joined by Justice Blackmun, concurring in part and dissenting in part.
I agree that the evidence was sufficient to sustain the mail fraud convictions and therefore join Part III of the Court’s *454opinion. I also agree that the Court of Appeals erred in holding that misjoinder under Rule 8 of the Federal Rules of Criminal Procedure is prejudicial per se. I write separately, however, because my reasons for reaching this conclusion differ from the Court’s, and because I agree with Justice Stevens that the harmless-error inquiry should be made in the first instance by the Court of Appeals.
I
The Act of February 26, 1919 (1919 Act), 40 Stat. 1181, amended § 269 of the Judicial Code. It provided in part:
“On the hearing of any appeal, certiorari, writ of error, or motion for a new trial, in any case, civil or criminal, the court shall give judgment after an examination of the entire record before the court, without regard to technical errors, defects, or exceptions which do not affect the substantial rights of the parties.” 28 U. S. C. §391 (1925-1926 ed.).
In 1949, this provision was reenacted in its current form as 28 U. S. C. §2111, and now instructs appellate courts to “give judgment after an examination of the record without regard to errors or defects which do not affect the substantial rights of the parties.” The 1919 Act was also incorporated in the Federal Rules of Criminal Procedure, and Rule 52(a) provides that “[a]ny error, defect, irregularity or variance which does not affect substantial rights shall be disregarded.” See also, Fed. Rule Civ. Proc. 61 (“The court at every stage of the proceeding must disregard any error or defect in the proceeding which does not affect the substantial rights of the parties”). Although § 2111 and Rule 52(a) refer to “errors or defects” without the qualifying word “technical,” this change did not alter the substantive legal test. See H. R. Rep. No. 352, 81st Cong., 1st Sess., 18 (1949) (§2111 “[ijncorpo-rates” former harmless-error statute); Advisory Committee’s *455Notes on Fed. Rule Crim. Proc. 52(a), 18 U. S. C. App., p. 657 (Rule is a “restatement of existing law”).
The 1919 Act, §2111, and Rule 52(a) all provide that an error is to be disregarded unless it “affects the substantial rights of the parties.” This litigation thus presents a straightforward question of statutory construction: what does the phrase “affects the substantial rights of the parties” mean? Respondents in No. 84-744 contend that the term “substantial rights” refers to a particular class of rights which are essential to a fair trial and argue that errors which “affect” these rights cannot be disregarded on appeal. According to respondents, the 1919 Act, as reenacted in §2111 and Rule 52(a), incorporated our holding in McElroy v. United States, 164 U. S. 76 (1896), that joinder is one of these “substantial rights,” so that misjoinder is per se reversible.
For the reasons which follow, I conclude that the question whether a particular error “affects the substantial rights of the parties” does not entail a process of classification, whereby some rights are deemed “substantial” and errors affecting these rights are automatically reversible. Rather, an error “affects substantial rights” only if it casts doubt on the outcome of the proceeding. In other words, subject to the exceptions discussed in Part II (most importantly the exception for constitutional errors), I read § 2111 and Rule 52(a) to require harmless-error inquiry for all procedural errors. As none of these exceptions is applicable to misjoinder in violation of Rule 8, I concur in the Court’s result on this issue.
Reference to whether error “affected the substantial rights of the parties” was not invented by Congress in 1919. The phrase was commonly used by courts throughout the 19th century to express the conclusion that particular claims of error did or did not warrant reversal. However, as used by these courts, error which “affected the substantial rights of the parties” was generally understood to refer, not to errors respecting a particular class of rights, but rather to any error which affected the fairness of the trial as a whole by calling *456into question the reliability of the result. See, e. g., Connors v. United States, 158 U. S. 408, 411, 414 (1895); Maish v. Arizona, 164 U. S. 599, 602 (1896); Williams v. United States, 168 U. S. 382, 390-398 (1897); American Surety Co. v. Pauly, 170 U. S. 133, 159 (1898); McCabe & Steen Constr. Co. v. Wilson, 209 U. S. 275, 279 (1908); Holmgren v. United States, 217 U. S. 509, 523-524 (1910). In other words, the statement that an error did not “affect the substantial rights of the parties” was a way of stating the conclusion that the error was not prejudicial.
