Justice SOTOMAYOR delivered the opinion of the Court.

This case interprets two provisions of the Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA). The first, 28 U.S.C. § 2253(c), provides that a habeas petitioner must obtain a certificate of appealability (COA) to appeal a federal district court's final order in a habeas proceeding. § 2253(c)(1). The COA may issue only if the petitioner has made a "substantial showing of the denial of a constitutional right," § 2253(c)(2), and "shall indicate which specific issue" satisfies that showing. § 2253(c)(3). We hold that § 2253(c)(3) is not a jurisdictional requirement. Accordingly, a judge's failure to "indicate" the requisite constitutional issue in a COA does not deprive a court of appeals of subject-matter jurisdiction to adjudicate the habeas petitioner's appeal.

The second provision, 28 U.S.C. § 2244(d)(1)(A), establishes a 1-year limitations period for state prisoners to file federal habeas petitions, running from "the date on which the judgment became final by the conclusion of direct review or the expiration of the time for seeking such review." We hold that, for a state prisoner who does not seek review in a State's highest court, the judgment becomes "final" on the date that the time for seeking such review expires.

I

Petitioner Rafael Gonzalez was convicted of murder in Texas state court. The intermediate state appellate court, the Texas Court of Appeals, affirmed Gonzalez's conviction on July 12, 2006. Gonzalez then allowed his time for seeking discretionary review with the Texas Court of Criminal Appeals (Texas CCA)—the State's highest court for criminal appeals—to expire on August 11, 2006. Tex. Rule App. Proc. 68.2(a) (2011). The Texas Court of Appeals issued its mandate on September 26, 2006.

After Gonzalez, proceeding pro se, petitioned unsuccessfully for state habeas relief, he filed a federal habeas petition under 28 U.S.C. § 2254 on January 24, 2008, in the U.S. District Court for the Northern District of Texas. His petition alleged, inter alia, that the nearly 10-year delay between his indictment and trial violated his Sixth Amendment right to a speedy trial. The District Court, without discussing Gonzalez's constitutional claims, dismissed Gonzalez's petition as time barred by the 1-year statute of limitations in § 2244(d)(1)(A). Although Gonzalez argued that his judgment had not become final until the Texas Court of Appeals issued its mandate, the District Court held that Gonzalez's judgment had become final when his time for seeking discretionary review in the Texas CCA expired on August 11, 2006. Counting from that date, and tolling the limitations period for the time during which Gonzalez's state habeas petition was pending, Gonzalez's limitations period elapsed on December 17, 2007—over a month before he filed his federal habeas petition. The District Court denied a COA.

Gonzalez applied to the U.S. Court of Appeals for the Fifth Circuit for a COA on two grounds: (1) his habeas petition was timely, and (2) his Sixth Amendment speedy-trial right was violated. A Court of Appeals judge granted a COA on the question "whether the habeas application was timely filed because Gonzalez's conviction became final, and thus the limitations period commenced, on the date the intermediate state appellate court issued its mandate." App. 347. The COA did not mention the Sixth Amendment question.

The Court of Appeals affirmed. 623 F.3d 222 (2010). Acknowledging that a sister Circuit had run the limitations period from the date of a state court's issuance of a mandate, the Court of Appeals deemed the mandate's issuance "irrelevant" to determining finality under § 2244(d)(1)(A). Id., at 224, 226 (disagreeing with Riddle v. Kemna, 523 F.3d 850 (C.A.8 2008) (en banc)). The Court of Appeals held that because a judgment becomes final at "the conclusion of direct review or the expiration of the time for seeking such review," § 2244(d)(1)(A), the limitations period begins to run for petitioners who fail to appeal to a State's highest court when the time for seeking further direct review in the state court expires. The Court of Appeals therefore concluded that Gonzalez's conviction became final on August 11, 2006, and his habeas petition was time barred.

The Court of Appeals did not address Gonzalez's Sixth Amendment claim or discuss whether the COA had been improperly issued. Nor did the State allege any defect in the COA or move to dismiss for lack of jurisdiction.

Gonzalez petitioned this Court for a writ of certiorari. In its brief in opposition, the State argued for the first time that the Court of Appeals lacked jurisdiction to adjudicate Gonzalez's appeal because the COA identified only a procedural issue, without also "indicat[ing]" a constitutional issue as required by § 2253(c)(3). We granted certiorari to decide two questions, both of which implicate splits in authority: (1) whether the Court of Appeals had jurisdiction to adjudicate Gonzalez's appeal, notwithstanding the § 2253(c)(3) defect;1 and (2) whether Gonzalez's habeas petition was time barred under § 2244(d)(1) due to the date on which his judgment became final.2 564 U.S. ___, 131 S.Ct. 2989, 180 L.Ed.2d 820 (2011).

II

We first consider whether the Court of Appeals had jurisdiction to adjudicate Gonzalez's appeal.

A

Section 2253, as amended by AEDPA, governs appeals in habeas corpus proceedings. The first subsection, § 2253(a), is a general grant of jurisdiction, providing that district courts' final orders in habeas proceedings "shall be subject to review, on appeal, by the court of appeals." 28 U.S.C. § 2253(a). The second, § 2253(b), limits jurisdiction over a particular type of final order. See § 2253(b) ("There shall be no right of appeal from a final order in a proceeding to test the validity of a warrant [of] remov[al] ..."). This case concerns the third, § 2253(c), which provides:

When, as here, the district court denies relief on procedural grounds, the petitioner seeking a COA must show both "that jurists of reason would find it debatable whether the petition states a valid claim of the denial of a constitutional right and that jurists of reason would find it debatable whether the district court was correct in its procedural ruling." Slack v. McDaniel, 529 U.S. 473, 484, 120 S.Ct. 1595, 146 L.Ed.2d 542 (2000).

In this case, the Court of Appeals judge granted a COA that identified a debatable procedural ruling, but did not "indicate" the issue on which Gonzalez had made a substantial showing of the denial of a constitutional right, as required by § 2253(c)(3). The question before us is whether that defect deprived the Court of Appeals of the power to adjudicate Gonzalez's appeal. We hold that it did not.

