances. Second, even if dismissals under § 1997e(c)(2) typically occur when the opportunity to pursue administrative remedies has passed, § 1997e(c)(2) still serves a useful function by making it clear that the PLRA exhaustion requirement is not jurisdictional, and thus allowing a district court to dismiss plainly meritless claims without first addressing what may be a much more complex question, namely, whether the prisoner did in fact properly exhaust available administrative remedies.
Questions regarding the timeliness of prisoner filings occur frequently. See, e. g., Wallace v. Burbury, 305 F. Supp. 2d 801, 806 (ND Ohio 2003); Pusey v. Belanger, No. Civ.02-351-SLR, 2004 WL 2075472 (D. Del., Sept. 14, 2004); Eakle v. Tennis, No. Civ. 4:CV-04-2040, 2005 WL 2266270 (MD Pa., Sept. 16, 2005); Williams v. Briley, No. 04 C 5701, 2005 WL 1498865 (ND Ill., June 21, 2005); Isaac v. Nix, No. Civ.A.2:04CV172RWS, 2006 WL 861642 (ND Ga., Mar. 30, 2006).
Respondent next argues that the similarity between the wording of the PLRA exhaustion provision and the AEDPA exhaustion provision, 28 U.S.C. § 2254(c), shows that the PLRA provision was meant to incorporate the narrow technical definition of exhaustion that applies in habeas. We reject this argument for two reasons.
First, there is nothing particularly distinctive about the wording of the habeas and PLRA exhaustion provisions. They say what any exhaustion provision must say — that a judicial remedy may not be sought or obtained unless, until, or before certain other remedies are exhausted. It is, therefore, unrealistic to infer from the wording of the PLRA provision that Congress framed and adopted that provision with habeas law and not administrative law in mind. Indeed, the wording of the PLRA provision (a prisoner may not bring an action with respect to prison conditions " until such administrative remedies as are available are exhausted") is strikingly similar to our description of the doctrine of administrative exhaustion ("`no one is entitled to judicial relief for a supposed or threatened injury until the prescribed administrative remedy has been exhausted,'" McKart, 395 U. S., at 193 (emphasis added)).
Second, respondent's suggestion that the PLRA was meant to incorporate the same technical distinction that exists in habeas law without providing any sanction to prevent willful noncompliance — not even the deliberate bypass standard of Fay — would produce a scheme that in practical terms is radically different from the habeas scheme. Copying habeas' narrow definition of exhaustion without furnishing any sanction to promote compliance would be like copying the design for an airplane but omitting one of the wings.
Respondent contends that requiring proper exhaustion will lead prison administrators to devise procedural requirements that are designed to trap unwary prisoners and thus to defeat their claims. Respondent does not contend, however, that anything like this occurred in his case, and it is speculative that this will occur in the future. Corrections officials concerned about maintaining order in their institutions have a reason for creating and retaining grievance systems that provide — and that are perceived by prisoners as providing — a meaningful opportunity for prisoners to raise meritorious grievances. And with respect to the possibility that prisons might create procedural requirements for the purpose of tripping up all but the most skillful prisoners, while Congress repealed the "plain, speedy, and effective" standard, see 42 U.S.C. § 1997e(a)(1) (1994 ed.) (repealed 1996), we have no occasion here to decide how such situations might be addressed.
Respondent argues that requiring proper exhaustion is harsh for prisoners, who generally are untrained in the law and are often poorly educated. This argument overlooks the informality and relative simplicity of prison grievance systems like California's, as well as the fact that prisoners who litigate in federal court generally proceed pro se and are forced to comply with numerous unforgiving deadlines and other procedural requirements.
* * *
For these reasons, we reverse the judgment of the Court of Appeals for the Ninth Circuit and remand the case for proceedings consistent with this opinion.
It is so ordered.
I agree with the Court that, in enacting the Prison Litigation Reform Act (PLRA), 42 U.S.C. § 1997e(a), Congress intended the term "exhausted" to "mean what the term means in administrative law, where exhaustion means proper exhaustion." Ante, at 93. I do not believe that Congress desired a system in which prisoners could elect to bypass prison grievance systems without consequences. Administrative law, however, contains well-established exceptions to exhaustion. See Sims v. Apfel, 530 U. S. 103, 115 (2000) (BREYER, J., joined by Rehnquist, C. J., and SCALIA and KENNEDY, JJ., dissenting) (constitutional claims); Shalala v. Illinois Council on Long Term Care, Inc., 529 U. S. 1, 13 (2000) (futility); McKart v. United States, 395 U. S. 185, 197-201 (1969) (hardship); McCarthy v. Madigan, 503 U. S. 140, 147-148 (1992) (inadequate or unavailable administrative remedies); see generally II R. Pierce, Administrative Law Treatise § 15 (4th ed. 2002). Moreover, habeas corpus law, which contains an exhaustion requirement that is "substantively similar" to administrative law's and which informs the Court's opinion, ante, at 92-93, also permits a number of exceptions. See post, at 109, n. 5 (STEVENS, J., dissenting) (noting that habeas corpus law permits "petitioners to overcome procedural defaults if they can show that the procedural rule is not firmly established and regularly followed, if they can demonstrate cause and prejudice to overcome a procedural default, or if enforcing the procedural default rule would result in a miscarriage of justice" (citation omitted)).
