AYESTAS v. DAVIS
NOTE: Where it is feasible, a syllabus (headnote) will be released, as is being done in connection with this case, at the time the opinion is issued.The syllabus constitutes no part of the opinion of the Court but has been prepared by the Reporter of Decisions for the convenience of the reader.See United States v. Detroit Timber & Lumber Co., 200 U. S. 321.
SUPREME COURT OF THE UNITED STATES
Ayestas, aka Zelaya Corea v. Davis, Director, Texas Department of Criminal Justice, Correctional Institutions Division
certiorari to the united states court of appeals for the fifth circuit
Petitioner Ayestas was convicted of murder and sentenced to death in a Texas state court. He secured new counsel, but his conviction and sentence were affirmed on appeal. A third legal team sought, unsuccessfully, state habeas relief, claiming trial-level ineffective assistance of counsel but not counsel’s failure to investigate petitioner’s mental health and alcohol and drug abuse during the trial’s penalty phase. His fourth set of attorneys did raise that failure in a federal habeas petition, but because the claim had never been raised in state court, the District Court held, it was barred by procedural default. That decision was vacated and remanded for reconsideration in light of Martinez v. Ryan, 566 U. S. 1—where this Court held that an Arizona prisoner seeking federal habeas relief could overcome the procedural default of a trial-level ineffective-assistance-of-counsel claim by showing that the claim is substantial and that state habeas counsel was also ineffective in failing to raise the claim in a state habeas proceeding—and Trevino v. Thaler, 569 U. S. 413—which extended that holding to Texas prisoners. Petitioner filed an ex parte motion asking the District Court for funding to develop his claim that both his trial and state habeas counsel were ineffective, relying on 18 U. S. C. §3599(f), which provides, in relevant part, that a district court “may authorize” funding for “investigative, expert, or other services . . . reasonably necessary for the representation of the defendant.” The court found his claim precluded by procedural default and thus denied his funding request. The Fifth Circuit also rejected the funding claim under its precedent: that a §3599(f) funding applicant must show that he has a “substantial need” for investigative or other services, and that funding may be denied when an applicant fails to present “a viable constitutional claim that is not procedurally barred.” 817 F. 3d 888, 895–896.
1. The District Court’s denial of petitioner’s funding request was a judicial decision subject to appellate review under the standard jurisdictional provisions. Pp. 7–14.
(a) Title 28 U. S. C. §§1291, 2253, and 1254 confer jurisdiction to review decisions made by a district court in a judicial capacity. “Administrative” decisions—about, e.g., facilities, personnel, equipment, supplies, and rules of procedure—are “not subject to [this Court’s] review,” Hohn v. United States, 524 U. S. 236, but the District Court’s ruling here does not remotely resemble such decisions. Petitioner’s request was made by motion in his federal habeas proceeding, which is indisputably a judicial proceeding. And resolution of the funding question requires the application of a legal standard—whether the funding is “reasonably necessary” for effective representation—that demands an evaluation of petitioner’s prospects of obtaining habeas relief. Pp. 8–10.
(b) Respondent’s arguments in support of her claim that §3599’s funding requests are nonadversarial and administrative are unpersuasive. First, that the requests can be decided ex parte does not make the proceeding nonadversarial. The habeas proceeding here was clearly adversarial. And petitioner and respondent plainly have adverse interests on the funding question and have therefore squared off as adversaries. The mere fact that a §3599 funding request may sometimes be made ex parte is thus hardly dispositive. Second, nothing in §3599 even hints that the funding decisions may be revised by the Director of the Administrative Office of the Courts. Lower court cases that appear to have accepted Administrative Office review of certain Criminal Justice Act (CJA) payments, even if a proper interpretation of the CJA, are inapposite. Finally, the fact that §3599(g)(2) requires funding in excess of the generally applicable statutory cap to be approved by the circuit’s chief judge or another designated circuit judge, instead of by a panel of three, does not make the proceeding administrative. If Congress wishes to make certain rulings reviewable by a single circuit judge, the Constitution does not stand in the way. Pp. 10–14.
2. The Fifth Circuit did not apply the correct legal standard in affirming the denial of petitioner’s funding request. Section 3599 authorizes funding for the “reasonably necessary” services of experts, investigators, and the like. But the Fifth Circuit’s requirement that applicants show a “substantial need” for the services is arguably a more demanding standard. Section 3599 appears to use the term “necessary” to mean something less than essential. Because it makes little sense to refer to something as being “reasonably essential,” the Court concludes that the statutory phrase calls for the district court to determine, in its discretion, whether a reasonable attorney would regard the services as sufficiently important, guided by considerations detailed in the opinion. The term “substantial” in the Fifth Circuit’s test, however, suggests a heavier burden. And that court exacerbated the difference by also requiring a funding applicant to present “a viable constitutional claim that is not procedurally barred.” That rule that is too restrictive after Trevino, see 569 U. S. at 429, because, in cases where funding stands a credible chance of enabling a habeas petitioner to overcome the procedural default obstacle, it may be error for a district court to refuse funding. That being said, district courts were given broad discretion in assessing funding requests when Congress changed the phrase “shall authorize” in §3599’s predecessor statute, see 21 U. S. C. §848(q)(9), to “may authorize” in §3599(f). A funding applicant must not be expected to prove that he will be able to win relief if given the services, but the “reasonably necessary” test does require an assessment of the likely utility of the services requested.
Respondent’s alternative ground for affirmance—that funding is never “reasonably necessary” where a habeas petitioner seeks to present a procedurally defaulted ineffective-assistance-of-trial-counsel claim that depends on facts outside the state-court record—remains open for the Fifth Circuit to consider on remand. Pp. 14–19.
817 F. 3d 888, vacated and remanded.
Alito, J., delivered the opinion for a unanimous Court. Sotomayor, J., filed a concurring opinion, in which Ginsburg, J., joined.
NOTICE: This opinion is subject to formal revision before publication in the preliminary print of the United States Reports. Readers are requested to notify the Reporter of Decisions, Supreme Court of the United States, Washington, D. C. 20543, of any typographical or other formal errors, in order that corrections may be made before the preliminary print goes to press.
SUPREME COURT OF THE UNITED STATES
CARLOS MANUEL AYESTAS, aka DENNIS ZELAYA COREA, PETITIONER v. LORIE DAVIS, DIRECTOR, TEXAS DEPARTMENT OF CRIMINAL JUSTICE, CORRECTIONAL INSTITUTIONS DIVISION
on writ of certiorari to the united states court of appeals for the fifth circuit
Justice Alito delivered the opinion of the Court.
Petitioner Carlos Ayestas, who was convicted of murder and sentenced to death in a Texas court, argues that he was wrongfully denied funding for investigative services needed to prove his entitlement to federal habeas relief. Petitioner moved for funding under 18 U. S. C. §3599(f ), which makes funds available if they are “reasonably necessary,” but petitioner’s motion was denied. We hold that the lower courts applied the wrong legal standard, and we therefore vacate the judgment below and remand for further proceedings.
