UNITED STATES OF AMERICA v. ADRIAN CASTRO
No. 18-10137
United States Court of Appeals for the Fifth Circuit
March 17, 2022
Appeal from the United States District Court for the Northern District of Texas, USDC No. 3:16-cv-1761
Before HO, OLDHAM, and WILSON, Circuit Judges.
The original opinion in this case issued on July 14, 2021. The petition for rehearing is DENIED. We withdraw our previous opinion and substitute the following in its place.
A judge on our court granted Adrian Castro a certificate of appealability (“COA“). It‘s undisputed that the COA is invalid under
I.
Adrian Castro plotted and executed a spree of violent thefts against United States postal workers. The Government indicted Castro and charged him with, inter alia, violating
Twelve years later, Castro filed his first motion under
Castro then asked our court for a COA. One judge of our court granted a COA on a single procedural ground: “whether the district court erred by denying Castro‘s § 2255 motion as untimely.”
II.
Prisoners challenging their custody are not like ordinary litigants. For over a century, Congress has required prisoners—unlike anyone else appealing a judgment—to receive permission before appealing. We first explain that permission requirement. Then we vacate Castro‘s COA.
A.
In 1908, Congress took away the appeal-as-of-right from state prisoners. See An Act restricting in certain cases the right of appeal to the Supreme Court in habeas proceedings, 35 Stat. 40, 40 (1908). In its place, Congress instituted the certificate of probable cause (“CPC“) procedure. The CPC procedure required a state prisoner to obtain certification from “the United States court by which the final decision was rendered or a justice of the Supreme Court” that “probable cause for an appeal” existed. Ibid. This prerequisite to appeal served to preempt frivolous petitions and prevent the expenditure of precious judicial resources on meritless cases. See Davis v. Jacobs, 454 U.S. 911, 917 (1981) (Rehnquist, J., dissenting) (“[C]ongress[] . . . impose[d] th[e] [CPC] requirement as a means of terminating frivolous appeals in habeas corpus proceedings.“).
In 1948, Congress broadened the types of judicial officers empowered to grant CPCs to include circuit judges in addition to Supreme Court justices. See An Act to revise, codify, and enact into law title 28 of the United States Code entitled “Judicial Code and Judiciary,” 62 Stat. 869, 967 (1948). And although the statute did not designate the substantive standard for probable cause, the Supreme Court held that a prisoner seeking certification must offer a “substantial showing of the denial of a federal right.” Barefoot v. Estelle, 463 U.S. 880, 893 (1983) (quotation omitted).
A tidal shift occurred in 1996 when Congress enacted AEDPA, which overhauled the statutory framework governing habeas corpus with an eye towards “eliminat[ing] delays in the federal habeas review process.” Holland v. Florida, 560 U.S. 631, 648 (2010); see Pub. L. No. 104-132, 110 Stat. 1214 (1996). Section 2253(c)(2), as amended by AEDPA, retained the certification requirement but changed the name to a “certificate of appealability.” And instead of permitting an appeal anytime a prisoner made “a substantial showing of the denial of a federal right,” Barefoot, 463 U.S. at 893 (emphasis added), AEDPA elevated the standard and limited appeals to only those cases in which an applicant makes “a substantial showing of
Today, state and federal prisoners face the same hurdle to noticing an appeal: The applicant must obtain a COA by making “a substantial showing of the denial of a constitutional right.”
The must-identify-a-constitutional-issue requirement is not diminished where a district court denies relief on procedural grounds. See Slack v. McDaniel, 529 U.S. 473, 484 (2000). To obtain a COA in such a circumstance, an applicant must show “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.” Id. at 478. The rationale for these rules is simple: If a prisoner must eventually prove a constitutional violation to secure release from custody, his appeal should proceed only if he can prove a debatable constitutional issue at the outset. A procedural-only appeal is much ado about nothing. See, e.g., id. at 483–84 (holding that a COA applicant “must make a substantial showing of the denial of a constitutional right“); Miller-El v. Cockrell, 537 U.S. 322, 336 (2003) (“[Section] 2253(c) permits the issuance of a COA only where a petitioner has made a ‘substantial showing of the denial of a constitutional right.‘“).
Most recently, the Supreme Court confronted a COA issued by our court that is materially identical to the COA our court issued in this case. See Gonzalez, 565 U.S. at 138. Here are the two Fifth Circuit COAs side-by-side:
| Gonzalez v. Thaler | United States v. Castro |
|---|---|
| “whether the habeas application was timely filed” | “whether the district court erred by denying Castro‘s § 2255 motion as untimely” |
The Supreme Court unanimously agreed that such a COA is invalid because it says nothing at all about the Constitution. See Gonzalez, 565 U.S. at 141 (eight justices agreeing that a procedural-only COA is invalid); id. at 155 (Scalia, J., dissenting) (arguing that a procedural-only COA is invalid and also constitutes a jurisdictional defect). And although an invalid COA does not deprive us of jurisdiction, the Court nevertheless held that the commands in
Given Gonzalez and the Court‘s unanimous judgment, both sides unsurprisingly agree that Castro‘s COA is invalid. The Government asks us to vacate it. The Federal Public Defender (“FPD“) says that once a COA is issued, it cannot be vacated—no matter how badly it conflicts with the COA requirements enacted by Congress and affirmed by the unanimous judgment of the Supreme Court.
We agree with the Government. The Supreme Court has repeatedly admonished us that procedural-only COAs are invalid. We‘ve refused to follow those instructions before, and we‘ve been reversed for the refusal. Today we resolve to follow the statute that Congress wrote and to forswear procedural-only COAs. “Having sworn off the habit of venturing beyond Congress‘s intent, we will not accept [the] invitation to have one last drink.” Alexander v. Sandoval, 532 U.S. 275, 287 (2001).
