542 U.S. 74
SCOTUS
2004

Lead Opinion

Justice Souter

delivered the opinion of the Court.

Respondent claims the right to withdraw his plea of guilty as a consequence of the District Court’s failure to give one of the warnings required by Federal Rule of Criminal Procedure-11. Because the claim of Rule 11 error was not preserved by timely objection, the plain-error standard of Rule 52(b) applies, with its requirement to prove effect on substantial rights. The question is what showing must thus be made to obtain relief for an unpreserved Rule 11 failing, and we hold that a defendant is obliged to show a reasonable probability that, but for the error, he would not have entered the plea.

I

In early May 1999, a confidential informant working with law enforcement arranged through respondent Carlos Dominguez Benitez (hereinafter Dominguez) to buy several pounds of methamphetamine. First, the informant got a sample from Dominguez, and a week later Dominguez went to a restaurant in Anaheim, California, to consummate the sale in the company of two confederates, one of whom brought a shopping bag with over a kilogram of the drugs. The meeting ended when the informant gave a signal and officers arrested the dealers. Dominguez confessed to selling the methamphetamine and gave information about his supplier and confederates.

*77A federal grand jury indicted Dominguez on two counts: conspiracy to possess more than 500 grams of methamphetamine, and possession of 1,391 grams of a methamphetamine mixture, both with intent to distribute. On the conspiracy count, Dominguez faced a statutory, mandatory minimum sentence of 10 years, with a maximum of life. 84 Stat. 1260, 21 U. S. C. §§ 841(b)(1)(A), 846. The District Court appointed counsel, who began talking with the Government about a plea agreement.

In September 1999, the District Court received the first of several letters from Dominguez,1 in which he asked for a new lawyer and expressed discomfort with the plea agreement his counsel was encouraging him to sign. On counsel’s motion, the court held a status conference, at which Dominguez spoke to the judge. Again he said he was dissatisfied with his representation, and wanted a “better deal.” The court asked whether he was “talking about a disposition ... other than trial,” and Dominguez answered, “At no time have I decided to go to any trial.” App. 46-47. Counsel spoke to the same effect later in the proceeding, when he said that he had “told [the prosecutor] all along that there won’t be a trial on the [date set] based on my client’s representations that he doesn’t want a trial.” Id., at 51. The court explained to Dominguez that it could not help him in plea negotiations, and found no reason to change counsel.

Shortly after that, the parties agreed that Dominguez would plead guilty to the conspiracy, and the Government would dismiss the possession charge. The Government stipulated that Dominguez would receive what is known as a safety-valve reduction of two levels. See United States Sentencing Commission, Guidelines Manual §§2Dl.l(b)(6), *785C1.2 (Nov. 1999) (hereinafter USSG).2 The safety valve was important because it would allow the court to invoke 18 U. S. C. § 3553(f), authorizing a sentence below the otherwise mandatory minimum in certain cases of diminished culpability, the only chance Dominguez had for a sentence under 10 years. That chance turned on satisfying five conditions, one going to Dominguez’s criminal history, which the agreement did not address. The agreement did, however, warn Dominguez that it did not bind the sentencing court, and that Dominguez could not withdraw his plea if the court did not accept the Government’s stipulations or recommendations. At a hearing the next day, Dominguez changed his plea to guilty. In the plea colloquy, the court gave almost all the required Rule 11 warnings, including the warning that the plea agreement did not bind the court, but the judge failed to mention that Dominguez could not withdraw his plea if the court did not accept the Government’s recommendations. See Fed. Rule Crim. Proc. 11(c)(3)(B).3

When the Probation Office subsequently issued its report, it found that Dominguez had three prior convictions, two of them under other names, which neither defense counsel nor the prosecutor had known at the time of the plea negotiations. The upshot was that Dominguez was ineligible for the safety valve, and so had no chance to escape the sentence of 10 years. After receiving two more letters from Dominguez complaining about the quality of counsel’s representa*79tion, the District Court sentenced Dominguez to the mandatory minimum. At the sentencing hearing, all counsel told the court that they had thought Dominguez might at least have been eligible for the safety-valve mitigation, but agreed that with three convictions, he was not. Dominguez told the court that he had “never had any knowledge about the points of responsibility, the safety valve, or anything like that.” App. 109. The court replied that in light of the “lengthy change of plea proceedings” it was “difficult ... to accept what” Dominguez said. Id., at 112.