In 1997, petitioner was convicted of capital murder in a Texas court. Evidence at trial showed that he and two accomplices invaded the home of a 67-year-old Houston woman, Santiaga Paneque, bound her with duct tape and electrical cord, beat and strangled her, and then made off with a stash of her belongings.
The jury also heard testimony from Henry Nuila regarding an incident that occurred about two weeks after the murder. Petitioner was drunk at the time, and he revealed to Nuila that he had recently murdered a woman in Houston. Petitioner then brandished an Uzi machinegun and threatened to murder Nuila if he did not help petitioner kill his two accomplices. Fortunately for Nuila, petitioner kept talking until he eventually passed out; Nuila then called the police, who arrested petitioner, still in possession of the gun.
After the jury found petitioner guilty, it was asked to determine whether he should be sentenced to death or to life in prison. In order to impose a death sentence, Texas law required the jury to answer the following three questions. First, would petitioner pose a continuing threat to society? Second, had he personally caused the death of the victim, intended to kill her, or anticipated that she would be killed? Third, in light of all the evidence surrounding the crime and petitioner’s background, were there sufficient mitigating circumstances to warrant a sentence of life without parole instead of death? Tex. Code Crim. Proc. Ann., Art. 37.071, §§2(b), (e) (Vernon Cum. Supp. 2017). Only if the jury gave a unanimous yes to the first two questions, and a unanimous no to the third question, could a death sentence be imposed; otherwise, petitioner would receive a sentence of life without parole. See §§2(d)(2), (f )(2), (g).
In asking the jury to impose a death sentence, the prosecution supplemented the trial record with evidence of petitioner’s criminal record and his encounter with a man named Candelario Martinez a few days after the murder. Martinez told the jury that he was standing in a hotel parking lot waiting for a friend when petitioner approached and began to make small talk. Before long, petitioner pulled out a machinegun and forced Martinez into a room where two of petitioner’s compatriots were holding Martinez’s friend at knifepoint. Ordered to lie down on the bathroom floor and await his execution, Martinez begged for his life while petitioner and his cohorts haggled about who would carry out the killing. Finally, petitioner relented, but he threatened to kill Martinez and his family if he contacted the police. Petitioner then stole Martinez’s truck.
Petitioner’s trial counsel presented very little mitigation evidence. This was due, at least in part, to petitioner’s steadfast refusal for many months to allow his lawyers to contact his family members, who were living in Honduras and might have testified about his character and upbringing. Petitioner gave in on the eve of trial, and at that point, according to the state habeas courts, his lawyers “made every effort to contact [his] family.” App. 171. They repeatedly contacted petitioner’s family members and urged them to attend the trial; they requested that the U. S. Embassy in Honduras facilitate family members’ travel to the United States; and they met in person with the Honduran Consulate to seek assistance. But these efforts were to no avail. Petitioner’s sister told his legal team that the family would not leave Honduras because the journey would create economic hardship and because their father was ill and had killed one of their neighbors. A defense attorney who spoke to petitioner’s mother testified that she seemed unconcerned about her son’s situation. In general, the state habeas courts found, petitioner “did nothing to assist counsel’s efforts to contact his family and did not want them contacted by the consulate or counsel.” Id., at 174.
In the end, the only mitigation evidence introduced by petitioner’s trial counsel consisted of three letters from petitioner’s English instructor. The letters, each two sentences long, described petitioner as “a serious and attentive student who is progressing well in English.” Ibid.
The jury unanimously concluded that petitioner should be sentenced to death, and a capital sentence was imposed. Petitioner secured new counsel to handle his appeal, and his conviction and sentence were affirmed by the Texas Court of Criminal Appeals in 1998. Ayestas v. State, No. 72,928, App. 115. Petitioner did not seek review at that time from this Court.
While petitioner’s direct appeal was still pending, a third legal team filed a habeas petition on his behalf in state court. This petition included several claims of trial-level ineffective assistance of counsel, but the petition did not assert that trial counsel were ineffective for failing to investigate petitioner’s mental health and abuse of alcohol and drugs. Petitioner’s quest for state habeas relief ended unsuccessfully in 2008. Ex parte Ayestas, No. WR–69,674–01 (Tex. Ct. Crim. App., Sept. 10, 2008), 2008 WL 4151814 (per curiam) (unpublished).
In 2009, represented by a fourth set of attorneys, petitioner filed a federal habeas petition under 28 U. S. C. §2254, and this time he did allege that his right to the effective assistance of counsel at trial was violated because his attorneys failed to conduct an adequate search for mitigation evidence. As relevant here, petitioner argued that trial counsel overlooked evidence that he was mentally ill and had a history of drug and alcohol abuse. Ayestasv. Thaler, Civ. Action No. H–09–2999 (SD Tex., Jan. 26, 2011), 2011 WL 285138, *4. Petitioner alleged that he had a history of substance abuse, and he noted that he had been diagnosed with schizophrenia while the state habeas proceeding was still pending. See Pet. for Writ of Habeas Corpus in Ayestas v. Quarterman, No. 4:09–cv–2999 (SD Tex.), Doc. 1, pp. 21–23. Petitioner claimed that trial counsel’s deficient performance caused prejudice because there was a reasonable chance that an adequate investigation would have produced mitigation evidence that would have persuaded the jury to spare his life.
Among the obstacles standing between petitioner and federal habeas relief, however, was the fact that he never raised this trial-level ineffective-assistance-of-counsel claim in state court. The District Court therefore held that the claim was barred by procedural default, Ayestas v. Thaler, 2011 WL 285138, *4–*7, and the Fifth Circuit affirmed, Ayestas v. Thaler, 462 Fed. Appx. 474, 482 (2012) (per curiam).
Petitioner sought review in this Court, and we vacated the decision below and remanded for reconsideration in light of two of our subsequent decisions, Martinez v. Ryan, 566 U. S. 1 (2012) , and Trevino v. Thaler, 569 U. S. 413 (2013) . Ayestas v. Thaler, 569 U. S. 1015 (2013) . Martinez held that an Arizona prisoner seeking federal habeas relief could overcome the procedural default of a trial-level ineffective-assistance-of-counsel claim by showing that the claim is substantial and that state habeas counsel was also ineffective in failing to raise the claim in a state habeas proceeding. 566 U. S., at 14. Trevino extended that holding to Texas prisoners, 569 U. S., at 416–417, and on remand, petitioner argued that he fell within Trevino because effective state habeas counsel would have uncovered evidence showing that trial counsels’ investigative efforts were deficient.