This approach accords with other habeas doctrines. Take for example Teague v. Lane, 489 U.S. 288 (1989). Like the COA requirement, the nonretroactivity doctrine serves the interests of judicial economy, efficiency, and administration. See, e.g., Mackey v. United States, 401 U.S. 667, 676 (1971) (Harlan, J., concurring in the judgment) (noting the nonretroactivity “doctrine was the product of the Court‘s disquietude with the impacts of its fast-moving pace of constitutional innovation in the criminal field” and “a technique that provided an impetus for the implementation of long overdue reforms, which otherwise could not be practicably effected” (quotation omitted)). Like an invalid COA, the nonretroactivity of a Supreme Court decision under Teague is a non-jurisdictional defense. And where the State fails to raise Teague, the federal court can raise it sua sponte and dismiss the habeas petition. See Caspari v. Bohlen, 510 U.S. 383, 389 (1994).
Or take AEDPA‘s one-year time bar. See
Third, take procedural default. Like the COA requirement, the procedural-default doctrine serves the interests of judicial economy, efficiency, and administration. See Magourik v. Phillips, 144 F.3d 348, 358 (5th Cir. 1998). Like an invalid COA, procedural default is a non-jurisdictional defense. And where the State fails to raise a default, the federal court can raise it sua sponte and dismiss the habeas petition. Ibid.; see also Brewer v. Marshall, 119 F.3d 993, 999 (1st Cir. 1997); Rosario v. United States, 164 F.3d 729, 732 (2d Cir. 1998); Sweger v. Chesney, 294 F.3d 506, 520 (3d Cir. 2002); Yeatts v. Angelone, 166 F.3d 255, 261 (4th Cir. 1999); Sowell v. Bradshaw, 372 F.3d 821, 830 (6th Cir. 2004); Kurzawa v. Jordan, 146 F.3d 435, 440 (7th Cir. 1998); King v. Kemna, 266 F.3d 816, 822 (8th Cir. 2001) (en banc); Vang v. Nevada, 329 F.3d 1069, 1073 (9th Cir. 2003); United States v. Wiseman, 297 F.3d 975, 979 (10th Cir. 2002); Moon v. Head, 285 F.3d 1301, 1315 n.17 (11th Cir. 2002).
Given the plain text of
C.
The FPD nonetheless says Gonzalez is somehow inconsistent with vacating Castro‘s COA. That misreads the Supreme Court‘s decision.
Gonzalez reiterated the principle—well-settled since the landmark decisions in Slack and Miller-El—that a COA comports with the mandatory language used by Congress only if it “indicates which specific issue or issues” constitute “a substantial showing of the denial of a constitutional right.”
Nor does Gonzalez‘s reasoning require that result. In Gonzalez, no one identified the invalidity of the COA until after briefing in our court, after argument in our court, after a precedential decision from our court, and after a cert petition in the Supreme Court. It was not until the State‘s brief in opposition that anyone noticed the COA problem. See Gonzalez, 565 U.S. at 145. On those facts, vacating the COA at such a late date would serve no “gatekeeping” function whatsoever. Ibid. To the contrary, it would create serious inefficiencies—the same ones the COA is designed to prevent—to vacate a COA after our court has already rendered its decision. Ibid.
Those concerns do not apply here. The parties conceded the invalidity of our COA
That does not mean—as the FPD seems to misunderstand—that our court must always raise COA defects sua sponte, as we must with jurisdictional defects. To hold otherwise, we would have to equate the two types of defects in a way that Gonzalez forbids. Rather, we hold only that our court can raise COA defects sua sponte—just as it can raise numerous other defects in postconviction proceedings. See supra pp. 6–7 (discussing retroactivity, the statute of limitations, procedural default, and exhaustion). We reject the FPD‘s “extreme position[]” that COA defects alone are somehow beyond our cognizance to recognize sua sponte. Granberry, 481 U.S. at 133.
III.
In some cases, our court and not the movant is to blame for a COA defect. See Gonzalez, 565 U.S. at 144 (“A petitioner, having successfully obtained a COA, has no control over how the judge drafts the COA and . . . may have done everything required of him by law.“). How, then, do we prevent our mistakes from hurting otherwise-faultless movants?
The answer is simple. When we spot a defective COA, on our own initiative or otherwise, it should be vacated. Then we can, in our discretion, consider issuing a valid COA. See United States v. Smith, 945 F.3d 860, 863 (5th Cir. 2019) (“We can correct this nonjurisdictional defect ‘by considering an amendment to the COA.‘” (quoting Gonzalez, 565 U.S. at 146)). Of course, we are bound by the well-settled rule that a prisoner cannot ask for a COA here on any ground different from the one(s) submitted to and rejected by the district court. See Black v. Davis, 902 F.3d 541, 545 (5th Cir. 2018). But assuming the prisoner did his part correctly and asked for a valid COA in the district court and then again in ours, the rule we announce today will not hurt him. He‘ll either get the COA in the first instance or, if we make a mistake and issue an invalid certificate, we can recognize our mistake and fix it with a valid one.
Castro asks us to issue a valid COA on “whether the residual clause found in
Even so, we cannot grant Castro‘s request. Castro was not sentenced under the residual clause in
* * *
The COA we originally granted was invalid, so we now vacate it. And because Castro has not made “a substantial showing of the denial of a constitutional right,” see Slack, 529 U.S. at 483–84, we cannot grant an amended COA, see Smith, 945 F.3d at 863. We therefore dismiss for lack of jurisdiction. See Gonzalez, 565 U.S. at 142 (lack of a COA destroys jurisdiction).
PETITION FOR REHEARING DENIED; COA VACATED; APPEAL DISMISSED.