At least two Circuits that have interpreted the statute in a manner similar to that which the Court today adopts have concluded that the PLRA's proper exhaustion requirement is not absolute. See Spruill v. Gillis, 372 F. 3d 218, 232 (CA3 2004); Giano v. Goord, 380 F. 3d 670, 677 (CA2 2004). In my view, on remand, the lower court should similarly consider any challenges that respondent may have concerning whether his case falls into a traditional exception that the statute implicitly incorporates.
The citizen's right to access an impartial tribunal to seek redress for official grievances is so fundamental and so well established that it is sometimes taken for granted. A state statute that purported to impose a 15-day period of limitations on the right of a discrete class of litigants to sue a state official for violation of a federal right would obviously be unenforceable in a federal court. The question in this case is whether, by enacting the exhaustion requirement in the Prison Litigation Reform Act of 1995 (PLRA), Congress intended to authorize state correction officials to impose a comparable limitation on prisoners' constitutionally protected right of access to the federal courts. The text of the statute, particularly when read in the light of our well-settled jurisprudence, provides us with the same unambiguous negative answer that common sense would dictate.
Congress enacted the following exhaustion requirement in the PLRA:
"No action shall be brought with respect to prison conditions under section 1983 of this title, or any other Federal law, by a prisoner confined in any jail, prison, or other correctional facility until such administrative remedies as are available are exhausted." 42 U.S.C. § 1997e(a).
This provision requires prisoners to exhaust informal remedies before filing a lawsuit under federal law. They must file an administrative grievance and, if the resolution of that grievance is unsatisfactory to them, they must exhaust available administrative appeals. The statute, however, says nothing about the reasons why a grievance may have been denied; it does not distinguish between a denial on the merits and a denial based on a procedural error. It does not attach any significance to a prison official's decision that a prisoner has made procedural missteps in exhausting administrative remedies. In the words of federal courts jurisprudence, the text of the PLRA does not impose a sanction of waiver or procedural default upon those prisoners who make such procedural errors. See Engle v. Isaac, 456 U. S. 107, 125-126, n. 28 (1982) (explaining that "the problem of waiver is separate from the question whether a state prisoner has exhausted state remedies"). The plain text of the PLRA simply requires that "such administrative remedies as are available" be exhausted before the prisoner can take the serious step of filing a federal lawsuit against the officials who hold him in custody.
Because we have used the term "waiver" in referring to this sanction in the habeas corpus context, I use that term in this opinion. Strictly speaking, it would be more accurate to characterize this sanction as a "forfeiture" sanction, as there is no question that prisoners do not, by making a procedural error in the course of exhausting administrative remedies, purposefully relinquish their right to bring constitutional claims in federal court.
Today, however, the Court concludes that the "PLRA exhaustion requirement requires proper exhaustion," ante, at 93. The absence of textual support for that conclusion is a sufficient reason for rejecting it. Unlike 28 U.S.C. § 2244(d)(2), a tolling provision of the Antiterrorism and Effective Death Penalty Act of 1996, which was signed into law just two days before the PLRA, 42 U.S.C. § 1997e(a) lacks any textual requirement of proper exhaustion. See Artuz v. Bennett, 531 U. S. 4, 8 (2000) (explaining the importance of the textual requirement that an application be " properly filed" under 28 U.S.C. § 2244(d)(2)). Instead, just as in the habeas context, under the PLRA a prisoner "who has [procedurally] defaulted his federal claims in [a state prison grievance proceeding] meets the technical requirements for exhaustion; there are no state remedies any longer `available' to him." Coleman v. Thompson, 501 U. S. 722, 732 (1991). Accordingly, under the plain text of 42 U.S.C. § 1997e(a), respondent satisfied his duty to exhaust available administrative remedies before filing a federal lawsuit.
The majority essentially ignores the PLRA's text, suggesting instead that general administrative law principles, which allow courts in certain circumstances to impose procedural default sanctions as a matter of federal common law, suggest we should read waiver into the PLRA. However, as discussed in Part III, infra, our cases make clear that such extratextual waiver sanctions are only appropriate if a statute directs a federal court to act as an appellate tribunal directly reviewing the decision of a federal agency. Because actions brought under Rev. Stat. § 1979, 42 U.S.C. § 1983, such as respondent's, are de novo proceedings in federal district court, the majority's invocation of these common-law principles is seriously misguided.