To assist in developing these claims, petitioner filed an ex parte motion asking the District Court for $20,016 in funding to conduct a search for evidence supporting his petition. He relied on 18 U. S. C. §3599(f ), which provides in relevant part as follows:
“Upon a finding that investigative, expert, or other services are reasonably necessary for the representation of the defendant, whether in connection with issues relating to guilt or the sentence, the court may authorize the defendant’s attorneys to obtain such services on behalf of the defendant and, if so authorized, shall order the payment of fees and expenses therefor.”
Petitioner averred that the funds would be used to conduct an investigation that would show that his trial counsel and his state habeas counsel were ineffective. Accordingly, he claimed, the investigation would establish both thathis trial-level ineffective-assistance-of-counsel claim was not barred by procedural default and that he was entitled to resentencing based on the denial of his Sixth Amendment right to the effective assistance of trial counsel.
The District Court refused the funding request and ultimately denied petitioner’s habeas petition. Ayestas v. Stephens, Civ. Action No. H–09–2999, (SD Tex., Nov. 18, 2014), 2014 WL 6606498, *6–*7. On the merits of petitioner’s new ineffective-assistance-of-trial-counsel claim, the District Court held that petitioner failed both prongs of the Strickland test. See Strickland v. Washington, 466 U. S. 668 (1984) . Noting that most of the evidence bearing on petitioner’s mental health had emerged only after he was sentenced, the court concluded that petitioner’s trial lawyers were not deficient in failing to find such evidence in time for the sentencing proceeding. 2014 WL 6606498, *5. In addition, the court found that state habeas counsel did not render deficient performance by failing to investigate petitioner’s history of substance abuse, and that, in any event, petitioner was not prejudiced at the sentencing phase of the trial or during the state habeas proceedings because the potential mitigation evidence at issue would not have made a difference to the jury in light of “the extremely brutal nature of [the] crime and [petitioner’s] history of criminal violence.” Ibid.
With respect to funding, the District Court pointed to Fifth Circuit case law holding that a §3599(f ) funding applicant cannot show that investigative services are “ ‘reasonably necessary’ ” unless the applicant can show that he has a “ ‘substantial need’ ” for those services. Id., at *6. In addition, the court noted that “[t]he Fifth Circuit upholds the denial of funding” when, among other things, “a petitioner has . . . failed to supplement his funding request with a viable constitutional claim that is not procedurally barred.” Ibid. (internal quotation marks omitted).
Given its holding that petitioner’s new ineffective-assistance-of-counsel claim was precluded by procedural default, this rule also doomed his request for funding. The District Court denied petitioner’s habeas petition and refused to grant him a certificate of appealability (COA). Id., at *7. On appeal, the Fifth Circuit held that a COA was not needed for review of the funding issue, but it rejected that claim for essentially the same reasons as the District Court, citing both the “substantial need” test and the rule that funding may be denied when a funding applicant fails to present “a viable constitutional claim that is not procedurally barred.” Ayestas v. Stephens, 817 F. 3d 888, 895–896 (2016) (internal quotation marks omitted). With respect to petitioner’s other claims, including his claim of ineffective assistance of trial counsel, the Fifth Circuit refused to issue a COA. Id., at 898.
We granted certiorari to decide whether the lower courts applied the correct legal standard in denying the funding request. 581 U. S. ___ (2017).
Before we reach that question, however, we must consider a jurisdictional argument advanced by respondent, the Director of the Texas Department of Criminal Justice. 1 Respondent contends that the District Court’s denial of petitioner’s funding request was an administrative, not a judicial, decision and therefore falls outside the scope of the jurisdictional provisions on which petitioner relied in seeking review in the Court of Appeals and in this Court.
When the District Court denied petitioner’s funding request and his habeas petition, he took an appeal to the Fifth Circuit under 28 U. S. C. §§1291 and 2253, which grant the courts of appeals jurisdiction to review final “decisions” and “orders” of a district court. 2 And when the Fifth Circuit affirmed, petitioner sought review in this Court under §1254, which gives us jurisdiction to review “[c]ases” in the courts of appeals. 3 As respondent correctly notes, these provisions confer jurisdiction to review decisions made by a district court in a judicial capacity. But we have recognized that not all decisions made by a fed-eral court are “judicial” in nature; some decisions are prop-erly understood to be “administrative,” and in that case they are “not subject to our review.” Hohn v. United States, 524 U. S. 236, 245 (1998) .
The need for federal judges to make many administrative decisions is obvious. The Federal Judiciary, while tiny in comparison to the Executive Branch, is nevertheless a large and complex institution, with an annual budget exceeding $7 billion and more than 32,000 employees.See Administrative Office of the U. S. Courts, The Judiciary FY 2018 Congressional Budget Summary Revised 9–10 (June 2017). Administering this operation requires many “decisions” in the ordinary sense of the term—decisions about such things as facilities, personnel, equipment, supplies, and rules of procedure. In re Application for Exemption from Electronic Pub. Access Fees by Jennifer Gollan and Shane Shifflett, 728 F. 3d 1033, 1037 (CA9 2013). It would be absurd to suggest that every “final decision” on any such matter is appealable under §1291 or reviewable in this Court under §1254. See Hohn, supra; 15A C. Wright, A. Miller, & E. Cooper, Federal Practice and Procedure §3903, pp. 134–135 (2d ed. 1992). Such administrative decisions are not the kind of decisions or orders—i.e., decisions or orders made in a judicial capacity—to which the relevant jurisdictional provisions apply.
Respondent argues that the denial of petitioner’s funding request was just such an administrative decision, but the District Court’s ruling does not remotely resemble the sort of administrative decisions noted above. Petitioner’s request was made by motion in his federal habeas proceeding, which is indisputably a judicial proceeding. And as we will explain, resolution of the funding question requires the application of a legal standard—whether the funding is “reasonably necessary” for effective representation—that demands an evaluation of petitioner’s prospects of obtaining habeas relief. We have never held that a ruling like that is administrative and thus not subject to appellate review under the standard jurisdictional provisions.
Respondent claims that two factors support the conclusion that the funding decision was administrative, but her argument is unpersuasive.
Respondent first argues as follows: Judicial proceedings must be adversarial; 18 U. S. C. §3599(f ) funding adjudications are not adversarial because the statute allows requests to be decided ex parte; therefore, §3599(f ) funding adjudications are not judicial in nature. This reasoning is flawed.
It is certainly true that cases and controversies in our legal system are adversarial in nature, e.g., Bond v. United States, 564 U. S. 211, 217 (2011) ; Aetna Life Ins. Co. v. Haworth, 300 U. S. 227–241 (1937), but here, both the habeas proceeding as a whole and the adjudication of the specific issue of funding were adversarial. That the habeas proceeding was adversarial is beyond dispute. And on the funding question, petitioner and respondent plainly have adverse interests and have therefore squared off as adversaries. The motion for funding was formally noted as “opposed” on the Disrict Court’s docket. App. 341. That is not surprising: On one side, petitioner is seeking funding that he hopes will prevent his execution. On the other, respondent wants to enforce the judgment of the Texas courts and to do so without undue delay. Petitioner and respondent have vigorously litigated the funding question all the way to this Court.