On appeal, Dominguez argued that the District Court’s failure to warn him, as Rule 11(c)(3)(B) instructs, that he could not withdraw his guilty plea if the court did not accept the Government’s recommendations required reversal. After waiting for United States v. Vonn, 535 U. S. 55 (2002), a divided panel of the Court of Appeals for the Ninth Circuit agreed, 310 F. 3d 1221 (2002), and cited United States v. Olano, 507 U. S. 725 (1993), in applying the plain-error standard. The court held that the District Court had indeed erred; and that the error was plain, affected Dominguez’s substantial rights, and required correction in the interests of justice.

To show that substantial rights were affected, the Court of Appeals required Dominguez to “prove that the court’s error was not minor or technical and that he did hot understand the rights at issue when he entered his guilty plea.” 310 F. 3d, at 1225.4 The court rejected the Government’s arguments that the written plea agreement or the District Court’s other statements in the plea colloquy sufficiently advised Dominguez of his rights, given Dominguez’s inability to speak English and the assurances of both counsel that he would likely qualify under the safety-valve provision. Judge Tallman dissented, with the warning that the majori*80ty’s analysis followed neither Vonn nor Circuit precedent. 310 F. 3d, at 1227-1228.

We granted certiorari, 540 U. S. 1072 (2003), on the question “[wjhether, in order to show that a violation of Federal Rule of Criminal Procedure 11 constitutes reversible plain error, a defendant must demonstrate that he would not have pleaded guilty if the violation had not occurred.” Pet. for Cert. (I). We now reverse.

II

A

Because the Government agreed to make a nonbinding sentencing recommendation, Rule 11(c)(3)(B) required the court to “advise the defendant that the defendant has no right to withdraw the plea if the court does not follow the recommendation or request.” Rule 11, however, instructs that not every violation of its terms calls for reversal of conviction by entitling the defendant to withdraw his guilty plea. “A variance from the requirements of this rule is harmless error if it does not affect substantial rights.” Fed. Rule Crim. Proc. 11(h).5

In Vonn, we considered the standard that applies when a defendant is dilatory in raising Rule 11 error, and held that reversal is not in order unless the error is plain. 535 U. S., at 63; see Olano, supra, at 731-737. Although we explained that in assessing the effect of Rule 11 error, a reviewing court must look to the entire record, not to the plea proceedings alone, Vonn, supra, at 74-75, we did not formulate the standard for determining whether a defendant has shown, as the plain-error standard requires, Olano, supra, at 734-735, an effect on his substantial rights.

*81B

It is only for certain structural errors undermining the fairness of a criminal proceeding as a whole that even preserved error requires reversal without regard to the mistake’s effect on the proceeding. See Arizona v. Fulmi-nante, 499 U. S. 279, 309-310 (1991) (giving examples). Dominguez does not argue that either Rule 11 error generally or the Rule 11 error here is structural in this sense.6

Otherwise, relief for error is tied in some way to prejudicial effect, and the standard phrased as “error that affects substantial rights,” used in Rule 52, has previously been taken to mean error with a prejudicial effect on the outcome of a judicial proceeding. See Kotteakos v. United States, 328 U. S. 750 (1946). To affect “substantial rights,” see 28 U. S. C. §2111, an error must have “substantial and injurious effect or influence in determining the . . . verdict.” Kot-teakos, supra, at 776.7 In cases where the burden of demonstrating prejudice (or materiality) is on the defendant seeking relief, we have invoked a standard with similarities to the Kotteakos formulation in requiring the showing of *82“a reasonable probability that, but for [the error claimed], the result of the proceeding would have been different.” United States v. Bagley, 473 U. S. 667, 682 (1985) (opinion of Blackmun, J.) (adopting the prejudice standard of Strickland v. Washington, 466 U. S. 668, 694 (1984), for claims under Brady v. Maryland, 373 U. S. 83 (1963) (internal quotation marks omitted)); 473 U. S., at 685 (White, J., concurring in part and concurring in judgment) (same).8