A careful reading of McElroy demonstrates that it is consistent with this understanding of the phrase “affects the substantial rights of the parties.” In McElroy, five defendants were charged in two indictments with separate assaults and in a third indictment with arson. Three of the defendants were also charged in yet a fourth indictment with another assault. After explaining these charges, the Court noted that “it is the settled rule ... to confine the indictment to one distinct offence or restrict the evidence to one transaction” because “[i]n cases of felony, the multiplication of distinct charges has been considered so objectionable as tending to confound the accused in his defence, or to prejudice him as to his challenges . . . .” 164 U. S., at 80. The Court then stated: “Necessarily where the accused is deprived of a substantial right by the action of the trial court, such action, having been properly objected to, is revisable on error.” Ibid. In context, this merely restates the common-law understanding that an error is reversible if it prejudices the defendant. The Court did not state that joinder is a “substantial right” and, for this reason, any error respecting joinder is reversible. Rather, the Court held that “[i]t cannot be said in [a case of improper joinder] that all the defendants may not have been embarrassed and prejudiced in their defence, or that the attention of the jury may not have been distracted to their injury in passing upon distinct and independent transactions.” Id., at 81. In other words, the *457Court concluded that misjoinder is the kind of error which must be presumed to have prejudiced the accused and, for that reason, misjoinder affects his “substantial rights.” As discussed in Part II, the irrebuttable presumption that mis-joinder is prejudicial is inconsistent with the Court’s subsequent harmless-error jurisprudence and can be overruled. For the moment, however, it is important only to note that nothing in McElroy suggests that the requirement that error have “affect[ed] the substantial rights of the parties” refers to anything other than that the error have been prejudicial.
Absent some contrary indication, then, it would seem logical to conclude that when Congress used the phrase “affect[s] the substantial rights of the parties” in the 1919 Act, Congress meant to require an inquiry into whether an error cast doubt on the verdict, not to create a class of rights as to which error was per se reversible. The legislative history of the 1919 Act confirms that this was in fact what Congress intended.
The primary impetus for the enactment of the 1919 Act was the practice in some jurisdictions of reversing convictions on appeal for any procedural error at trial, without regard to whether the error was prejudicial. See Kotteakos v. United States, 328 U. S. 750, 758-759 (1946). There was also concern over the inconsistent application of harmless-error analysis by other courts, this Court in particular. See H. R. Rep. No. 913, 65th Cong., 3d Sess., 2 (1919) (quoting H. R. Rep. No. 611, 62d Cong., 2d Sess., 2 (1912)). The large number of reversals which resulted from failure to scrutinize errors for their prejudicial effect was criticized by leaders of the legal profession, including Taft, Pound, Wigmore, and Hadley. See Kotteakos, supra, at 758-759. After prolonged consideration, Congress responded to this criticism by passing the 1919 Act. The House Report accompanying the Act explained:
“ Tt is the purpose of the . . . bill to enact, in so far as the appellate courts are concerned, that in the consideration *458in an appellate court of a writ of error or an appeal judgment shall be rendered upon the merits without permitting reversals for technical defects in the procedure below and without presuming that any error which may appear had been of necessity prejudicial to the complaining party.”’ H. R. Rep. No. 913, supra, at 2 (quoting H. R. Rep. No. 611, supra, at 2) (emphasis added).