In arguing that the funding dispute is nonadversarial, respondent attaches too much importance to the fact that the request was made ex parte. As we have noted, the “ex parte nature of a proceeding has not been thought to imply that an act otherwise within a judge’s lawful jurisdiction was deprived of its judicial character.” Forrester v. White, 484 U. S. 219, 227 (1988) .
In our adversary system, ex parte motions are disfavored, but they have their place. See, e.g., Hohn, supra, at 248 (application for COA); Dalia v. United States, 441 U. S. 238, 255 (1979) (application for a search warrant); 50 U. S. C. §1805(a) (application to conduct electronic surveillance for foreign intelligence); 18 U. S. C. §2518(3) (applications to intercept “wire, oral, or electronic communications”); 15 U. S. C. §1116(d)(1)(A) (application to seize certain goods and counterfeit marks involved in violations of the trademark laws); Fed. Rule Crim. Proc. 17(b) (application for witness subpoena); Fed. Rule Crim. Proc. 47(c) (generally recognizing ex parte motions and applications); Ullmann v. United States, 350 U. S. 422–424, 434 (1956) (application for an order granting a witness immunity in exchange for self-incriminating testimony); United States v. Monsanto, 491 U. S. 600–604 (1989) (motion to freeze defendant’s assets pending trial).
Respondent’s second argument is based on the vener-able principle “that Congress cannot vest review of the decisions of Article III courts in” entities other than “superior courts in the Article III hierarchy.” Plaut v. Spendthrift Farm, Inc., 514 U. S. 211–219 (1995) (citing Hayburn’s Case, 2 Dall. 409 (1792)). Respondent claims that §3599 funding decisions may be revised by the Director of the Administrative Office of the Courts and that this shows that such decisions must be administrative. This argument, however, rests on a faulty premise. Nothing in §3599 even hints that review by the Director of the Administrative Office is allowed.
Respondent’s argument rests in part on a handful of old lower court cases that appear to have accepted Administrative Office review of Criminal Justice Act of 1964 (CJA) payments that had been authorized by a District Court and approved by the chief judge of the relevant Circuit. See United States v. Aadal, 282 F. Supp. 664, 665 (SDNY 1968); United States v. Gast, 297 F. Supp. 620, 621–622 (Del. 1969); see also United States v. Hunter, 385 F. Supp. 358, 362 (DC 1974). The basis for these decisions was a provision of the CJA, 18 U. S. C. §3006A(h) (1964 ed.), stating that CJA payments “shall be made under the supervision of the Director of the Administrative Office of the United States Courts.” 4
It is not clear whether these decisions correctly interpreted the CJA, 5 but in any event, no similar language appears in §3599. And respondent has not identified a single instance in which the Director of the Administrative Office or any other nonjudicial officer has attempted to review or alter a §3599 decision.
Moreover, attorneys’ requests for CJA funds are markedly different from the funding application at issue here. Attorneys appointed under the CJA typically submit those requests after the conclusion of the case, and the prosecution has no stake in the resolution of the matter. The judgment in the criminal case cannot be affected by a decision on compensation for services that have been completed, and any funds awarded come out of the budget of the Judiciary, not the Executive. See 18 U. S. C. §3006A(i) (2012 ed.). Thus, the adversaries in the criminal case are not pitted against each other. In this case,on the other hand, as we have explained, petitionerand respondent have strong adverse interests. Forthese reasons, we reject respondent’s argument that the adjudication of the funding issue is nonadversarial and administrative.
Respondent, however, claims that the funding decision is administrative for an additional reason. “A §3599(f ) funding determination is properly deemed administrative,” she contends, “because it . . . may be revised outside the traditional Article III judicial hierarchy.” Brief for Respondent 23. The basis for this argument is a provision of §3599 stating that funding in excess of the generally applicable statutory cap of $7,500 must be approved by the chief judge of the circuit or another designated circuit judge. §3599(g)(2). If a funding decision is judicial and not administrative, respondent suggests, it could not be reviewed by a single circuit judge as opposed to a panel of three.
This argument confuses what is familiar with what is constitutionally required. Nothing in the Constitution ties Congress to the typical structure of appellate review established by statute. If Congress wishes to make certain rulings reviewable by a single circuit judge, rather than a panel of three, the Constitution does not stand in the way.
Satisfied that we have jurisdiction, we turn to the question whether the Court of Appeals applied the correct legal standard when it affirmed the denial of petitioner’s funding request.
Section 3599(a) authorizes federal courts to provide funding to a party who is facing the prospect of a death sentence and is “financially unable to obtain adequate representation or investigative, expert, or other reason-ably necessary services.” The statute applies to defendants in federal cases, §3599(a)(1), as well as to state and federal prisoners seeking collateral relief in federal court, §3599(a)(2).
Here we are concerned not with legal representation but with services provided by experts, investigators, and the like. Such services must be “reasonably necessary for the representation of the [applicant]” in order to be eligible for funding. §3599(f ). If the statutory standard is met, a court “may authorize the [applicant’s] attorneys to obtain such services on [his] behalf.” Ibid.
The Fifth Circuit has held that individuals seeking funding for such services must show that they have a “substantial need” for the services. 817 F. 3d, at 896; Allen v. Stephens, 805 F. 3d 617, 626 (2015); Ward v. Stephens, 777 F. 3d 250, 266, cert. denied, 577 U. S. ___ (2015). Petitioner contends that this interpretation is more demanding than the standard—“reasonably necessary”—set out in the statute. And although the difference between the two formulations may not be great, petitioner has a point.
In the strictest sense of the term, something is “necessary” only if it is essential. See Webster’s Third New International Dictionary 1510 (1993) (something is necessary if it “must be by reason of the nature of things,” if it “cannot be otherwise by reason of inherent qualities”); 10 Oxford English Dictionary 275–276 (2d ed. 1989) (OED) (defining the adjective “necessary” to mean “essential”). But in ordinary speech, the term is often used more loosely to refer to something that is merely important or strongly desired. (“I need a vacation.” “I need to catch up with an old friend.”) The term is sometimes used in a similar way in the law. The term “necessary” in the Necessary and Proper Clause does not mean “absolutely necessary,” McCulloch v. Maryland, 4 Wheat. 316, 414–415 (1819), and a “necessary” business expense under the Internal Revenue Code, 26 U. S. C. §162(a), may be an expense that is merely helpful and appropriate, Commissioner v. Tellier, 383 U. S. 687, 689 (1966) . As Black’s Law Dictionary puts it, the term “may import absolute physical necessity or inevitability, or it may import that which is only convenient, useful, appropriate, suitable, proper, or conducive to the end sought.” Black’s Law Dictionary 928 (5th ed. 1979) (Black’s).