No reason has appeared for treating the phrase “affecting substantial rights” as untethered to a prejudice requirement when applying Olano to this nonstructural error, or for doubting that Bagley is a sensible model to follow. As Vonn makes clear, the burden of establishing entitlement to relief for plain error is on the defendant claiming it, and for several reasons, we think that burden should not be too easy for defendants in Dominguez’s position. First, the standard should enforce the policies that underpin Rule 52(b) generally, to encourage timely objections and reduce wasteful reversals by demanding strenuous exertion to get relief for unpreserved error. See Vonn, 535 U. S., at 73. Second, it should respect the particular importance of the finality of guilty pleas, which usually rest, after all, on a defendant’s *83profession of guilt in open court, and are indispensable in the operation of the modern criminal justice system. See United States v. Timmreck, 441 U. S. 780, 784 (1979). And, in this case, these reasons are complemented by the fact, worth repeating, that the violation claimed was of Rule 11, not of due process.

We hold, therefore, that a defendant who seeks reversal of his conviction after a guilty plea, on the ground that the district court committed plain error under Rule 11, must show a reasonable probability that, but for the error, he would not have entered the plea. A defendant must thus satisfy the judgment of the reviewing court, informed by the entire record, that the probability of a different result is “sufficient to undermine confidence in the outcome” of the proceeding. Strickland, supra, at 694; Bagley, supra, at 682 (opinion of Blackmun, J. (internal quotation marks omitted)).9

*84c

What we have already said points to why the test applied by the Court of Appeals in this case fell short. Its first element was whether the error was “minor or technical,” 310 F. 3d, at 1225, a phrase it took from United States v. Graibe, 946 F. 2d 1428 (CA9 1991), which in turn found it in the 1983 commentary that accompanied the amendment to Rule 11(h). 946 F. 2d, at 1433. But this element requires no examination of the effect of the omitted warning on a defendant’s decision, a failing repeated to a significant extent by the second element of the Ninth Circuit’s test, taken from United States v. Minore, 292 F. 3d 1109 (CA9 2002), which asks whether the defendant understood “the rights at issue when he entered his.guilty plea.” 310 F. 3d, at 1225. True, this enquiry gets closer than the first to a consideration of the likely effect of Rule 11 error on the defendant’s decision to plead; assessing a claim that an error affected a defendant’s decision to plead guilty must take into account any indication that the omission of a Rule 11 warning misled him. But the standard of the Court of Appeals does not allow consideration of any record evidence tending to show that a misunderstanding was inconsequential to a defendant’s decision, or evidence indicating the relative significance of other facts that may have borne on his choice regardless of any Rule 11 error.10

Relevant evidence that the Court of Appeals thus passed over in this case included Dominguez’s statement to the District Court that he did not intend to go to trial, and his eoun-*85sel’s confirmation of that representation, made at the same hearing. The neglected but relevant considerations also included the implications raised by Dominguez’s protests at the sentencing hearing. He claimed that when he pleaded guilty he had “never had any knowledge about the points of responsibility, the safety valve, or anything like that.” App. 109. These statements, if credited, would show that Dominguez was confused about the law that applied to his sentence, about which the court clearly informed him, but they do not suggest any causal link between his confusion and the particular Rule 11 violation on which he now seeks relief.