The majority does not claim that the plain language of the statute dictates its decision, but rather that the text "strongly suggests" that the PLRA includes a procedural default sanction, ante, at 93. The majority then states: "Section 1997e(a) refers to `such administrative remedies as are available,' and thus points to the doctrine of exhaustion in administrative law." Ibid. The reference to "administrative remedies" simply addresses the fact that the review procedures provided by prison officials are administrative in character rather than judicial. At any rate, as discussed in Part III, infra, the doctrine of exhaustion in administrative law does not support the majority's engraftment of a procedural default sanction into the PLRA.
The majority's disregard of the plain text of the PLRA is especially unjustified in light of the backdrop against which the statute was enacted. We presume, of course, that Congress is familiar with this Court's precedents and expects its legislation to be interpreted in conformity with those precedents. See, e. g., Edelman v. Lynchburg College, 535 U. S. 106, 117, n. 13 (2002); Porter v. Nussle, 534 U. S. 516, 528 (2002); North Star Steel Co. v. Thomas, 515 U. S. 29, 34 (1995). This strong presumption is even more forceful when the underlying precedent is "`unusually important.'" Gebser v. Lago Vista Independent School Dist., 524 U. S. 274, 294, n. 1 (1998) (quoting Cannon v. University of Chicago, 441 U. S. 677, 699 (1979)). Consistent with this presumption, if we have already provided a definitive interpretation of the language in one statute, and Congress then uses nearly identical language in another statute, we will give the language in the latter statute an identical interpretation unless there is a clear indication in the text or legislative history that we should not do so. See, e. g., United States v. Wells, 519 U. S. 482, 495 (1997). Under these elementary principles of statutory interpretation, the PLRA's exhaustion requirement does not incorporate a procedural default component.
As the Solicitor General correctly points out in his brief supporting petitioners, "the PLRA's exhaustion provision is essentially identical to that of the habeas corpus statute." Brief for United States as Amicus Curiae 13. Specifically, a provision in the federal habeas statute, first enacted in 1948 as a codification of a previous judge-made rule, bars relief "unless it appears that . . . the applicant has exhausted the remedies available in the courts of the State," 28 U.S.C. § 2254(b)(1)(A). The PLRA similarly bars judicial relief "until such administrative remedies as are available are exhausted," 42 U.S.C. § 1997e(a). The only noteworthy distinction between the two provisions is that 28 U.S.C. § 2254(b)(1)(A) uses the word "unless," whereas 42 U.S.C. § 1997e(a) uses the word "until." If anything, this distinction suggests that the exhaustion requirement in the PLRA is less amenable to a waiver sanction than the comparable requirement in the habeas statute: The word "until" indicates a temporal condition whereas the word "unless" would have been more appropriate for a procedural bar.
See generally O'Sullivan v. Boerckel, 526 U.S. 838, 850-853 (1999) (STEVENS, J., joined by GINSBURG and BREYER, JJ., dissenting) (tracing history of exhaustion requirement in habeas law).
This language is, in relevant part, identical to the language as it was enacted in 1948. See 62 Stat. 967.
Notwithstanding the use of the word "unless" in 28 U.S.C. § 2254(b)(1)(A), as the majority correctly recognizes, we have held that state-court remedies are "exhausted" for the purposes of the federal habeas statute so long as "they are no longer available, regardless of the reason for their unavailability," ante, at 92-93. In other words, the exhaustion requirement in the federal habeas statute does not incorporate a procedural default sanction.
In habeas law it is a separate judge-made doctrine of procedural default, stemming from our decision in Wainwright v. Sykes, 433 U. S. 72 (1977), that may bar relief even though a claim has been exhausted. This procedural default doctrine is based on unique considerations of comity in the habeas context, including the need to ensure that the state criminal trial remains the "main event" rather than a "tryout on the road" for a later federal habeas proceeding. Id., at 90 (internal quotation marks omitted). Moreover, procedural default in habeas is closely related to the principle that this Court lacks certiorari jurisdiction to review a state-court judgment that rests on an adequate and independent state procedural ground. See id., at 81-82. It is undisputed that these unique considerations do not apply in the context of 42 U.S.C. § 1983 suits, because the "very purpose of § 1983 was to interpose the federal courts between the States and the people, as guardians of the people's federal rights." Mitchum v. Foster, 407 U. S. 225, 242 (1972). Accordingly, the majority correctly does not suggest that we incorporate our procedural default jurisprudence from the federal habeas context into prison conditions suits under § 1983.
Nonetheless, I fear that the majority's analysis may actually create a harsher procedural default regime under the PLRA than the judge-made procedural default doctrine in habeas law. But see Muhammad v. Close, 540 U. S. 749, 751 (2004) ( per curiam) (stating that "[p]risoners suing under § 1983 . . . generally face a substantially lower gate [than prisoners seeking habeas corpus relief], even with the requirement of the Prison Litigation Reform Act of 1995 that administrative opportunities be exhausted first" (citing 42 U.S.C. § 1997e(a))). Our habeas jurisprudence allows petitioners to overcome procedural defaults if they can show that the procedural rule is not firmly established and regularly followed, see James v. Kentucky, 466 U. S. 341, 348 (1984), if they can demonstrate cause and prejudice to overcome a procedural default, or if enforcing the procedural default rule would result in a miscarriage of justice, see Murray v. Carrier, 477 U. S. 478 (1986).