Section 3599 appears to use the term “necessary” to mean something less than essential. The provision applies to services that are “reasonably necessary,” but it makes little sense to refer to something as being “reasonably essential.” What the statutory phrase calls for, we conclude, is a determination by the district court, in the exercise of its discretion, as to whether a reasonable attorney would regard the services as sufficiently important, guided by the considerations we set out more fully below.
The Fifth Circuit’s test—“substantial need”—is arguably more demanding. We may assume that the term “need” is comparable to “necessary”—that is, that something is “needed” if it is “necessary.” But the term “substantial” suggests a heavier burden than the statutory term “reasonably.” Compare 13 OED 291 (defining “reasonably” to mean, among other things, “[s]ufficiently, suitably, fairly”; “[f ]airly or pretty well”) with 17 id., at 66–67 (defining “substantial,” with respect to “reasons, causes, evidence,” to mean “firmly or solidly established”); see also Black’s 1456 (10th ed. 2014) (defining “reasonable” to mean “[f ]air, proper, or moderate under the circumstances . . . See plausible”); id., at 1656 (defining “substantial” to mean, among other things, “[i]mportant, essential, and material”).
The difference between “reasonably necessary” and “substantially need[ed]” may be small, but the Fifth Circuit exacerbated the problem by invoking precedent to the effect that a habeas petitioner seeking funding must present “a viable constitutional claim that is not procedurally barred.” 817 F. 3d, at 895 (internal quotation marks omitted). See also, e.g., Riley v. Dretke, 362 F. 3d 302, 307 (CA5 2004) (“A petitioner cannot show a substantial need when his claim is procedurally barred from review”); Allen, supra, at 638–639 (describing “ ‘our rule that a prisoner cannot show a substantial need for funds when his claim is procedurally barred from review’ ” (quoting Crutsinger v. Stephens, 576 Fed. Appx. 422, 431 (CA5 2014) (per curiam)); Ward, supra, at 266 (“The denial of funding will be upheld . . . when the constitutional claim is procedurally barred”).
The Fifth Circuit adopted this rule before our decision in Trevino, but after Trevino, the rule is too restrictive. Trevino permits a Texas prisoner to overcome the failure to raise a substantial ineffective-assistance claim in state court by showing that state habeas counsel was ineffective, 569 U. S., at 429, and it is possible that investigation might enable a petitioner to carry that burden. In those cases in which funding stands a credible chance of enabling a habeas petitioner to overcome the obstacle of procedural default, it may be error for a district court to refuse funding.
Congress has made it clear, however, that district courts have broad discretion in assessing requests for funding. Section 3599’s predecessor declared that district courts “shall authorize” funding for services deemed “reasonably necessary.” 21 U. S. C. §848(q)(9) (1988 ed.). Applying this provision, courts of appeals reviewed district court funding decisions for abuse of discretion. E.g., Bonin v. Calderon, 59 F. 3d 815, 837 (CA9 1995); In re Lindsey, 875 F. 2d 1502, 1507, n. 4 (CA11 1989); United States v. Alden, 767 F. 2d 314, 319 (CA7 1984). Then, as part of the Antiterrorism and Effective Death Penalty Act of 1996, 110Stat. 1226, Congress changed the verb from “shall” to “may,” and thus made it perfectly clear that determining whether funding is “reasonably necessary” is a decision as to which district courts enjoy broad discretion. See Kingdomware Technologies, Inc. v. United States, 579 U. S. ___, ___ (2016) (slip op., at 9).
A natural consideration informing the exercise of that discretion is the likelihood that the contemplated services will help the applicant win relief. After all, the proposed services must be “reasonably necessary” for the applicant’s representation, and it would not be reasonable—in fact, it would be quite unreasonable—to think that services are necessary to the applicant’s representation if, realistically speaking, they stand little hope of helping him win relief. Proper application of the “reasonably necessary” standard thus requires courts to consider the potential merit of the claims that the applicant wants to pursue, the likelihood that the services will generate useful and admissible evidence, and the prospect that the applicant will be able to clear any procedural hurdles standing in the way.
To be clear, a funding applicant must not be expected to prove that he will be able to win relief if given the services he seeks. But the “reasonably necessary” test requires an assessment of the likely utility of the services requested, and §3599(f ) cannot be read to guarantee that an applicant will have enough money to turn over every stone.
Petitioner does not deny this. He agrees that an applicant must “articulat[e] specific reasons why the services are warranted”—which includes demonstrating that the underlying claim is at least “ ‘plausible’ ”—and he acknowledges that there may even be cases in which it would be within a court’s discretion to “deny funds after a finding of ‘reasonable necessity.’ ” Brief for Petitioner 43.
These interpretive principles are consistent with the way in which §3599’s predecessors were read by the lower courts. See, e.g., Alden, supra, at 318–319 (explaining that it was “appropriate for the district court to satisfy itself that [the] defendant may have a plausible defense before granting the defendant’s . . . motion for psychiatric assistance to aid in that defense,” and that it is not proper to use the funding statute to subsidize a “ ‘fishing expedition’ ”); United States v. Hamlet, 480 F. 2d 556, 557 (CA5 1973) (per curiam) (upholding District Court’s refusal to fund psychiatric services based on the District Court’s conclusion that “the request for psychiatric services was . . . lacking in merit” because there was “no serious possibility that appellant was legally insane at any time pertinent to the crimes committed”). This abundance of precedent shows courts have plenty of experience making the determinations that §3599(f) contemplates.
Perhaps anticipating that we might not accept the Fifth Circuit’s reading of §3599(f), respondent devotes a substantial portion of her brief to an alternative ground for affirmance that was neither presented nor passed on below.
Respondent contends that whatever “reasonably necessary” means, funding is never “reasonably necessary” in a case like this one, where a habeas petitioner seeks to present a procedurally defaulted ineffective-assistance-of-trial-counsel claim that depends on facts outside the state-court record. Citing 28 U. S. C. §2254(e)(2), respondent contends that the fruits of any such investigation would be inadmissible in a federal habeas court.
We decline to decide in the first instance whether respondent’s reading of §2254(e)(2) is correct. Petitioner agrees that the argument remains open for the Fifth Circuit to consider on remand. Tr. of Oral Arg. 6.
* * *
We conclude that the Fifth Circuit’s interpretation of §3599(f ) is not a permissible reading of the statute. We therefore vacate the judgment below and remand the case for further proceedings consistent with this opinion.
It is so ordered.