This Court has endeavored in recent years to "bring some discipline" to the use of the term "jurisdictional." Henderson v. Shinseki, 562 U.S. ___, ___, 131 S.Ct. 1197, 1202-1203, 179 L.Ed.2d 159 (2011). Recognizing our "less than meticulous" use of the term in the past, we have pressed a stricter distinction between truly jurisdictional rules, which govern "a court's adjudicatory authority," and nonjurisdictional "claim-processing rules," which do not. Kontrick v. Ryan, 540 U.S. 443, 454-455, 124 S.Ct. 906, 157 L.Ed.2d 867 (2004). When a requirement goes to subject-matter jurisdiction, courts are obligated to consider sua sponte issues that the parties have disclaimed or have not presented. See United States v. Cotton, 535 U.S. 625, 630, 122 S.Ct. 1781, 152 L.Ed.2d 860 (2002). Subject-matter jurisdiction can never be waived or forfeited. The objections may be resurrected at any point in the litigation, and a valid objection may lead a court midway through briefing to dismiss a complaint in its entirety. "[M]any months of work on the part of the attorneys and the court may be wasted." Henderson, 562 U.S., at ___, 131 S.Ct., at 1202. Courts, we have said, should not lightly attach those "drastic" consequences to limits Congress has enacted. Ibid.

We accordingly have applied the following principle: A rule is jurisdictional "[i]f the Legislature clearly states that a threshold limitation on a statute's scope shall count as jurisdictional." Arbaugh v. Y & H Corp., 546 U.S. 500, 515, 126 S.Ct. 1235, 163 L.Ed.2d 1097 (2006). But if "Congress does not rank a statutory limitation on coverage as jurisdictional, courts should treat the restriction as nonjurisdictional." Id., at 516, 126 S.Ct. 1235.3 That clear-statement principle makes particular sense in this statute, as we consider— against the backdrop of § 2253(a)'s clear jurisdictional grant to the courts of appeals and § 2253(b)'s clear limit on that grant— the extent to which Congress intended the COA process outlined in § 2253(c) to further limit the courts of appeals' jurisdiction over habeas appeals.

Here, the only "clear" jurisdictional language in § 2253(c) appears in § 2253(c)(1). As we explained in Miller-El v. Cockrell, 537 U.S. 322, 123 S.Ct. 1029, 154 L.Ed.2d 931 (2003), § 2253(c)(1)'s plain terms—"Unless a circuit justice or judge issues a certificate of appealability, an appeal may not be taken to the court of appeals"—establish that "until a COA has been issued federal courts of appeals lack jurisdiction to rule on the merits of appeals from habeas petitioners." Id., at 336, 123 S.Ct. 1029. The parties thus agree that § 2253(c)(1) is jurisdictional.

The parties also agree that § 2253(c)(2) is nonjurisdictional.4 That is for good reason. Section 2253(c)(2) speaks only to when a COA may issue—upon "a substantial showing of the denial of a constitutional right." It does not contain § 2253(c)(1)'s jurisdictional terms. See Russello v. United States, 464 U.S. 16, 23, 104 S.Ct. 296, 78 L.Ed.2d 17 (1983) ("[W]here Congress includes particular language in one section of a statute but omits it in another section of the same Act, it is generally presumed that Congress acts intentionally ..."). And it would be passing strange if, after a COA has issued, each court of appeals adjudicating an appeal were dutybound to revisit the threshold showing and gauge its "substantial[ity]" to verify its jurisdiction. That inquiry would be largely duplicative of the merits question before the court.

It follows that § 2253(c)(3) is nonjurisdictional as well. Like § 2253(c)(2), it too reflects a threshold condition for the issuance of a COA—the COA's indication of "which specific issue or issues satisfy the showing required by paragraph (2)." It too "does not speak in jurisdictional terms or refer in any way to the jurisdiction of the [appeals] courts." Arbaugh, 546 U.S., at 515, 126 S.Ct. 1235 (internal quotation marks omitted). The unambiguous jurisdictional terms of §§ 2253(a), (b), and (c)(1) show that Congress would have spoken in clearer terms if it intended § 2253(c)(3) to have similar jurisdictional force. Instead, the contrast underscores that the failure to obtain a COA is jurisdictional, while a COA's failure to indicate an issue is not. A defective COA is not equivalent to the lack of any COA.

It is telling, moreover, that Congress placed the power to issue COAs in the hands of a "circuit justice or judge."5 It would seem somewhat counterintuitive to render a panel of court of appeals judges powerless to act on appeals based on COAs that Congress specifically empowered one court of appeals judge to grant. Indeed, whereas § 2253(c)(2)'s substantial-showing requirement at least describes a burden that "the applicant" seeking a COA bears, § 2253(c)(3)'s indication requirement binds only the judge issuing the COA. Notably, Gonzalez advanced both the timeliness and Sixth Amendment issues in his application for a COA. A petitioner, having successfully obtained a COA, has no control over how the judge drafts the COA and, as in Gonzalez's case, may have done everything required of him by law. That fact would only compound the "unfai[r] prejudice" resulting from the sua sponte dismissals and remands that jurisdictional treatment would entail. Henderson, 562 U.S., at ___, 131 S.Ct., at 1202.6

Treating § 2253(c)(3) as jurisdictional also would thwart Congress' intent in AEDPA "to eliminate delays in the federal habeas review process." Holland v. Florida, 560 U.S. ___, ___, 130 S.Ct. 2549, 2562, 177 L.Ed.2d 130 (2010). The COA process screens out issues unworthy of judicial time and attention and ensures that frivolous claims are not assigned to merits panels. Once a judge has made the determination that a COA is warranted and resources are deployed in briefing and argument, however, the COA has fulfilled that gatekeeping function. Even if additional screening of already-issued COAs for § 2253(c)(3) defects could further winnow the cases before the courts of appeals, that would not outweigh the costs of further delay from the extra layer of review. This case, in which the alleged defect would be dispositive, exemplifies those inefficiencies; the State requests that we vacate and remand with instructions to dismiss the appeal based on a § 2253(c)(3) defect that it raised for the first time in response to a petition for certiorari. And delay would be particularly fruitless in the numerous cases where, as here, the district court dismissed the petition on procedural grounds and the court of appeals affirms, without having to address the omitted constitutional issue at all.

B

The State, aided by the United States as amicus curiae, makes several arguments in support of jurisdictional treatment of § 2253(c)(3). None is persuasive.