Between Congress' codification of the exhaustion requirement in federal habeas law and Congress' adoption of an essentially identical exhaustion requirement in the PLRA, we decided no fewer than six cases in which we stated explicitly that a habeas petitioner satisfies the statutory exhaustion requirement so long as state-court remedies are no longer available to him at the time of the federal-court filing, regardless of the reason for their unavailability. See Coleman, 501 U. S., at 731; Castille v. Peoples, 489 U. S. 346, 351 (1989); Teague v. Lane, 489 U. S. 288, 298 (1989); Engle, 456 U. S., at 125, n. 8; Humphrey v. Cady, 405 U. S. 504, 516 (1972); Fay v. Noia, 372 U. S. 391, 434-435 (1963).
The Court rejects the obvious analogy to habeas law because the wording of the PLRA's exhaustion provision is also "strikingly similar to our description of the doctrine of administrative exhaustion (`"no one is entitled to judicial relief for a supposed or threatened injury until the prescribed administrative remedy has been exhausted"')," ante, at 102 (quoting McKart v. United States, 395 U. S. 185, 193 (1969), in turn citing Myers v. Bethlehem Shipbuilding Corp., 303 U. S. 41, 50-51 (1938)). The language quoted by the majority from our case law is indeed similar to the language of the PLRA (and the habeas corpus statute). But this provides no help to the majority: We clearly used this language to describe only an exhaustion requirement, not a procedural default sanction.
The quoted language originally appeared in Justice Brandeis' opinion in Myers, 303 U. S., at 50-51. Myers is a simple exhaustion case: The question presented was whether an employer could seek the immediate intervention of federal courts in response to a complaint filed with the National Labor Relations Board that it had engaged in unfair labor practices, or whether it had to await the conclusion of the Board's proceedings to avail itself of judicial review. The case was purely about timing — there was no discussion whatever of procedural default.
McKart clearly recognized that the language of Myers concerned only exhaustion, not procedural default. Immediately after quoting Myers, the McKart Court discussed the benefits of exhaustion (primarily avoiding premature interruption of the agency process), and drew an analogy to judicial rules that limit interlocutory appeals, without making any reference to procedural default. See 395 U. S., at 193-194. It was not until later in the opinion that the McKart Court turned to a discussion of the considerations underlying the imposition of a procedural default sanction in cases "where the administrative process is at an end and a party seeks judicial review of a decision that was not appealed through the administrative process." Id., at 194.
In sum, the language the majority quotes from McKart further supports the presumption that Congress intended the exhaustion requirement in the PLRA to be read in conformity with our decisions interpreting the exhaustion requirement in the federal habeas statute — that is, to require exhaustion, but not to impose a waiver sanction for procedural errors made in the course of exhaustion.
Absent any support for a procedural default sanction in the text of the PLRA, the Court turns to background principles of administrative law in an effort to justify its holding. See ante, at 89-91. The Court's discussion of these background administrative law principles misapprehends our precedent.
As a general rule in the administrative law context, courts should not "`topple over administrative decisions unless the administrative body has not only erred, but has erred against objection made at the appropriate time under its practice.'" Ante, at 90 (quoting United States v. L. A. Tucker Truck Lines, Inc., 344 U. S. 33, 37 (1952)). This doctrine is, "like most judicial doctrines, subject to numerous exceptions. Application of the doctrine to specific cases requires an understanding of its purposes and of the particular administrative scheme involved." McKart, 395 U. S., at 193 (footnote omitted); see id., at 198-201 (declining to apply waiver doctrine in the circumstances of the case before it).
The waiver doctrine in administrative law is "largely [a] creatur[e] of statute." Sims v. Apfel, 530 U. S. 103, 107 (2000). In other words, many statutes explicitly prohibit courts from considering claims "`that ha[ve] not been urged'" before the administrative agency. Id., at 108 (quoting National Labor Relations Act, 29 U.S.C. § 160(e) (1982 ed.)). See L. A. Tucker Truck Lines, 344 U. S., at 36, n. 6 (collecting statutes). It is important to emphasize that statutory waiver requirements always mandate, by their plain terms, that courts shall not consider arguments not properly raised before the agency; we have never suggested that the word "exhaustion," standing alone, imposes a statutory waiver requirement. Accordingly, the Court's claim that a procedural default sanction is mandated by simply "interpreting and applying the statutory requirement set out in the PLRA exhaustion provision," ante, at 91, n. 2, is patently erroneous.