1 We also consider a jurisdictional issue not raised by the parties, namely, whether we have jurisdiction even though no COA has yet been issued. We do not have jurisdiction if jurisdiction was lacking in the Court of Appeals, and the jurisdiction of a court of appeals to entertain an appeal from a final order in a habeas proceeding is dependent on the issuance of a COA. See 28 U. S. C. §2253(c)(l); Gonzalez v. Thaler, 565 U. S. 134, 142 (2012) .In this case, petitioner appealed an order of the District Court that denied both his request for funding under 18 U. S. C. §3599 and his underlying habeas claims. The Court of Appeals denied a COA as to the merits of his request for habeas relief but held that a COA was not required insofar as petitioner challenged the District Court’s denial of funding under §3599. The Fifth Circuit relied on Harbison v. Bell, 556 U. S. 180 (2009) , in which a prisoner appealed from an order that denied counsel under §3599 for a state clemency proceeding but that did not address the merits of any habeas petition. This Court held that a COA was not required. Here, petitioner took his appeal from the final order in his habeas proceeding.The parties have not briefed whether that difference between Harbison and the present case is relevant or whether an appeal from a denial of a §3599 request for funding would fit within the COA framework, and we find it unnecessary to resolve the issue. Though we take no view on the merits, we will assume for the sake of argument that the Court of Appeals could not entertain petitioner’s §3599 claim without the issuance of a COA.We may review the denial of a COA by the lower courts. See, e.g., Miller-El v. Cockrell, 537 U. S. 322–327 (2003). When the lower courts deny a COA and we conclude that their reason for doing so was flawed, we may reverse and remand so that the correct legal standard may be applied. See Slack v. McDaniel, 529 U. S. 473–486, 489–490 (2000). We take that course here. As we will explain, the correctness of the rule applied by the District Court in denying the funding request was not only debatable; it was erroneous.
2 In relevant part §1291 declares that “[t]he courts of appeals . . . shall have jurisdiction of appeals from all final decisions of the district courts of the United States, the United States District Court for the District of the Canal Zone, the District Court of Guam, and the District Court of the Virgin Islands, except where a direct review may be had in the Supreme Court.”Similarly, §2253 provides, as relevant, that “[i]n a habeas corpus proceeding or a proceeding under section 2255 before a district judge, the final order shall be subject to review, on appeal, by the court of appeals for the circuit in which the proceeding is held.” §2253(a).
3 “Cases in the courts of appeals may be reviewed by the Supreme Court by . . . writ of certiorari granted upon the petition of any party to any civil or criminal case, before or after rendition of judgment or decree.” §1254(1).
5 As far as we are aware, neither the Administrative Office nor any other nonjudicial entity currently claims the power to revise or reject a CJA compensation order issued by a court. Nothing in the CJA Guidelines suggests such a policy. See generally 7A Guide to Judiciary Policy (May 17, 2017).
SUPREME COURT OF THE UNITED STATES
CARLOS MANUEL AYESTAS, aka DENNIS ZELAYA COREA, PETITIONER v. LORIE DAVIS, DIRECTOR, TEXAS DEPARTMENT OF CRIMINAL JUSTICE, CORRECTIONAL INSTITUTIONS DIVISION
on writ of certiorari to the united states court of appeals for the fifth circuit
Justice Sotomayor, with whom Justice Ginsburg joins, concurring.
The Court correctly concludes that the Fifth Circuit applied the wrong legal standard in evaluating a request for funding for investigative services under 18 U. S. C. §3599(f). That should come as no surprise, as the Fifth Circuit required capital habeas petitioners to show a “ ‘substantial need’ ” for services, when the statute requires only a showing that the services are “ ‘reasonably necessary.’ ” Ante, at 16. “Substantial,” of course, imposes a higher burden than “reasonable.” Ante, at 16. The Fifth Circuit “exacerbated the problem” by requiring a showing of “a viable constitutional claim that is not procedurally barred,” which ignores “that investigation might enable a petitioner . . . to overcome the obstacle of procedural default.” Ante, at 16–17 (internal quotation marks omitted). I therefore join the opinion of the Court in full holding that to satisfy §3599(f), a petitioner need only show that “a reasonable attorney would regard the services as sufficiently important.” Ante, at 16.
Having answered the question presented of what is the appropriate §3599(f) standard, the Court remands Ayestas’ case for the lower courts to consider the application of the standard in the first instance. Ante, at 19. 1 I write separately to explain why, on the record before this Court, there should be little doubt that Ayestas has satisfied §3599(f).
At the center of the §3599(f) funding request in this case is Ayestas’ claim that his trial counsel was ineffective for failing to investigate mitigation. Specifically, Ayestas claims that his trial counsel was deficient in failing to conduct an investigation of his mental health and substance abuse, which could have been presented at the penalty phase of the trial to convince the jury to spare his life. As the Court notes, however, Ayestas faces a hurdle in presenting this ineffective-assistance-of-trial-counsel claim in his federal habeas petition, as his state postconviction counsel never presented that claim in the Texas collateral proceedings. See ante, at 5.
To overcome that procedural default, Ayestas relies on Martinez v. Ryan, 566 U. S. 1 (2012) , and Trevino v. Thaler, 569 U. S. 413 (2013) . In those cases, this Court recognized a “particular concern” in the application of a procedural default rule that would prevent a petitioner from “present[ing] a claim of trial error,” especially “when the claim is one of ineffective assistance of counsel.” Martinez, 566 U. S., at 12. “The right to the effective assistance of counsel,” the Court reasoned, “is a bedrock principle in our justice system.” Ibid. The Court thus held that where the “state procedural framework, by reason of its design and operation, makes it highly unlikely in a typical case that a defendant will have a meaningful opportunity to raise a claim of ineffective assistance of trial counsel on direct appeal,” then “ ‘a procedural default will not bar a federal habeas court from hearing a substantial claim of ineffective assistance at trial if, in the initial-review collateral proceeding, there was no counsel or counsel . . . was ineffective.’ ” Trevino, 569 U. S., at 429 (quoting Martinez, 566 U. S., at 17; alteration omitted). 2
Therefore, the fact that Ayestas’ postconviction counsel failed to raise his ineffective-assistance-of-trial-counsel claim in state court does not bar federal review of that claim if Ayestas can show that the “attorney in his first collateral proceeding was ineffective” and that “his claim of ineffective assistance of trial counsel is substantial.” Id., at 18. The substantiality of the ineffective-assistance-of-trial-counsel claim and the ineffectiveness of postconviction counsel are both analyzed under the familiar framework set out in Strickland v. Washington, 466 U. S. 668 (1984) . “Ineffective assistance under Strickland is deficient performance by counsel resulting in prejudice, with performance being measured against an objective standard of reasonableness.” Rompilla v. Beard, 545 U. S. 374, 380 (2005) (citation and internal quotation marks omitted).