First, the State notes that although § 2253(c)(3) does not speak in jurisdictional terms, it refers back to § 2253(c)(1), which does. The State argues that it is as if § 2253(c)(1) provided: "Unless a circuit justice or judge issues a certificate of appealability that shall indicate the specific issue or issues that satisfy the showing required by paragraph (2), an appeal may not be taken to the court of appeals." The problem is that the statute provides no such thing. Instead, Congress set off the requirements in distinct paragraphs and, rather than mirroring their terms, excluded the jurisdictional terms in one from the other. Notably, the State concedes that § 2253(c)(2) is nonjurisdictional, even though it too cross-references § 2253(c)(1) and is cross-referenced by § 2253(c)(3).

Second, the State seizes on the word "shall" in § 2253(c)(3), arguing that an omitted indication renders the COA no COA at all. But calling a rule nonjurisdictional does not mean that it is not mandatory or that a timely objection can be ignored. If a party timely raises the COA's failure to indicate a constitutional issue, the court of appeals panel must address the defect by considering an amendment to the COA or remanding to the district judge for specification of the issues.7 This Court, moreover, has long "rejected the notion that `all mandatory prescriptions, however emphatic, are ... properly typed jurisdictional.'" Henderson, 562 U.S., at ___, 131 S.Ct., at 1205; see also Dolan v. United States, 560 U.S. ___, ___, 130 S.Ct. 2533, 2539, 177 L.Ed.2d 108 (2010) (statute's reference to "shall" alone does not render statutory deadline jurisdictional). Nothing in § 2253(c)(3)'s prescription establishes that an omitted indication should remain an open issue throughout the case.

Third, the United States argues that the placement of § 2253(c)(3) in a section containing jurisdictional provisions signals that it too is jurisdictional. In characterizing certain requirements as nonjurisdictional, we have on occasion observed their "`separat[ion]'" from jurisdictional provisions. E.g., Reed Elsevier, Inc. v. Muchnick, 559 U.S. ___, ___, 130 S.Ct. 1237, 1244-1245, 176 L.Ed.2d 18 (2010); Arbaugh, 546 U.S., at 515, 126 S.Ct. 1235. The converse, however, is not necessarily true: Mere proximity will not turn a rule that speaks in nonjurisdictional terms into a jurisdictional hurdle. In fact, § 2253(c)(3)'s proximity to §§ 2253(a), (b), and (c)(1) highlights the absence of clear jurisdictional terms in § 2253(c)(3).

Finally, the State analogizes a COA to a notice of appeal, pointing out that both a notice and its contents are jurisdictional prerequisites. Federal Rule of Appellate Procedure 3(c)(1) provides that a notice of appeal must: "(A) specify the party or parties taking the appeal"; "(B) designate the judgment, order, or part thereof being appealed"; and "(C) name the court to which the appeal is taken." We have held that "Rule 3's dictates are jurisdictional in nature." Smith v. Barry, 502 U.S. 244, 248, 112 S.Ct. 678, 116 L.Ed.2d 678 (1992).

We reject this analogy. We construed the content requirements for notices of appeal as jurisdictional because we were "convinced that the harshness of our construction [wa]s `imposed by the legislature.'" Torres v. Oakland Scavenger Co., 487 U.S. 312, 318, 108 S.Ct. 2405, 101 L.Ed.2d 285 (1988). Rule 4, we noted, establishes mandatory time limits for filing a notice of appeal. Excusing a failure to name a party in a notice of appeal, in violation of Rule 3, would be "equivalent to permitting courts to extend the time for filing a notice of appeal," in violation of Rule 4. Id., at 315, 108 S.Ct. 2405. And "time limits for filing a notice of appeal have been treated as jurisdictional in American law for well over a century." Bowles v. Russell, 551 U.S. 205, 209, n. 2, 127 S.Ct. 2360, 168 L.Ed.2d 96 (2007). Accordingly, the Advisory Committee Note "makes no distinction among the various requirements of Rule 3 and Rule 4," treating them "as a single jurisdictional threshold." Torres, 487 U.S., at 315, 108 S.Ct. 2405; see also id., at 316, 108 S.Ct. 2405 ("the Advisory Committee viewed the requirements of Rule 3 as jurisdictional in nature"). Here, we find no similar basis for treating the paragraphs of § 2253(c) as a single jurisdictional threshold.

Moreover, in explaining why the naming requirement was jurisdictional in Torres, we reasoned that an unnamed party leaves the notice's "intended recipient[s]"—the appellee and court—"unable to determine with certitude whether [that party] should be bound by an adverse judgment or held liable for costs or sanctions." Id., at 318, 108 S.Ct. 2405. The party could sit on the fence, await the outcome, and opt to participate only if it was favorable. That possibility of gamesmanship is not present here. Unlike the party who fails to submit a compliant notice of appeal, the habeas petitioner who obtains a COA cannot control how that COA is drafted.8 And whereas a party's failure to be named in a notice of appeal gives absolutely no "notice of [his or her] appeal," a judge's issuance of a COA reflects his or her judgment that the appeal should proceed and supplies the State with notice that the habeas litigation will continue.

Because we conclude that § 2253(c)(3) is a nonjurisdictional rule, the Court of Appeals had jurisdiction to adjudicate Gonzalez's appeal.

III

We next consider whether Gonzalez's habeas petition was time barred. AEDPA establishes a 1-year limitations period for state prisoners to file for federal habeas relief, which "run[s] from the latest of" four specified dates.9 § 2244(d)(1). This case concerns the first of those dates: "the date on which the judgment became final by the conclusion of direct review or the expiration of the time for seeking such review." § 2244(d)(1)(A). The question before us is when the judgment becomes "final" if a petitioner does not appeal to a State's highest court.

A

In construing the language of § 2244(d)(1)(A), we do not write on a blank slate. In Clay v. United States, 537 U.S. 522, 123 S.Ct. 1072, 155 L.Ed.2d 88 (2003), we addressed AEDPA's statute of limitations for federal prisoners seeking postconviction relief. See § 2255(f)(1) (2006 ed., Supp. III) (beginning 1-year period of limitations from "the date on which the judgment of conviction becomes final"). We held that the federal judgment becomes final "when this Court affirms a conviction on the merits on direct review or denies a petition for a writ of certiorari," or, if a petitioner does not seek certiorari, "when the time for filing a certiorari petition expires." Id., at 527, 123 S.Ct. 1072. In so holding, we rejected the argument that, if a petitioner declines to seek certiorari, the limitations period "starts to run on the date the court of appeals issues its mandate." Id., at 529, 123 S.Ct. 1072.