Other matters that may be relevant but escape notice under the Ninth Circuit’s test are the overall strength of the Government’s case and any possible defenses that appear from the record, subjects that courts are accustomed to considering in a Strickland or Brady analysis. When the record made for a guilty plea and sentencing reveals evidence, as this one does, showing both a controlled sale of drugs to an informant and a confession, one can fairly ask a defendant seeking to withdraw his plea what he might ever have thought he could gain by going to trial. The point of the question is not to second-guess a defendant’s actual decision; if it is reasonably probable he would have gone to trial absent the error, it is no matter that the choice may have been foolish. The point, rather, is to enquire whether the omitted warning would have made the difference required by the standard of reasonable probability; it is hard to see here how the warning could have had an effect on Dominguez’s assessment of his strategic position. And even if there were reason to think the warning from the bench could have mattered, there was the plea agreement, read to Dominguez in his native Spanish, which specifically warned that he could not withdraw his plea if the court refused to accept the Government’s recommendations. This fact, uncontested by Dominguez, tends to show that the Rule 11 error made no difference to the outcome here.

*86* * *

We reverse the judgment of the Court of Appeals and remand the case for further proceedings consistent with this opinion.

It is so ordered.

Notes

Dominguez speaks and writes Spanish, not English. A certified translator was present for the hearings in court we describe, and for the plea agreement. Some of the letters are in English, and the record does not show who translated them or assisted Dominguez in writing them.

The agreement also contemplated that Dominguez’s total offense level under the Guidelines would be 27, after considering the safety valve and a downward adjustment for acceptance of responsibility. Assuming so, and assuming he had no (or minimal) criminal history, his sentence could have been as low as 70 months. See USSG eh. 5, pt. A (sentencing table).

At the time of the plea hearing, the requirement appeared at Federal Rule of Criminal Procedure 11(e)(2). It has not changed in substance. We refer to the current Rule in the text of this opinion, and do likewise for Rules 11(h) and 52(b), each of which has also received a stylistic amendment.

Other Courts of Appeals employed different tests. See n. 8, infra.

Congress gave the courts this instruction in 1983, in partial response to this Court’s decision in McCarthy v. United States, 394 U. S. 459 (1969), which it felt had caused too many reversals for reasons that were too insubstantial. See United States v. Vonn, 535 U. S. 55, 66-71 (2002) (discussing the history of Rule 11(h)).

The argument, if made, would not prevail. The omission of a single Rule 11 warning without more is not colorably structural. Cf. United States v. Timmreck, 441 U. S. 780, 783-784 (1979) (holding that Rule 11 error without more is not cognizable on collateral review).

When the Government has the burden of addressing prejudice, as in excusing preserved error as harmless on direct review of the criminal conviction, it is not enough to negate an effect on the outcome of the case. See Chapman v. California, 386 U. S. 18, 24 (1967) (“[T]he court must be able to declare a belief that [constitutional error] was harmless beyond a reasonable doubt”). When the Government has the burden of showing that constitutional trial error is harmless because it comes up on collateral review, the heightened interest in finality generally calls for the Government to meet the more lenient Kotteakos standard. Brecht v. Abrahamson, 507 U. S. 619, 638 (1993). If the burden is on a defendant to show prejudice in the first instance, of course, it would be easier to show a reasonable doubt that constitutional error affected a trial than to show a likely effect on the outcome or verdict.

This standard is similar to one already applied by some Courts of Appeals, though those courts have not drawn a direct connection to Strickland and Bagley, and in some cases understood themselves to be reviewing for harmless, rather than plain, error. See United States v. Martinez, 289 F. 3d 1023, 1029 (CA7 2002) (on plain-error review, asking “whether any Rule 11 violations would have likely affected [the defendant’s] willingness to plead guilty”); see also United States v. Johnson, 1 F. 3d 296, 302 (CA5 1993) (en banc) (on harmless-error review, asking “whether the defendant’s knowledge and comprehension of the full and correct information would have been likely to affect his willingness to plead guilty”); cf. United States v. Olano, 507 U. S. 725, 734-735 (1993) (the main difference as to substantial rights in the harmless- and plain-error analyses is that the burden of persuasion shifts from Government to defendant).