In the federal administrative law context we have also imposed waiver requirements even in the absence of explicit statutory directive. This judge-made rule, discussed extensively by the majority, see ante, at 88-91, however, is based on "an analogy to the rule that appellate courts will not consider arguments not raised before trial courts." Sims, 530 U. S., at 108-109. As amici curiae law professors explain, this is because, in the context of such appellate review proceedings, procedural errors in the course of exhaustion naturally create bars to review because the decision under review rests on a procedural ground. Brief for Law Professors 1. Moreover, the rule that appellate tribunals will not consider claims not properly exhausted below prevents parties from being unfairly surprised on appeal by resolution of issues about which they lacked an opportunity or incentive to introduce evidence at trial. See Sims, 530 U. S., at 109. Accordingly, whether a court should impose a procedural default sanction for issues not properly exhausted in a prior administrative proceeding "depends on the degree to which the analogy to normal adversarial litigation applies in a particular administrative proceeding." Ibid. (citing L. A. Tucker Truck Lines and Hormel v. Helvering, 312 U. S. 552 (1941)). If the analogy does not hold, we will not impose a procedural default sanction. See Sims, 530 U. S., at 108-110.
The majority's attempt to distinguish Sims as concerning "different questions," ante, at 91, n. 2, is perplexing, particularly in light of the fact that the United States, in its brief supporting petitioners, relies on Sims to argue that our administrative law decisions support the proposition that the Court should impose a waiver sanction into the PLRA. See Brief for United States as Amicus Curiae 11. Although the particular procedural error made during the exhaustion of administrative remedies was different in Sims than the procedural error at issue here, our analysis in Sims concerned the circumstances under which we should or should not engraft a waiver sanction into the administrative exhaustion process generally. See 530 U. S., at 108-112; id., at 112-113 (O'Connor, J., concurring in part and concurring in judgment); id., at 114-115 (BREYER, J., dissenting).
Applying these principles, it is clear that ordinary principles of administrative law do not justify engrafting procedural default into the PLRA. The purpose of a 42 U.S.C. § 1983 action such as that filed by respondent is not to obtain direct review of an order entered in the grievance procedure, but to obtain redress for an alleged violation of federal law committed by state corrections officials. See, e. g., Mitchum v. Foster, 407 U. S. 225, 242 (1972). It is undisputed that the PLRA does nothing to change the nature of the federal action under § 1983; prisoners who bring such actions after exhausting their administrative remedies are entitled to de novo proceedings in the federal district court without any deference (on issues of law or fact) to any ruling in the administrative grievance proceedings. In sum, because federal district court proceedings in prison condition litigation bear no resemblance to appellate review of lower court decisions, the administrative law precedent cited by the majority makes clear that we should not engraft a judge-made procedural default sanction into the PLRA. The majority's mis-apprehension of our precedent is especially troubling because, as the American Bar Association points out, we should be particularly hesitant to impose "judicially-created procedural technicalities . . . `in a statutory scheme in which laymen, unassisted by trained lawyers, initiate the process.'" Brief as Amicus Curiae 11 (quoting Oscar Mayer Co. v. Evans, 441 U. S. 750, 765, n. 13 (1979)).
The majority's suggestion that habeas law indicates otherwise, see ante, at 91-92, n. 2, is incorrect. As explained above, see n. 5, supra, the judge-made procedural default sanction in habeas law is based on unique considerations that do not apply to § 1983 suits. Our precedent concerning judicial review of administrative proceedings, upon which the majority purports to rely, see ante, at 93, makes clear that we will not impose a waiver sanction when judicial review of the administrative decision does not resemble appellate review of lower court decisions.
The majority notes that many prisoners proceed pro se in federal court, where there are also time limits and other procedural requirements. See ante, at 102. However, the timeliness and other procedural requirements of prison grievance systems are generally far more stringent than those imposed by federal courts. See Brief for American Civil Liberties Union et al. as Amici Curiae 6, n. 1, 25-27; Brief for Jerome N. Frank Legal Services Organization of Yale Law School as Amicus Curiae A1-A7.
Finally, the majority's invocation of judge-made administrative law principles fails for an entirely separate reason: An "established exception" to the judge-made doctrine of procedural default in review of administrative proceedings permits individuals to raise constitutional complaints for the first time in federal court, even if they failed to raise those claims properly before the agency. Sims, 530 U. S., at 115 (BREYER, J., joined by Rehnquist, C. J., and SCALIA and KENNEDY, JJ., dissenting) (citing Mathews v. Eldridge, 424 U. S. 319, 329, n. 10 (1976)). Because respondent has raised constitutional claims, under our precedent, the Court may not, as a matter of federal common law, apply an extrastatutory waiver requirement against him.
The principal arguments offered by the Court in support of its holding are policy arguments that, in its view, are grounded in the purposes of the PLRA. The majority correctly identifies two of the principal purposes of the PLRA: (1) affording corrections officials time and opportunity to address complaints internally before the initiation of a federal lawsuit; and (2) reducing the quantity, and improving the quality, of prison litigation. Both of these purposes would be served by the PLRA, even if the Court did not engraft a procedural default sanction into the statute.
Of course, if the majority were serious that "what matters is not whether proper exhaustion was necessary to reach [policy goals], but whether proper exhaustion was mandated by Congress," ante, at 94, n. 4, its opinion would not rest almost entirely on policy arguments.