In Jimenez v. Quarterman, 555 U.S. 113, 129 S.Ct. 681, 172 L.Ed.2d 475 (2009), we described Clay's interpretation as comporting "with the most natural reading of the statutory text" and saw "no reason to depart" from it in "construing the similar language of § 2244(d)(1)(A)." 555 U.S., at 119, 129 S.Ct. 681. The state court had permitted Jimenez to file an out-of-time direct appeal. We held that this "reset" the limitations period; Jimenez's judgment would now become final at "the conclusion of the out-of-time direct appeal, or the expiration of the time for seeking review of that [out-of-time] appeal." Id., at 120-121, 129 S.Ct. 681. Because Jimenez did not seek certiorari, we made no mention of when the out-of-time appeal "conclu[ded]." Rather, we held that his judgment became final when his "time for seeking certiorari review in this Court expired." Id., at 120, 129 S.Ct. 681. Nor did we mention the date on which the state court issued its mandate. Both Clay and Jimenez thus suggested that the direct review process either "concludes" or "expires," depending on whether the petitioner pursues or forgoes direct appeal to this Court.

We now make clear what we suggested in those cases: The text of § 2244(d)(1)(A), which marks finality as of "the conclusion of direct review or the expiration of the time for seeking such review," consists of two prongs. Each prong—the "conclusion of direct review" and the "expiration of the time for seeking such review"—relates to a distinct category of petitioners. For petitioners who pursue direct review all the way to this Court, the judgment becomes final at the "conclusion of direct review"—when this Court affirms a conviction on the merits or denies a petition for certiorari. For all other petitioners, the judgment becomes final at the "expiration of the time for seeking such review"—when the time for pursuing direct review in this Court, or in state court, expires. We thus agree with the Court of Appeals that because Gonzalez did not appeal to the State's highest court, his judgment became final when his time for seeking review with the State's highest court expired.

B

Gonzalez offers an alternative reading of § 2244(d)(1)(A): Courts should determine both the "conclusion of direct review" and the "expiration of the time for seeking such review" for every petitioner who does not seek certiorari, then start the 1-year clock from the "latest of" the two dates. Gonzalez rejects our uniform definition of the "conclusion of direct review" as the date on which this Court affirms a conviction on the merits or denies a petition for certiorari. In his view, whenever a petitioner does not seek certiorari, the "conclusion of direct review" is the date on which state law marks finality—in Texas, the date on which the mandate issues. Ex parte Johnson, 12 S.W.3d 472, 473 (Crim. App. 2000) (per curiam). Applying this approach, Gonzalez contends that his habeas petition was timely because his direct review "concluded" when the mandate issued (on September 26, 2006), later than the date on which his time for seeking Texas CCA review "expired" (August 11, 2006). We find his construction of the statute unpersuasive.

First, Gonzalez lacks a textual anchor for his later-in-time approach. The words "latest of" do not appear anywhere in § 2244(d)(1)(A). Rather, they appear in § 2244(d)(1) and refer to the "latest of" the dates in subparagraphs (A), (B), (C), and (D)—the latter three of which are inapplicable here. Nothing in § 2244(d)(1)(A) contemplates any conflict between the "conclusion of direct review" and the "expiration of the time for seeking such review," much less instructs that the later of the two shall prevail.

Nor is Gonzalez's later-in-time reading necessary to give both prongs of § 2244(d)(1)(A) full effect. Our reading does so by applying one "or" the other, depending on whether the direct review process concludes or expires. Treating the judgment as final on one date "or" the other is consistent with the disjunctive language of the provision.

Second, Gonzalez misreads our precedents. Gonzalez asserts that in Jimenez, we made a later-in-time choice between the two prongs. That is mistaken. Rather, we chose between two "expiration" dates corresponding to different appeals: Jimenez initially failed to appeal to the Texas Court of Appeals and that appeal became final when his "time for seeking discretionary review ... expired." 555 U.S., at 117, 119, 129 S.Ct. 681. When Jimenez was later allowed to file an out-of-time appeal, he pursued appeals with both the Texas Court of Appeals and Texas CCA; the out-of-time appeal thus became final when his "[t]ime for seeking certiorari review ... with this Court expired." Id., at 116, 120, 129 S.Ct. 681. We adopted the out-of-time appeal's date of finality over the initial appeal's date of finality. Id., at 119-121, 129 S.Ct. 681. Critically, by deeming the initial appeal final at the expiration of time for seeking review in state court, and the out-of-time appeal final at the expiration of time for seeking certiorari in this Court, we reinforced Clay's suggestion that the "expiration" prong governs all petitioners who do not pursue direct review all the way to this Court.10

Third, Gonzalez argues that AEDPA's federalism concerns and respect for state-law procedures mean that we should not read § 2244(d)(1)(A) to disregard state law. We agree. That is why a state court's reopening of direct review will reset the limitations period. 555 U.S., at 121, 129 S.Ct. 681. That is also why, just as we determine the "expiration of the time for seeking [direct] review" from this Court's filing deadlines when petitioners forgo certiorari, we look to state-court filing deadlines when petitioners forgo state-court appeals. Referring to state-law procedures in that context makes sense because such deadlines are inherently court specific. There is no risk of relying on "state-law rules that may differ from the general federal rule." Clay, 537 U.S., at 531, 123 S.Ct. 1072.