Remember, however, the specific context in which ineffective assistance is being considered in Ayestas’ case: a request under §3599(f) for investigative services, which requires a showing only that “a reasonable attorney would regard the services as sufficiently important.” Ante, at 16. Ayestas is not “expected to prove that he will be able to win relief if given the services he seeks.” Ante, at 18 (emphasis in original). A court simply must consider at this stage “the potential merit of the claims that the applicant wants to pursue, the likelihood that the services will generate useful and admissible evidence, and the prospect that the applicant will be able to clear any procedural hurdles standing in the way.” Ante, at 17–18. Thus, the inquiry is not whether Ayestas can prove that his trial counsel was ineffective under Strickland or whether he will succeed in overcoming the procedural default under Martinez and Trevino. Rather, at this §3599(f) request stage, the focus is on the potential merit of these claims.
With this framework in mind, the focus first is on the evidence of the deficient performance of Ayestas’ state-appointed counsel. 3 Trial counsel secured the appointment of an investigator, who met with Ayestas shortly after the appointment. For nearly 15 months, however, there was apparently no investigation into Ayestas’ history in preparation for trial. Counsel instructed the investi-gator “to resume investigation” only about a month before jury selection. Record 878. The investigator then subpoenaed psychological and disciplinary prison records and had Ayestas fill out a questionnaire, in response to which Ayestas revealed that he had experienced multiple head traumas and had a history of substance abuse. Jail records also noted a rules infraction for possession of home-made intoxicants. Trial counsel never followed up on any of this information, sought further related records, or had Ayestas evaluated by a mental health professional.
About two weeks before jury selection, trial counsel for the first time reached out to Ayestas’ family in Honduras. Shortly thereafter, five days before trial, counsel wrote Ayestas’ family stating that she needed them to come testify. Ayestas’ family agreed, but they indicated that they could not obtain visas because a letter that trial counsel was supposed to have sent to the U. S. Embassy to facilitate their travel never arrived, and ultimately no family members appeared at Ayestas’ trial.
The guilt phase lasted two days, and trial counsel presented no witnesses. The penalty phase lasted less than a day, and trial counsel presented two minutes of mitigation evidence consisting of three letters from an instructor who taught English classes to Ayestas in prison, attesting that he was “a serious and attentive student.” App. 41–43. 4
On this record, Ayestas has made a strong showing that trial counsel was deficient. “It is unquestioned that under the prevailing professional norms at the time of [Ayestas’] trial, counsel had an obligation to conduct a thorough investigation of [his] background.” Porter v. McCollum, 558 U. S. 30, 39 (2009) (per curiam) (internal quotation marks omitted). Here, Ayestas’ trial counsel “clearly did not satisfy those norms.” Ibid. With a client facing a possible death sentence, counsel and her investigator did not start looking into Ayestas’ personal history until the eve of trial. The little the investigator uncovered—head trauma and a history of substance abuse—should have prompted further inquiry. Yet trial counsel did nothing. Even if Ayestas prohibited counsel from contacting his family in Honduras until the start of trial was imminent, see ante, at 3, 5 that still would not explain why counsel failed to perform any other mitigation investigation, see Porter, 558 U. S., at 40 (noting that even if the defendant is “uncooperative, . . . that does not obviate the need for defense counsel to conduct some sort of mitigation investigation (emphasis in original)). In the end, the decision to sentence Ayestas to death was made in less than one day, and his counsel spent less than two minutes presenting mitigation to the jury. Two minutes.
This Court has recognized that the decision not to present mitigation may be supported in certain cases by “strategic judgments,” provided the reviewing court is satisfied with “the adequacy of the investigations supporting those judgments.” Wiggins v. Smith, 539 U. S. 510, 521 (2003) . But this does not appear to be one of those cases. There is nothing in the record that would support the conclusion that counsel chose the two-minutes-of-mitigation strategy after careful investigation and consideration of Ayestas’ case. Instead, counsel for the most part “did not even take the first step of interviewing witnesses or requesting records” and “ignored pertinent avenues for investigation of which [they] should have been aware.” Porter, 558 U. S., at 39–40.
In evaluating the potential merit of Ayestas’ claim, the Fifth Circuit misapplied Strickland and the §3599(f) standard. It reasoned that Ayestas had not presented a viable claim that trial counsel was deficient in failing to investigate Ayestas’ mental illness because, as he was not diagnosed with schizophrenia until his time in prison, there was nothing that flagged mental illness issues prior to trial. 6 See Ayestas v. Stephens, 817 F. 3d 888, 895–897 (2016) (per curiam). The absence of a documented diagnosis, however, did not excuse trial counsel from their “obligation to conduct a thorough investigation of [Ayestas’] background.” Porter, 558 U. S., at 39 (internal quotation marks omitted). In fact, the obligation to investigate exists in part precisely because it is all too common for individuals to go years battling an undiagnosed and untreated mental illness.
In any event, the Fifth Circuit failed to consider that one of the purposes of the §3599(f) investigation was to look at Ayestas’ life around the time of the crime and trial to determine if there were mitigating circumstances that trial counsel could have discovered, such as whether symptoms of his schizophrenia had begun to manifest even before his diagnosis. The Court makes clear today that in evaluating §3599(f) funding requests, courts must consider “the likelihood that the services will generate useful and admissible evidence.” Ante, at 17. It was error, therefore, for the Fifth Circuit to evaluate the merit of the ineffective-assistance-of-trial-counsel claim and to deny §3599(f) funding based solely on an evaluation of the evidence in the record at the time of the request, without evaluating the potential evidence that Ayestas sought. Ante, at17–18.
The evidence concerning the deficiency of Ayestas’ state postconviction counsel is similarly strong. State postconviction counsel retained the services of a mitigation specialist, who prepared an investigation plan noting that it was “obvious no social history investigation was conducted” and that the jury had “heard nothing about [Ayestas’] . . . mental health, possible mental illness, [or] substance abuse history.” App. 81, 266. The plan also noted that it was “clear that [Ayestas] had a history of substance abuse.” Record 721; see also App. 267. The specialist recommended a comprehensive investigation into Ayestas’ biological, psychological, and social history to explore, inter alia, issues related to addiction and mental health.
State postconviction counsel failed to follow these recommendations. He did nothing to investigate issues related to Ayestas’ mental health or substance abuse. Notably, Ayestas suffered a psychotic episode and was diagnosed with schizophrenia while his state postconviction application was pending. Moreover, in 2003, a counsel-arranged evaluation pursuant Atkins v. Virginia, 536 U. S. 304 (2002) , noted concerns about Ayestas’ “delusional thinking.” App. 139–140. These events still did not prompt counsel to investigate Ayestas’ mental health history.