The first policy concern identified by the majority does not even arguably justify either a timeliness requirement or a procedural default sanction. Prison officials certainly have the opportunity to address claims that were filed in some procedurally defective manner; indeed, California, like the vast majority of state prison systems, explicitly gives prison administrators an opportunity to hear untimely or otherwise procedurally defective grievances. Cal. Code Regs., tit. 15, § 3084.3(c). See generally Roosevelt, Exhaustion Under the Prison Litigation Reform Act: The Consequence of Procedural Error, 52 Emory L. J. 1771, 1810, and n. 192 (2003) (hereinafter Roosevelt). Because it is undisputed that the PLRA mandates that prisoners exhaust their administrative remedies before filing a federal lawsuit, prison officials will have the opportunity to address prisoners' claims before a suit is filed.
In this regard, the majority's reference to Coleman v. Thompson, 501 U. S. 722, 735, n. (1991), see ante, at 96, is perplexing. If a prison regulation explicitly grants prison officials discretion to consider untimely or otherwise procedurally defective grievances, of course prison grievance remedies would still be "available," and thus unexhausted, if a prisoner had not even tried to file a grievance simply because it was untimely or otherwise procedurally defective.
Second, the PLRA has already had the effect of reducing the quantity of prison litigation, without the need for an extrastatutory procedural default sanction. As petitioners themselves point out, the number of civil rights suits filed by prisoners in federal court dropped from 41,679 in 1995 to 25,504 in 2000, and the rate of prisoner filing dropped even more dramatically during that period, from 37 prisoner suits per 1,000 inmates to 19 suits per 1,000 inmates. By contrast, between 2000 and 2004, the rate of filing remained relatively constant, dropping only "slight[ly]" to approximately 16 suits per 1,000 inmates. See Brief for Petitioners 21-22. The sharp drop in prison litigation between 1995 and 2000 occurred before the Seventh Circuit's opinion in Pozo v. McCaughtry, 286 F. 3d 1022 (2002), which was the first appellate decision engrafting a procedural default sanction into the PLRA. Prior to Pozo, the federal courts had regularly assumed that the PLRA did not create any procedural default sanction, and dismissals for failure to exhaust were without prejudice. See Roosevelt 1780-1781 (discussing cases). Thus, the PLRA, including its simple exhaustion requirement, was sufficient to reduce the quantity of prisoner suits without any procedural default requirement. This is not surprising: Because the exhaustion requirement always ensures that prison officials have an opportunity to address claims brought by prisoners before a federal lawsuit, some prisoners will be "successful in the administrative process, and others are persuaded by the proceedings not to file an action in federal court," ante, at 94, in part because "the very fact of being heard . . . can mollify passions," Booth v. Churner, 532 U. S. 731, 737 (2001).
Without any support, the majority speculates that the drop in suits filed by prisoners between 1995 and 2000 resulted from other provisions of the PLRA. See ante, at 94, n. 4. Regardless, the aforementioned statistics demonstrate that the procedural default sanction imposed by the PLRA is unnecessary to reduce the quantity of prison litigation.
Ordinary exhaustion also improves the quality of prisoner suits. By giving prison officials an opportunity to address a prisoner's grievance before the initiation of the lawsuit, ordinary exhaustion "often results in the creation of an administrative record that is helpful to the court," ante, at 95.
The majority also argues that ensuring strict compliance with strict prison timeliness requirements (generally ranging from 48 hours to a month, see n. 15, infra) will improve the quality of prisoner litigation because if "a grievance is filed shortly after the event giving rise to the grievance, witnesses can be identified and questioned while memories are still fresh, and evidence can be gathered and preserved." Ante, at 95. While these are advantages to filing grievances soon after the alleged injury occurs, courts regularly resolve § 1983 (and other) litigation without such Draconian time limitations. At any rate, as discussed below, legislation does not pursue any one purpose at all costs, and the marginal advantages of encouraging compliance with such short time limitations do not justify judicially rewriting the PLRA's exhaustion requirement by engrafting a procedural default sanction into the statute.
I acknowledge, of course, that the majority's creation of a waiver sanction for procedural missteps during the course of exhaustion will have an even more significant effect in reducing the number of lawsuits filed by prisoners. However, "no legislation pursues its purposes at all costs. Deciding what competing values will or will not be sacrificed to the achievement of a particular objective is the very essence of legislative choice — and it frustrates rather than effectuates legislative intent simplistically to assume that whatever furthers the statute's primary objective must be the law." Rodriguez v. United States, 480 U. S. 522, 525-526 (1987) ( per curiam) (emphasis deleted).
The competing values that Congress sought to effectuate by enacting the PLRA were reducing the number of frivolous filings, on one hand, while preserving prisoners' capacity to file meritorious claims, on the other. As explained by Senator Hatch when he introduced the legislation on the Senate floor, the PLRA was needed because the quantity of frivolous suits filed by prisoners was, in Senator Hatch's view, making it difficult for "courts to consider meritorious claims." 141 Cong. Rec. 27042 (1995). He continued: "Indeed, I do not want to prevent inmates from raising legitimate claims. This legislation will not prevent those claims from being raised." Ibid. Similarly, as Senator Thurmond, a cosponsor of the bill, stated: "[The PLRA] will allow meritorious claims to be filed, but gives the judge broader discretion to prevent frivolous and malicious lawsuits filed by prison inmates." Id., at 27044.