By contrast, Gonzalez urges us to scour each State's laws and cases to determine how it defines finality for every petitioner who forgoes a state-court appeal. That approach would usher in state-by-state definitions of the conclusion of direct review. It would be at odds with the uniform definition we adopted in Clay and accepted in the § 2244(d)(1)(A) context in Jimenez. And it would pose serious administrability concerns. Even if roughly "half of the States define the conclusion of direct review as the issuance of the mandate or similar process," Brief for Petitioner 40, that still leaves half with either different rules or no settled rules at all.11

Fourth, Gonzalez speculates that our reading will rob some habeas petitioners of the full 1-year limitations period. Gonzalez asserts that our reading starts the clock running from the date that his time for seeking Texas CCA review expired, even though, under Texas law, he could not file for state habeas relief until six weeks later, on the date the Texas Court of Appeals issued its mandate. Tex.Code Crim. Proc. Ann., Art. 11.07, § 3(a) (Vernon Supp. 2011). His inability to initiate state habeas proceedings during those six weeks, he argues, reduced his 1-year federal habeas filing period by six weeks. We expect, however, that it will be a rare situation where a petitioner confronting similar state laws faces a delay in the mandate's issuance so excessive that it prevents him or her from filing a federal habeas petition within a year.12 A petitioner who has exhausted his or her claims in state court need not await state habeas proceedings to seek federal habeas relief on those claims. To the extent a petitioner has had his or her federal filing period severely truncated by a delay in the mandate's issuance and has unexhausted claims that must be raised on state habeas review, such a petitioner could file a request for a stay and abeyance from the federal district court. See Rhines v. Weber, 544 U.S. 269, 277, 125 S.Ct. 1528, 161 L.Ed.2d 440 (2005).

Finally, Gonzalez argues, as an alternative to his later-in-time construction, that his petition should be considered timely because it was filed within a year of when his time for seeking this Court's review— as opposed to the Texas CCA's review— expired. We can review, however, only judgments of a "state court of last resort" or of a lower state court if the "state court of last resort" has denied discretionary review. This Court's Rule 13.1; see also 28 U.S.C. § 1257(a) (2006 ed.). Because Gonzalez did not appeal to the Texas CCA, this Court would have lacked jurisdiction over a petition for certiorari from the Texas Court of Appeals' decision affirming Gonzalez's conviction. We therefore decline to incorporate the 90-day period for seeking certiorari in determining when Gonzalez's judgment became final.

* * *

In sum, we hold that § 2253(c)(3) is a mandatory but nonjurisdictional rule. Here, the COA's failure to "indicate" a constitutional issue did not deprive the Court of Appeals of jurisdiction to adjudicate Gonzalez's appeal. We further hold that, with respect to a state prisoner who does not seek review in a State's highest court, the judgment becomes "final" under § 2244(d)(1)(A) when the time for seeking such review expires—here, August 11, 2006. We thus agree with the Court of Appeals that Gonzalez's federal habeas petition was time barred.

For the reasons stated, the judgment of the Court of Appeals for the Fifth Circuit is

Affirmed.

Justice SCALIA, dissenting.

The obvious, undeniable, purpose of 28 U.S.C. § 2253(c) is to spare three-judge courts of appeals the trouble of entertaining (and the prosecution the trouble of defending against) appeals from the denials of relief in habeas and § 2255 proceedings, unless a district or circuit judge has identified an issue on which the applicant has made a substantial showing of a constitutional violation. Where no such constitutional issue has been identified, an appeal on other, nonconstitutional, issues (such as the statute of limitations issue that the Court decides today) will not lie.

Today's opinion transforms this into a provision that allows appeal so long as a district or circuit judge, for whatever reason or for no reason at all, approves it. This makes a hash of the statute. The opinion thinks this alchemy required by the Court's previously expressed desire to "`bring some discipline' to the use of the term `jurisdictional,'" ante, at 648 (quoting Henderson v. Shinseki, 562 U.S. ___, ___, 131 S.Ct. 1197, 1202, 179 L.Ed.2d 159 (2011)). If that is true, discipline has become a code word for eliminating inconvenient statutory limits on our jurisdiction. I would reverse the judgment below for want of jurisdiction.

I

Fair Meaning of the Text

Congress amended § 2253 to its current form in the Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA). In its entirety, the section reads as follows:

As the Court acknowledges, ante, at 648-649, all three subsections—(a), (b), and (c)—clearly speak to the jurisdiction of the courts of appeals. Subsection (a) gives appellate jurisdiction to "the court of appeals for the circuit in which the proceeding is held"; subsection (b) carves out certain classes of cases from that appellate jurisdiction; and subsection (c) imposes a procedural hurdle to the exercise of that appellate jurisdiction—a judge's issuance of a certificate of appealability, see Miller-El v. Cockrell, 537 U.S. 322, 336, 123 S.Ct. 1029, 154 L.Ed.2d 931 (2003).

Paragraph 2253(c)(3) says that a certificate of appealability must "indicate" which issue or issues in the case involve a substantial showing of a constitutional violation. Everyone agrees that the certificate issued below contains no such indication. See ante, at 648. It appears, in fact, that the issuing judge never considered whether any of Gonzalez's constitutional claims satisfied paragraph (2). As far as we know, no federal judge has ever determined that Gonzalez "has made a substantial showing of the denial of a constitutional right." § 2253(c)(2). The Court does not even suggest that he has—but it goes on to decide the statute-of-limitations issue in the case.

Its basis for proceeding in this fashion is the remarkable statement that "[a] defective COA is not equivalent to the lack of any COA." Ante, at 649. That is simply not true with respect to a significant defect in a legal document. Would one say that a deed which lacks the words of conveyance is not equivalent to the lack of a deed? Or that a passport which lacks the Secretary of State's affirmance of the bearer's citizenship is not equivalent to the lack of a passport? Minor technical defects are one thing, but a defect that goes to the whole purpose of the instrument is something else. And the whole purpose of the certificate-of-appealability procedure is to make sure that, before a case can proceed to the court of appeals, a judge has made the determination that it presents a substantial showing of the denial of a constitutional right. To call something a valid certificate of appealability which does not contain the central finding that is the whole purpose of a certificate of appealability is quite absurd.

The Court says that "[o]nce a judge has made the determination that a COA is warranted and resources are deployed in briefing and argument, ... the COA has fulfilled [its] gatekeeping function." Ante, at 650. But of course it has not done so— it has performed no gatekeeping function whatever—if "the determination that a COA is warranted" has not been accompanied by the issuing judge's opinion required to support the determination: that there is an issue as to which the applicant has made a "substantial showing of the denial of a constitutional right," § 2253(c)(2). As the very next sentence of today's opinion discloses, what the Court means by "has fulfilled [its] gatekeeping function" is simply that it will not be worth the trouble of going back, since that would "not outweigh the costs of further delay," ante, at 650.