Instead, state postconviction counsel explored the circumstances of Ayestas’ arrest, conducted some juror interviews, and interviewed Ayestas’ mother and sisters, obtaining affidavits regarding Ayestas’ upbringing in Honduras and their interactions with trial counsel. Postconviction counsel eventually filed an application that contained a narrow claim of ineffective assistance of trial counsel with respect to mitigation regarding the attorneys’ failure to secure the attendance of Ayestas’ family members at trial. The Texas Court of Criminal Appeals denied the application, relying on the affidavit submitted by trial counsel, see n. 4, supra, to find no ineffectiveness in failing to get Ayestas’ family to attend trial.
The Fifth Circuit concluded that Ayestas’ state postconviction counsel was not ineffective because, in its view, Ayestas had not established any deficiency at trial in the failure to investigate mental health and substance abuse mitigation. See 817 F. 3d, at 898. That conclusion, as noted in Part II–A, supra, was based on a misapplication of Strickland and the §3599(f) standard, and thus cannot support a finding that the failure to present the claim in postconviction proceedings was “strategic.” 817 F. 3d, at 898. Nor is there anything else in the record that would excuse that deficiency. State postconviction counsel ignored his own mitigation specialist, who alerted him to a serious failing in the trial because the jury heard virtually no mitigation and to the serious failings of trial counsel because of the failure to conduct a social history investigation of Ayestas. Even after Ayestas’ psychotic episode, schizophrenia diagnosis, and documented tendencies of “delusional thinking” during the course of the representation, state postconviction counsel did nothing. As with trial counsel, the record provides no support for any “strategic justification” to disregard completely a mitigation investigation of Ayestas’ mental health and substance abuse.
Strickland next requires consideration of prejudice. To establish prejudice, this Court has held that a “defendant must show that there is a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different,” meaning “a probability sufficient to undermine confidence in the outcome.” 466 U. S., at 694. In cases alleging a failure to investigate mitigation, as here, the Court must “reweigh the evidence in aggravation against the totality of available mitigating evidence.” Wiggins, 539 U. S., at 534.
Even with the scant evidence in the record at this time as to what Ayestas could have presented to the jury in the form of mitigation, Ayestas has made a strong showing that his claim has potential merit. That trial counsel presented only two minutes of mitigation already goes a long way to establishing prejudice. In fact, the State emphasized to the jury at sentencing:
“Does he have anything there that would lead you to conclude there is some type of mitigation, anything at all? There is no drug problem . . . no health problem . . . no alcohol problem. . . . [O]nly . . . these three pieces of paper . . . . Making steps to learn a second language does not lessen his moral blameworthiness . . . .” Record 4747.
The State, in contrast, presented evidence of Ayestas’ criminal history as well as victim impact testimony. After deliberating for only 25 minutes, the jury assessed a punishment of death against Ayestas, finding that he was a future danger, that he intended to cause death or anticipated the loss of life, and that there were no mitigating circumstances that warranted imposition of a life sentence over a death sentence. Had just one juror dissented on a single one of these findings, no death sentence could have been imposed. See Tex. Code Crim. Proc. Ann., Art. 37.071, §2(g) (Vernon Cum. Supp. 2017); see also ante, at 2. With even minimal investigation by trial counsel, at least one may well have, as this Court has held that evidence of mental illness and substance abuse is relevant to assessing moral culpability. See Rompilla, 545 U. S., at 393; Porter, 558 U. S., at 43–44. Instead, the jury “heard almost nothing that would humanize [him] or allow them to accurately gauge his moral culpability.” Id., at 41. There is thus good reason to believe that, were Ayestas’ §3599(f) motion granted, he could establish prejudice under Strickland.
The Fifth Circuit held otherwise based on its belief that no amount of mitigation would have changed the outcome of the sentencing given the “brutality of the crime.” 817 F. 3d, at 898. That “brutality of the crime” rationale is simply contrary to our directive in case after case that, in assessing prejudice, a court must “consider the totality of the available mitigation evidence . . . and reweigh it against the evidence in aggravation.” Porter, 558 U. S., at 41 (internal quotation marks and alterations omitted); see also Williams v. Taylor, 529 U. S. 362–398 (2000); Wiggins, 539 U. S., at 534. By considering aggravation in isolation, the Fifth Circuit directly contravened this fundamental principle. 7
In sum, Ayestas has made a strong showing that he is entitled to §3599(f) funding. As the Court notes, the statute affords district courts some discretion in these funding determinations, even where a petitioner shows the services are “ ‘reasonably necessary.’ ” Ante, at 17–18. Exercise of that discretion may be appropriate if there is a showing of gamesmanship or where the State has provided funding for the same investigation services, as Ayestas conceded at argument. See Tr. of Oral Arg. 13. Nonetheless, the troubling failures of counsel at both the trial and state postconviction stages of Ayestas’ case are exactly the types of facts that should prompt courts to afford investigatory services to ensure that trial errors that go to a “bedrock principle in our justice system” do not go unaddressed. Martinez, 566 U. S., at 12.
1 The Court also declines to consider arguments that respondent advanced that were neither presented nor passed on below. Ante, at 19.
2 The reason for this exception is evident. Excusing the procedural default “acknowledges, as an equitable matter, that the initial-review collateral proceeding, if undertaken without counsel or with ineffective counsel, may not have been sufficient.” Martinez, 566 U. S., at 14. “Claims of ineffective assistance at trial often require investigative work and an understanding of trial strategy,” and “the prisoner is in no position to develop the evidentiary basis for a claim of ineffective assistance, which often turns on evidence outside the trial record.” Id., at 11–12; see also Trevino, 569 U. S., at 423–424, 428.
3 The State appointed two attorneys to represent Ayestas at trial. I refer to them together as “trial counsel.”
4 Trial counsel also attempted to introduce evidence that Ayestas had no criminal history in Honduras, but failed to link Ayestas to the records, which were under his given name, “Dennis Zelaya Corea.” See Ayestas v. Stephens, 817 F. 3d 888, 892, n. 1 (CA5 2016) (per curiam).
5 During postconviction proceedings, trial counsel filed an affidavit asserting that Ayestas did not allow contact with his family in Honduras until after jury selection had commenced. When the record evidence contradicted that assertion, counsel submitted another affidavit with a revised timeline. Ayestas disputes having instructed trial counsel not to contact his family in Honduras.
6 It is unclear whether the Fifth Circuit ultimately relied on its determination that trial counsel was not deficient in rejecting Ayestas’ claims. In its panel opinion, it incorrectly stated that trial counsel had conducted a psychological evaluation of Ayestas. 817 F. 3d, at 897. After Ayestas corrected the record in his petition for rehearing, the panel issued an order reaffirming its holding, relying on its finding of no prejudice. See Ayestas v. Stephens, 826 F. 3d 214, 215 (2016) ( per curiam). Still, the Fifth Circuit never disavowed its conclusion regarding trial-counsel deficiency. Ibid.
7 Notably, application of this “brutality of the crime” rule is particularly irrational in the §3599(f) context, where the court is unaware of what the undiscovered evidence of mitigation looks like.