But the procedural default sanction created by this Court, unlike the exhaustion requirement created by Congress, bars litigation at random, irrespective of whether a claim is meritorious or frivolous. Consider, for example, an inmate who has been raped while in prison. Such a scenario is far from hypothetical; in enacting the Prison Rape Elimination Act of 2003, 42 U.S.C. § 15601 et seq. (2000 ed., Supp. III), Congress estimated that some one million people have been sexually assaulted in the Nation's prisons over the last 20 years, § 15601(2). Although not all of these tragic incidents result in constitutional violations, the sovereign does have a constitutional duty to "provide humane conditions of confinement," Farmer v. Brennan, 511 U. S. 825, 832 (1994). Accordingly, those inmates who are sexually assaulted by guards, or whose sexual assaults by other inmates are facilitated by guards, have suffered grave deprivations of their Eighth Amendment rights. Yet, the Court's engraftment of a procedural default sanction into the PLRA's exhaustion requirement risks barring such claims when a prisoner fails, inter alia, to file her grievance (perhaps because she correctly fears retaliation) within strict time requirements that are generally no more than 15 days, and that, in nine States, are between 2 and 5 days.
Indeed, if anything, it will have a worse effect on meritorious claims; prisoners who file frivolous claims are probably more likely to be repeat filers, and to learn the ins and outs of all procedural requirements.
See, e. g., Daskalea v. District of Columbia, 227 F. 3d 433, 437, 439 (CADC 2000) (discussing how female prisoner had her underwear confiscated as "`contraband'" and was placed in solitary confinement without a mattress as a result of talking to prison officials about the sexual assaults and harassment to which guards had subjected her).
For a comprehensive discussion of state prison grievance system filing deadlines, see Brief for American Civil Liberties Union et al. as Amici Curiae 6, n. 1, and Brief for Jerome N. Frank Legal Services Organization of Yale Law School as Amicus Curiae A1-A7.
Much of the majority opinion seems to assume that, absent the creation of a waiver sanction, prisoners will purposely circumvent prison grievance proceedings. However, prisoners generally lack both the incentive and the capacity to engage in such evasive tactics. Because federal courts do not provide any deference to administrative decisions by prison officials and any later federal suit is de novo, prisoners — even prisoners who are acting in bad faith — lack an incentive to avoid filing an administrative grievance unless they fear retaliation. Moreover, because prisoners must exhaust administrative remedies, prison officials can always thwart efforts by prisoners to avoid the grievance process by simply exercising their discretion to excuse any procedural defect in the presentation of the prisoners' claims.
At any rate, there is a simple solution that would allow courts to punish prisoners who seek to deliberately bypass state administrative remedies, but that would not impose the Draconian punishment of procedural default on prisoners who make reasonable, good-faith efforts to comply with relevant administrative rules but, out of fear of retaliation, a reasonable mistake of law, or simple inadvertence, make some procedural misstep along the way. Federal courts could simply exercise their discretion to dismiss suits brought by the former group of litigants but not those brought by the latter.
The majority argues that imposing a sanction against prisoners who deliberately bypass administrative remedies "neither has a statutory basis nor refers to a concept of exhaustion from an existing body of law," ante, at 98. In fact, this criticism applies to the majority's engraftment of an overinclusive procedural default sanction into the PLRA. If this Court insists upon rewriting § 1997e(a) in light of its understanding of the statute's purposes, surely the majority should add to the statute no harsher a sanction for making a procedural error during exhaustion than is necessary to accomplish its policy goals.
Moreover, ordinary abstention principles allow federal district courts to dismiss suits brought by prisoners who have deliberately bypassed available state remedies. Federal courts have the power to decline jurisdiction in exceptional circumstances, including the need to promote "wise judicial administration." Quackenbush v. Allstate Ins. Co., 517 U. S. 706, 716 (1996) (internal quotation marks omitted). Indeed, in Fay, we emphasized the discretion of district court judges in embracing precisely such a deliberate bypass regime in the habeas corpus statute. See 372 U. S., at 438. Applying such a deliberate bypass sanction to the PLRA would ensure that prisoners who act in bad faith are penalized, while not interfering with the capacity of other inmates to litigate meritorious constitutional claims.
In sum, the version of the PLRA Congress actually enacted, which includes an exhaustion requirement but not a procedural default sanction, is plainly sufficient to advance the policy values identified by the Court. Moreover, if, as the Court worries, there are many prisoners who act in bad faith and purposely eschew administrative remedies, the imposition of a deliberate bypass standard would resolve that problem, without depriving litigants who act in good faith but nonetheless make a procedural error from obtaining judicial relief relating to their valid constitutional claims. The majority's holding is as unsupported by the policy concerns it discusses as it is by the text of the statute.