The theme that reversal be limited to prejudicial errors is found throughout the legislative history. For example, the Report accompanying the first version of the bill to pass the House of Representatives explained the meaning of the requirement that error be disregarded unless it “affect[s] the substantial rights of the parties” by quoting from an article by President Taft: “‘No judgment of the court below should be reversed except for an error which the court, after hearing [sic] the entire evidence, can affirmatively say would have led to a different verdict.’” H. R. Rep. No. 1949, 61st Cong., 3d Sess., 1 (1911) (quoting Taft, The Administration of Criminal Law, 15 Yale L. J. 1, 16 (1905)). The Report criticized the practice of reversing judgments for errors which “did not in the least affect the substantial rights of the parties, the real merits of the case having been properly adjudicated upon the first trial.” H. R. Rep. No. 1949, supra, at 2 (emphasis added). See also, ibid, (quoting Justice O’Gorman of the New York Supreme Court to the effect that “[o]ne of the gravest faults with our present mode of trial is the ease and frequency with which judgments are reversed on technicalities which do not affect the merits of the case, and which at no stage of the case have affected the merits”); H. R. Rep. No. 1218, 63d Cong., 3d Sess. (1914); H. R. Rep. No. 264, 64th Cong., 1st Sess. (1916).
Our decision in Kotteakos v. United States, supra, forecloses any remaining questions as to the interpretation of the phrase “affects substantial rights of the parties.” In Kotteakos, we expressly rejected the argument that the 1919 Act required a determination of “what are only technical, *459what substantial rights; and what really affects the latter hurtfully.” 328 U. S., at 761. We held instead that the Act’s command to disregard errors unless they “affect the substantial rights of the parties” was a command not to overturn a conviction unless, after examining the record as a whole, the court concludes that an error may have had “substantial influence” on the outcome of the proceeding. Id., at 765. Justice Rutledge’s explanation, which includes a description of the proper analysis to apply in evaluating the effect of procedural errors, is well worth repeating:
“It comes down on its face to a very plain admonition: ‘Do not be technical, where technicality does not really hurt the party whose rights in the trial and in its outcome the technicality affects.’. . .
“Easier was the command to make than it has been always to observe. This, in part because it is general; but in part also because the discrimination it requires is one of judgment transcending confinement by formula or precise rule. That faculty cannot ever be wholly imprisoned in words, much less upon such a criterion as what are only technical, what substantial rights; and what really affects the latter hurtfully. Judgment, the play of impression and conviction along with intelligence, varies with judges and also with circumstance. What may be technical for one is substantial for another; what minor and unimportant in one setting crucial in another.
“In the final analysis judgment in each case must be influenced by conviction resulting from examination of the proceedings in their entirety, tempered but not governed in any rigid sense of stare decisis by what has been done in similar situations. Necessarily the character of the proceeding, what is at stake upon its outcome, and the relation of the error asserted to casting the balance *460for decision on the case as a whole, are material factors in judgment.
“If, when all is said and done, the conviction is sure that the error did not influence the jury, or had but very slight effect, the verdict and the judgment should stand, except perhaps where the departure is from a constitutional norm or a specific command of Congress. But if one cannot say, with fair assurance, after pondering all that happened without stripping the erroneous action from the whole, that the judgment was not substantially swayed by the error, it is impossible to conclude that substantial rights were not affected. The inquiry cannot be merely whether there was enough to support the result, apart from the phase affected by the error. It is rather, even so, whether the error itself had substantial influence. If so, or if one is left in grave doubt, the conviction cannot stand.” Id., at 760-765 (citations and footnotes omitted).1
II
This interpretation of §2111 and Rule 52(a) as requiring examination of the prejudicial effect of all procedural errors is subject to several exceptions. First, and most importantly, constitutional errors are governed by the Due Process Clauses of the Fifth and Fourteenth Amendments rather than by §2111 and Rule 52(a). See Chapman v. California, 386 U. S. 18 (1967); United States v. Hasting, 461 U. S. 499 (1983). Thus, the test for harmless constitutional error is stricter than its statutory counterpart. Compare, Chapman, supra, at 24 (prosecution must establish that the error *461was “harmless beyond a reasonable doubt”), with Kotteakos, 328 U. S., at 765 (error is harmless unless it had “substantial influence” on the outcome or leaves one in “grave doubt” as to whether it had such effect).2 In addition, Congress may, of course, expressly provide that a particular right is excluded from the operation of the harmless-error rule. Neither of these exceptions applies to misjoinder in violation of Rule 8, however. Misjoinder does not ordinarily rise to the level of a constitutional violation,3 and nothing in the language or *462history of either the statutory harmless-error provisions or Rule 8 indicates that Congress chose to except misjoinder from harmless-error scrutiny.4
Justice Stevens’ partial dissent recognizes two further exceptions: (1) “when an independent value besides reliability of the outcome suggests that [harmless-error] analysis is inappropriate,” and (2) “when the harmlessness of the error cannot be measured with precision.” Post, at 474. Although the cases he cites to support these additional exceptions involved constitutional errors, Justice Stevens may well be correct in asserting that they also apply to errors governed by the statutory harmless-error provisions. I need not decide that question to conclude, as does Justice Stevens, that — like the first two exceptions — neither applies to misjoinder.