One significant difference, however, between Rule 11 claims and claims under Strickland and Brady v. Maryland, 373 U. S. 83 (1963), is that the latter may be raised in postconviction proceedings such as a petition for habeas corpus, or a motion to vacate a sentence under 28 U. S. C. §2255. Those proceedings permit greater development of the record. See Massaro v. United States, 538 U. S. 500 (2003) (Strickland claims are not procedurally defaulted when brought for the first time on § 2255, because of the advantages of that form of proceeding for hearing, such cases). For Rule 11 claims, by contrast, that way is open only in the most egregious cases. Timmreck, supra; see also Vonn, 535 U. S., at 64 (noting that Rule 11(h) was not meant to disturb Timmreck). A defendant will rarely, if ever, be able to obtain relief for Rule 11 violations under §2255; and relief on direct appeal, given the plain-error standard that will apply in many cases, will be difficult to get, as it should be.' Cf. United States v. Raineri, 42 F. 3d 36, 45 (CA1 1994) (Boudin, J.) (“[JJust as there are many fair trials but few perfect ones, so flaws are also to be expected in Rule 11 proceedings”).

Our rule does not, however, foreclose relief altogether. The reasonable-probability standard is not the same as, and should not be confused with, a requirement that a defendant prove by a preponderance of the evidence that but for error things would have been different. See Kyles v. Whitley, 514 U. S. 419, 434 (1995).

This is another point of contrast with the constitutional question whether a defendant’s guilty plea was knowing and voluntary. We have held, for example, that when the record of a criminal conviction obtained by guilty plea contains no evidence that a defendant knew of the rights he was putatively waiving, the conviction must be reversed. Boykin v. Alabama, 395 U. S. 238, 243 (1969). We do not suggest that such a conviction could be saved even by overwhelming evidence that the defendant would have pleaded guilty regardless.



Concurrence Opinion

Justice Scalia,

concurring in the judgment.

I agree with much of the Court’s opinion and concur in its disposition of the case. I do not, however, agree with its holding that respondent need not show prejudice by a preponderance of the evidence. Ante, at 83, n. 9.

By my count, this Court has adopted no fewer than, four assertedly different standards of probability relating to the assessment of whether the outcome of trial would have been different if error had not occurred, or if omitted evidence had been included. See Chapman v. California, 386 U. S. 18, 24 (1967) (adopting “harmless beyond a reasonable doubt” standard for preserving, on direct review, conviction obtained in a trial where constitutional error occurred); Brecht v. Abrahamson, 507 U. S. 619, 637 (1993) (rejecting Chapman in favor of the less defendant-friendly “ ‘substantial and injurious effect or influence’” standard of Kotteakos v. United States, 328 U. S. 750 (1946), for overturning conviction on collateral review); United States v. Agurs, 427 U. S. 97, 111-113 (1976) (rejecting Kotteakos for overturning conviction on the basis of violations of Brady v. Maryland, 373 U. S. 83 (1963), in favor of an even less defendant-friendly standard later described in Strickland v. Washington, 466 U. S. 668, 694 (1984), as a “reasonable probability”); id., at 693-694 (distinguishing the “reasonable probability” standard from the still yet less defendant-friendly “more likely than not” standard applicable to claims of newly discovered evidence). See generally Kyles v. Whitley, 514 U. S. 419, 434-436 (1995). Such ineffable gradations of probability seem to me quite beyond the ability of the judicial mind (or any mind) to grasp, and thus harmful rather than helpful *87to the consistency and rationality of judicial decisionmaking. That is especially so when they are applied to the hypothesizing of events that never in fact occurred. Such an enterprise is not factfinding, but closer to divination.

For purposes of estimating what would have happened, it seems to me that the only serviceable standards are the traditional “beyond a reasonable doubt” and “more likely than not.” We should not pretend to a higher degree of precision. I would not, therefore, extend our “reasonable probability” standard to the plain-error context. I would hold that, where a defendant has failed to object at trial, and thus has the burden of proving that a mistake he failed to prevent had an effect on his substantial rights,' he must show that effect to be probable, that is, more likely than not.