The majority leaves open the question whether a prisoner's failure to comply properly with procedural requirements that do not provide a "meaningful opportunity for prisoners to raise meritorious grievances" would bar the later filing of a suit in federal court. Ante, at 102. What the majority has in mind by a "meaningful opportunity" is unclear, and this question is sure to breed a great deal of litigation in federal courts in the years to come.
For example, in this case, respondent filed a second grievance after his first grievance was rejected, arguing that his first grievance was in fact timely because he was challenging petitioners' continuing prohibition on his capacity to participate in Catholic observances, such as Confession, Holy Week services, and Bible study. The prison again rejected this second grievance on timeliness grounds, even though the denial of respondent's capacity to engage in religious activities was clearly ongoing, and thus had occurred within the prison's 15-day statute of limitations. See 403 F. 3d 620, 622 (CA9 2005). Assuming respondent explicitly requested the restoration of his right to engage in religious activities within 15 days of the filing of his second grievance and prison officials denied the request, did petitioners' grievance procedures fail to provide respondent with a "meaningful opportunity" to raise his claim, because, in light of the continuing nature of the injury respondent is challenging, his grievance was in fact timely? Cf. Klehr v. A. O. Smith Corp., 521 U. S. 179, 189 (1997) (explaining that, under the Clayton Act, each overt act in the case of a "continuing violation," such as a price-fixing conspiracy, is sufficient to restart the statute of limitations).
What about cases involving other types of procedural mis-steps? Does a 48-hour limitations period furnish a meaningful opportunity for a prisoner to raise meritorious grievances in the context of a juvenile who has been raped and repeatedly assaulted, with the knowledge and assistance of guards, while in detention? See Minix v. Pazera, No. 1:04 CV 447 RM, 2005 WL 1799538, *2 (ND Ind., July 27, 2005). Does a prison grievance system provide such a meaningful opportunity when women prisoners fail to file timely grievances relating to a pattern of rape and sexual harassment throughout a city's prisons, because they correctly fear retaliation if they file such complaints? See Women Prisoners v. District of Columbia, 877 F. Supp. 634 (DC 1994). Are such remedies meaningful when a prisoner files a grievance concerning a prison official having encouraged him to commit suicide, which the prisoner reasonably thinks raises one claim, but which prison officials interpret to raise two separate claims — one related to the guard's comments and one related to the prisoner's failure to receive health care — and thus dismiss for violating a prison regulation against including more than one claim in a single grievance? See Harper v. Laufenberg, No. 04-C-699-C, 2005 WL 79009, *3 (WD Wis., Jan. 6, 2005). What if prison officials dismiss a timely filed appeal because the prisoner explains that the prison will take two weeks to finish making certain copies of relevant documents by sending a letter to the Secretary of the Department of Corrections, rather than to the Secretary of Inmate Grievances and Appeals, as he should have under the prison regulations? See Keys v. Craig, 160 Fed. Appx. 125 (CA3 2005) ( per curiam). More generally, are remedies meaningful when prison officials refuse to hear a claim simply because a prisoner makes some hypertechnical procedural error? See Spruill v. Gillis, 372 F. 3d 218, 232 (CA3 2004) (imposing a procedural default sanction in the PLRA, but stating that compliance with grievance proceedings need only be "`substantial'"); Giano v. Goord, 380 F. 3d 670, 676-678 (CA2 2004) (stating that failure to comply with procedural requirements in grievance proceedings may be excused based on special circumstances, such as a prisoner's reasonable, but mistaken, understanding of prison regulations).
Depending on the answer to questions like these, the majority's interpretation of the PLRA may cause the statute to be vulnerable to constitutional challenges. "[T]he right of access to the courts is an aspect of the First Amendment right to petition the Government for redress of grievances." Bill Johnson's Restaurants, Inc. v. NLRB, 461 U. S. 731, 741 (1983). Accordingly, the Constitution guarantees that prisoners, like all citizens, have a reasonably adequate opportunity to raise constitutional claims before impartial judges, see, e. g., Lewis v. Casey, 518 U. S. 343, 351 (1996). Moreover, because access to the courts is a fundamental right, see id., at 346, government-drawn classifications that impose substantial burdens on the capacity of a group of citizens to exercise that right require searching judicial examination under the Equal Protection Clause, see, e. g., Lyng v. Automobile Workers, 485 U. S. 360, 370 (1988).
The correct interpretation of the PLRA would obviate the need for litigation over any of these issues. More importantly, the correct interpretation of the statute would recognize that, in enacting the PLRA, Members of Congress created a rational regime designed to reduce the quantity of frivolous prison litigation while adhering to their constitutional duty "to respect the dignity of all persons," even "those convicted of heinous crimes." Roper v. Simmons, 543 U. S. 551, 560 (2005). Because today's decision ignores that duty, I respectfully dissent.