The applicability of the exception to protect values other than reliability is easily disposed of. Rules respecting joinder are based on recognition that the multiplication of charges or defendants may confuse the jury and lead to inferences of habitual criminality or guilt by association. McElroy, 164 U. S., at 80. Apart from this, however, joinder rules do not serve “an independent value besides reliability of the outcome” justifying an exception to the harmless-error principle. Surely it cannot be maintained that misjoinder affects a right so fundamental to a fair trial that it “ ‘infect[s] the validity of the underlying judgment itself, or the integrity of the process by which that judgment was obtained.’” Post, at 474, n. 15 (quoting Rose v. Lundy, 455 U. S. 509, 544 (1982) (Stevens, J., dissenting)).
*463The exception for errors as to which the prejudicial effect cannot be measured with precision requires closer consideration. As previously noted, McElroy held that misjoinder is per se reversible because a court can never safely conclude that it was not prejudicial. 164 U. S., at 81. However, trial courts routinely inquire into possible prejudice from joint trials when considering motions for severance under Federal Rule of Criminal Procedure 14, and appellate courts just as routinely perform that inquiry in reviewing Rule 14 rulings.5 To be sure, problems of jury confusion arising from misjoinder may be substantial. It is also quite easy for the jury to be prejudiced by evidence of other crimes or by inferences from an accused’s association with other defendants. Thus, it may be that, once the proper test for harmless error is applied, most misjoinders will in fact result in reversal. However, the prejudice that may result from mis-joinder is not so difficult to ascertain that it must always be presumed to be present. Whatever force the holding in McElroy may once have had, its precedential force has been greatly eroded by the 1919 Act, whose legislative history disapproves of such presumptions, supra, at 457-458, and by subsequent decisions such as Kotteakos.6 Today, adherence to the view that misjoinder is per se prejudicial would stand out as a stark and unjustified anomaly, leading to just the sort of unnecessary reversals that inspired enactment of the *4641919 Act. To the extent that McElroy states a contrary holding, I would overrule it.
1 — I I — I
The Court goes on to resolve the harmless-error question. I respectfully dissent. To begin with, I agree with Justice Stevens that “[undertaking a harmless-error analysis is perhaps the least useful function that this Court can perform.” Post, at 476. See United States v. Hasting, 461 U. S., at 520, n. 2 (opinion of Brennan, J.); see also, Connecticut v. Johnson, 460 U. S. 73, 102 (1983) (Powell, J., dissenting). Having concluded that a harmless-error inquiry is required, I, like Justice Stevens, think we should remand to the Court of Appeals, which is in a better position than we are to study the complete trial record with care.