That is doubtless true, and it demonstrates the hollowness of the Court's assurance that "calling a rule nonjurisdictional does not mean that it is not mandatory or that a timely objection can be ignored," ante, at 651. That statement is true enough as a general proposition: Calling the numerosity requirement in Arbaugh v. Y & H Corp., 546 U.S. 500, 126 S.Ct. 1235, 163 L.Ed.2d 1097 (2006), nonjurisdictional, for example, did not eliminate it, where protest was made, as a continuing mandatory requirement for relief on the merits, id., at 516, 126 S.Ct. 1235. Even the time-of-filing requirement in Eberhart v. United States, 546 U.S. 12, 126 S.Ct. 403, 163 L.Ed.2d 14 (2005) (per curiam), continued to have "bite" even though it was held nonjurisdictional: It prevented relief when the failure to observe it was properly challenged, id., at 19, 126 S.Ct. 403. But the Court has managed to create today a "mandatory" requirement which—precisely because it will not be worth the trouble of going back—has no practical, real-world effect.1 What is the consequence when the issuing judge, over properly preserved objection, produces a COA like the one here, which does not contain the required opinion? None whatever. The habeas petitioner already has what he wants, argument before the court of appeals. The government, for its part, is either confident in its view that there has been no substantial showing of denial of a constitutional right—in which case it is just as easy (if not easier) to win before three judges as it is before one; or else it is not—in which case a crusade to enforce § 2253(c) is likely to yield nothing but additional litigation expenses. As for the three-judge panel of the court of appeals, it remains free, as always, to choose whichever mandatory-but-not-jurisdictional basis it wishes for resolving the case. Cf. Steel Co. v. Citizens for Better Environment, 523 U.S. 83, 93-94, 118 S.Ct. 1003, 140 L.Ed.2d 210 (1998). Why not choose the one that is sure to be final and that might avoid embarrassing a colleague? No one has any interest in enforcing the "mandatory" requirement. Which is perhaps why, as I proceed to discuss, mandatory requirements for court-to-court appeal are always made jurisdictional.

Past Treatment of Similar Provisions

As the Court acknowledges, "`context, including this Court's interpretation of similar provisions in many years past, is relevant to whether a statute ranks a requirement as jurisdictional.'" Ante, at 648-649, n. 3 (quoting Reed Elsevier, Inc. v. Muchnick, 559 U.S. ___, ___, 130 S.Ct. 1237, 1248, 176 L.Ed.2d 18 (2010)). Thus, we have said that a requirement prescribed as a condition to obtaining judicial review of agency action is quite different (nonjurisdictional) from a requirement prescribed as a condition to appeal from one court to another (jurisdictional). See Henderson, 562 U.S., at ___ _ ___, 131 S.Ct., at 1203-1204. We have always— always, without exception—held that procedural conditions for appealing a case from one Article III court to another are jurisdictional. When an appeal is "not taken within the time prescribed by law," the "Court of Appeals [is] without jurisdiction." George v. Victor Talking Machine Co., 293 U.S. 377, 379, 55 S.Ct. 229, 79 S.Ct. 439 (1934) (per curiam); see also United States v. Robinson, 361 U.S. 220, 229-230, 80 S.Ct. 282, 4 L.Ed.2d 259 (1960). When a party's name is not listed in the notice of appeal, as the Federal Rules of Appellate Procedure require, the court has no jurisdiction over that party's appeal. Torres v. Oakland Scavenger Co., 487 U.S. 312, 314-315, 108 S.Ct. 2405, 101 L.Ed.2d 285 (1988).

When this Court reviewed cases by writ of error, the law required that the lower-court record be filed with the Court "before the end of the term next succeeding the issue of the writ." Edmonson v. Bloomshire, 7 Wall. 306, 309, 19 S.Ct. 91 (1869). The Court routinely dismissed cases that did not comply with that requirement. See, e.g., Mesa v. United States, 2 Black 721, 721-722, 17 S.Ct. 350 (1863) (per curiam); Edmonson, supra, at 309-310; Steamer Virginia v. West, 19 How. 182, 183, 15 S.Ct. 594 (1857). The same jurisdictional treatment was accorded to failure to serve notice on the defendant in error within the succeeding term, see, e.g., United States v. Curry, 6 How. 106, 112-113, 12 S.Ct. 363 (1848); Villabolos v. United States, 6 How. 81, 88, 91, 12 S.Ct. 352 (1848), and to failure to file the writ of error with the clerk of the lower court, see, e.g., Credit Co. v. Arkansas Central R. Co., 128 U.S. 258, 261, 9 S.Ct. 107, 32 S.Ct. 448 (1888); Scarborough v. Pargoud, 108 U.S. 567, 2 S.Ct. 877, 27 S.Ct. 824 (1883). Today, when a petition for certiorari in a civil case is not filed within the time prescribed by 28 U.S.C. § 2101(c), this Court lacks jurisdiction. Federal Election Comm'n v. NRA Political Victory Fund, 513 U.S. 88, 90, 115 S.Ct. 537, 130 L.Ed.2d 439 (1994) (citing Missouri v. Jenkins, 495 U.S. 33, 45, 110 S.Ct. 1651, 109 L.Ed.2d 31 (1990)); see also Matton S.S. Co. v. Murphy, 319 U.S. 412, 415, 63 S.Ct. 1126, 87 S.Ct. 1483 (1943) (per curiam).2

So strict has been the rule enforcing as jurisdictional those requirements attached to court-from-court appeals, that we have applied it to a requirement contained in a statute not even addressed to the courts. Section 518(a) of Title 28 charges the Solicitor General with "conduct[ing] and argu[ing] suits and appeals in the Supreme Court ... in which the United States is interested." We held that, absent independent statutory authority, an agency's petition for certiorari filed without authorization from the Solicitor General does not suffice to invoke our jurisdiction. NRA Political Victory Fund, supra, at 98-99.3

Jurisdictional enforcement of procedural requirements for appeal has deep roots in our jurisprudence. Chief Justice Taney dismissed an appeal in which the citation was not issued and served in time, because "we have no power to receive an appeal in any other mode than that provided by law." Villabolos, supra, at 90. And Chief Justice Chase wrote, in a case dismissing an appeal for failure to file in time:

Jurisdictional Nature of Predecessor Provision

But similarity to a general type of provision that has always been held jurisdictional is not all that supports the jurisdictional character of § 2253(c)(3). Its very predecessor statute made a judge's expression of opinion a condition of appellate jurisdiction. The certificate of probable cause, of which the COA was born, arrived on the scene over 100 years ago in "An Act Restricting in certain cases the right of appeal to the Supreme Court in habeas corpus proceedings," Act of Mar. 10, 1908, ch. 76, 35 Stat. 40:

The last version of this statute, before it was amended to its current form in AEDPA, provided for issuance of the certificate of probable cause by a circuit judge instead of a justice. See § 2253, 62 Stat. 967 (codified at 28 U.S.C. § 2253). Even applying the Court's simplistic rule that the jurisdictional restriction must be contained in the very same paragraph as the procedural requirement, there is no doubt that under this statute a judge's certification that there was probable cause for an appeal was jurisdictional. See, e.g., Ex parte Patrick, 212 U.S. 555, 29 S.Ct. 686, 53 S.Ct. 650 (1908) (per curiam); Bilik v. Strassheim, 212 U.S. 551, 29 S.Ct. 684, 53 S.Ct. 649 (1908) (per curiam). There is no reason whatever to think that Congress rendered the statement of opinion unnecessary for jurisdiction by (1) extending the requirement for it to § 2255 proceedings; (2) requiring the opinion to address a more specific point (not just probable cause for an appeal but presence of an issue presenting a "substantial showing of the denial of a constitutional right")4; and (3) giving the document in which the judge is required to express the opinion a name ("certificate of appealability")—so that now a "certificate of appealability" without opinion will suffice. Neither any one of these steps, nor all of them combined, suggest elimination of jurisdictional status for the required expression of opinion.5 It would be an entirely strange way of achieving that result. It was not a strange way, however, of dividing the now more complex and lengthy provision into manageable subsections.

Stare Decisis Effect of Torres

In addition to the fact that conditions attached to court-to-court appeal have always been held jurisdictional, and the fact that this statute's predecessor was held to be so, we have considered, and found to be jurisdictional, a statute presenting precisely what is at issue here: a provision governing court-to-court appeals which made particular content a required element of a document that the statute said was necessary for jurisdiction; and which did that in a separate section that "excluded the jurisdictional terms," ante, at 651. That case flatly contradicts today's holding. In Torres v. Oakland Scavenger Co., 487 U.S. 312, 108 S.Ct. 2405, 101 L.Ed.2d 285, we dealt with Rule 3(c)(1) of the Federal Rules of Appellate Procedure. Rule 3(a) of those Rules makes a notice of appeal necessary to appellate jurisdiction—just as § 2253(c)(1) makes a certificate of appealability necessary. And Rule 3(c)(1), which, like § 2253(c)(3), does not contain jurisdictional language, says what the requisite notice of appeal must contain—just as § 2253(c)(3) says what the requisite certificate of appealability must contain:

In Torres we held that the Court of Appeals lacked jurisdiction over the appeal of a party not properly named in the notice of appeal. 487 U.S., at 314-315, 108 S.Ct. 2405. The parallel is perfect.

The Court claims that the jurisdictional consequences of Rule 3(c) were "`imposed by the legislature,'" ante, at 652 (quoting Torres, supra, at 318, 108 S.Ct. 2405), which according to the Court's analysis "`clearly state[d],'" ante, at 648 (quoting Arbaugh, 546 U.S., at 515, 126 S.Ct. 1235), that Rule 3(c) is jurisdictional. But the legislature there did precisely what it did here: made a particular document necessary to jurisdiction and then specified what that document must contain.6 I certainly agree that that is a clear statement that a document with the requisite content is necessary to jurisdiction. But the Court does not. So to distinguish Torres it has to find something else in Rule 3(c) that provided a "clear statement" of what "Congress intended," ante, at 648-649. The best it can come up with, ante, at 652, is an un clear statement, and that not from Congress but from Advisory Committee Notes referred to in the Torres opinion. Such Notes are (of course) "the product of the Advisory Committee, and not Congress," and "they are transmitted to Congress before the rule is enacted into law." United States v. Vonn, 535 U.S. 55, 64, n. 6, 122 S.Ct. 1043, 152 L.Ed.2d 90 (2002). They are, in other words, a species of legislative history. I know of no precedent for the proposition that legislative history can satisfy a clear-statement requirement imposed by this Court's opinions. Does today's distinguishing of Torres mean that legislative history can waive the sovereign immunity of the United States? See United States v. Nordic Village, Inc., 503 U.S. 30, 33-34, 112 S.Ct. 1011, 117 L.Ed.2d 181 (1992). Or abrogate the sovereign immunity of the States? See Atascadero State Hospital v. Scanlon, 473 U.S. 234, 242, 105 S.Ct. 3142, 87 L.Ed.2d 171 (1985). Or give retroactive effect to new legislation? See Greene v. United States, 376 U.S. 149, 160, 84 S.Ct. 615, 11 L.Ed.2d 576 (1964). Or foreclose review of agency actions? See Abbott Laboratories v. Gardner, 387 U.S. 136, 141, 87 S.Ct. 1507, 18 L.Ed.2d 681 (1967). Today's opinion is in this respect a time-bomb.

To make matters worse, the Advisory Committee Note considered by the Torres Court—as "support for [its] view," 487 U.S., at 315, 108 S.Ct. 2405—did not clearly say that Rule 3(c)'s requirements were jurisdictional. It said this:

To say that timely filing of a notice of appeal is jurisdictional, and that placing within the notice of appeal what Rule 3 says it must contain is "of the utmost importance," does not remotely add up to a clear statement that placing within the notice of appeal what Rule 3 says it must contain is jurisdictional. There is simply no principled basis for saying that Torres satisfies the "clear-statement principle," ante, at 649, except the commonsense notion that when a document is made jurisdictional, and the required contents of that document specified, a document that does not contain those contents cannot confer jurisdiction.7

The Court is not willing to say that Torres is no longer good law, but I doubt whether future litigants will be so coy. They know that in the past, to avoid the uncongenial rigidity of the rule that procedures attending court-to-court appeals are jurisdictional, we have performed wondrous contortions to find compliance with those rules. For example, in Smith v. Barry, 502 U.S. 244, 248, 112 S.Ct. 678, 116 L.Ed.2d 678 (1992), we held that an "informal brief" filed after a defective notice of appeal counted as a valid notice of appeal. In Foman v. Davis, 371 U.S. 178, 181, 83 S.Ct. 227, 9 L.Ed.2d 222 (1962), we held that a notice of appeal from the denial of a motion to vacate the judgment was also a notice of appeal from the underlying judgment. And in Houston v. Lack, 487 U.S. 266, 270, 108 S.Ct. 2379, 101 L.Ed.2d 245 (1988), we held that a prisoner's notice of appeal was "filed" when it was delivered to prison authorities for forwarding to the district court. These (shall we say) creative interpretations of the procedural requirements were made necessary by the background principle that is centuries old: "[I]f the mode prescribed for removing cases by writ of error or appeal be too strict and technical, and likely to produce inconvenience or injustice, it is for Congress to provide a remedy by altering the existing laws; not for the court." United States v. Curry, 6 How. 106, 113, 12 S.Ct. 363 (1848). But if we have been willing to expose ourselves to ridicule in order to approve implausible compliance with procedural prerequisites to appeal, surely we may be willing to continue and expand the process of simply converting those obnoxious prerequisites into the now favored "claims processing rules," enabling us to avoid unseemly contortions by simply invoking the ever-judge-friendly principles of equity.

What began as an effort to "`bring some discipline' to the use of the term `jurisdictional,'" ante, at 648 (quoting Henderson, 562, U.S., at ___, 131 S.Ct., at 1202), shows signs of becoming a libertine, liberating romp through our established jurisprudence.

II

A few remaining points raised by the Court's opinion warrant response.

The Court holds that the requirement imposed by paragraph (c)(2) (that a COA may issue "only if the applicant has made a substantial showing of the denial of a constitutional right") is not jurisdictional, and says that "[i]t follows that § 2253(c)(3) is nonjurisdictional as well." Ante, at 649. I need not reach the issue whether (c)(2) is jurisdictional—though it seems to me that the Court disposes rather summarily of the Solicitor General's view that it is. And I need not confront the Court with the back-at-you argument that if (c)(3) is jurisdictional (as I think) then (c)(2) is as well. For whether one runs it backwards or forwards, the argument is a bad one. Assuming that (c)(2) is nonjurisdictional, it does not at all "follow" that (c)(3) is nonjurisdictional as well. Paragraph (c)(3) is jurisdictional not because it is located in subsection (c), but because it describes the required content of a COA. Paragraph (c)(2) does not; it sets forth the criterion for a COA's issuance. A judge may apply that criterion erroneously but still produce a COA that (as paragraph (c)(3) requires) "indicate[s] which specific issue or issues satisfy the showing required by paragraph (2)." It no more follows that the erroneousness of the judge's indication must destroy the jurisdiction that the COA creates, than it followed under the predecessor statute that the erroneousness of the certification of probable cause for an appeal destroyed the jurisdiction that the certification created.8 The two issues are quite separate: what the judge must find, and what the COA (or certification) must contain.

The Court points out that Gonzalez raised the Sixth Amendment issue in his application for a COA, that "[a] petitioner, having successfully obtained a COA, has no control over how the judge drafts the COA," and that the petitioner, "as in Gonzalez's case, may have done everything required of him by law." Ante, at 650. Perhaps it is true that the defective COA was not at all Gonzalez's fault—though he could have promptly moved to amend it. But no-fault elimination of jurisdiction is not forbidden. In Bowles v. Russell, 551 U.S. 205, 127 S.Ct. 2360, 168 L.Ed.2d 96 (2007), we enforced a time limit on notice of appeal where the district court had purported to extend the time to file and the appellant had complied with the court's order. Id., at 207, 213-214, 127 S.Ct. 2360. It did not matter that the fault lay with the court.

Finally, the Court points out that treating § 2253(c)(3) as jurisdictional would waste a lot of time. "Even if additional screening of already-issued COAs for § 2253(c)(3) defects could further winnow the cases before the courts of appeals, that would not outweigh the costs of further delay from the extra layer of review." Ante, at 650. But that is not an argument directed to the statute before us; it is an argument directed against enforcement of all jurisdictional requirements (all of which, I suspect, are the object of the Court's mounting disfavor). And the argument may not even be true, except in the (presumably rare) case where the jurisdictional prescription is disregarded. Over the long term, the time saved to judges and lawyers by an enforceable requirement that appeals be screened by a single judge may vastly outweigh the time wasted by the occasional need for enforcement. That, it seems to me, is what Congress believed.

* * *

Terminology is destiny. Today's holding, and the erosion of our prior jurisprudence that will perhaps follow upon it, is foreshadowed and facilitated by the unfortunate terminology with which we have chosen to accompany our campaign to "bring some discipline" to determinations of jurisdiction. We have said that the universe of rules placing limitations upon the courts is divided into (1) "claims processing rules," and (2) jurisdiction-removing rules. Unless our prior jurisprudence is to be repudiated, that is a false dichotomy. The requirement that the unsuccessful litigant file a timely notice of appeal, for example, is (if the term is to have any meaning) a claims-processing rule, ordering the process by which claims are adjudicated. Yet as discussed above, that, and all procedures that must be followed to proceed from one court to another, have always been deemed jurisdictional. The proper dichotomy is between claims processing rules that are jurisdictional, and those that are not. To put it otherwise suggests a test for jurisdiction that is not to be found in our cases.9

At the end of the day, the indication requirement in § 2253(c)(3) is "`imposed by the legislature and not by the judicial process.'" Torres, 487 U.S., at 318, 108 S.Ct. 2405 (quoting Schiavone v. Fortune, 477 U.S. 21, 31, 106 S.Ct. 2379, 91 L.Ed.2d 18 (1986)). Whether or not its enforcement leads to a harsh result, wastes time in this particular case, or (though the Court does not give this as a reason) prevents us from reaching a circuit conflict we are dying to resolve, we are obliged to enforce it. I respectfully dissent.