Moreover, it is apparent that the Court’s perfunctory effort to evaluate the effect of this error is inadequate. The Court tells us simply that the error is harmless “[i]n the face of overwhelming evidence of guilt shown here . . . .” Ante, at 450. But where is the “examination of the proceedings in their entirety” called for by Kotteakos? See 328 U. S., at 762. Kotteakos instructs the reviewing court to “ponde[r] all that happened without stripping the erroneous action from the whole,” and expressly states that “[t]he inquiry cannot be merely whether there was enough to support the result, apart from the phase affected by the error.” Id., at 765. Obviously, the existence of overwhelming evidence is relevant to determining the “effect the error had or reasonably may be taken to have had upon the jury’s decision.” Id., at 764. But I would have thought it equally obvious that, at the very least, consideration of the magnitude of the error in the context of the trial would also be called for; this the Court has not done. The Court also tells us that the error was harmless because the same evidence “would likely have been admissible” at a joint retrial of the defendants without the improper count. Ante, at 450. However, as I thought *465Kotteakos made clear, that is irrelevant. The crucial thing is the effect the error had in the proceedings which actually took place, not whether the same thing could have been done in hypothetical proceedings. See 328 U. S., at 762-765. Harmless-error analysis is not an excuse for overlooking error because the reviewing court is itself convinced of the defendant’s guilt. The determination of guilt is for the jury to make, and the reviewing court is concerned solely with whether the error may have had a “substantial effect” upon that body.
Justice Traynor of the California Supreme Court wrote that “the evaluation of an error as harmless or prejudicial is one of the most significant tasks of an appellate court, as well as one of the most complex.” R. Traynor, The Riddle of Harmless Error 80 (1970). It is a task this Court is manifestly ill-equipped to undertake. See United States v. Hasting, supra, at 516-518 (Stevens, J., concurring in judgment). I would remand the cases for the Court of Appeals to undertake the task.
It scarcely needs repeating that, since correction may come from the legislature, considerations of stare decisis are at their strongest when this Court confronts its previous constructions of a statute. Cf. Burnet v. Coronado Oil & Gas Co., 285 U. S. 393, 406-407 (1932) (Brandéis, J., dissenting).
Until Chapman v. California, 386 U. S. 18 (1967), harmless-error analysis was considered inapplicable to errors respecting constitutional rights. See id., at 42-44 (Stewart, J., concurring in result) (“[I]n a long line of cases, involving a variety of constitutional claims in both state and federal prosecutions, this Court has steadfastly rejected any notion that constitutional violations might be disregarded on the ground that they were ‘harmless’” (citing and discussing examples)); see also, Kotteakos, 328 U. S., at 764-765, and n. 19. In Chapman, we altered this practice and held that “there may be some constitutional errors which in the setting of a particular case are so unimportant and insignificant that they may, consistent with the Federal Constitution, be deemed harmless, not requiring the automatic reversal of the conviction.” 386 U. S., at 22. Although we have since held that the Chapman harmless-error test applies to “most constitutional violations,” United States v. Hasting, 461 U. S., at 509, harmless-error analysis remains inapplicable to many constitutional rights. E. g., Vasquez v. Hillery, ante, p. 254 (discrimination in grand jury selection); Connecticut v. Johnson, 460 U. S. 73, 84-88 (1983) (opinion of Blackmun, J.) (Sandstrom violation); Gideon v. Wainwright, 372 U. S. 335 (1963) (right to counsel); Tumey v. Ohio, 273 U. S. 510 (1927) (right to impartial tribunal).
Because the source and nature of the harmless-error test for constitutional errors does not derive from §2111 or Rule 52(a), our cases concerning constitutional errors do not affect, and are not affected by, our decision today, which applies only to the statutory harmless-error doctrine.
But cf. Bruton v. United States, 391 U. S. 123 (1968). It is also possible that a particular case of misjoinder may be so egregious as to constitute a deprivation of due process. If this were the case, the error would be governed by Chapman rather than by §2111 or Rule 52(a). See n. 4, infra. Of course, a joinder of claims or parties that was so improper as to violate the Due Process Clause would undoubtedly also be prejudicial.
As explained above, the 1919 Act was not intended to codify a rule of per se reversal for particular rights, much less for misjoinder. Similarly, as the majority points out, nothing in the legislative history of Rule 8 indicates an intent to do anything more then set forth the technical requirements for and limitations on the joinder of claims or defendants. Ante, at 444-445, n. 7.
The Court correctly notes in its opinion, see ante, at 449-450, n. 12, that while the nature of the inquiry under Rules 8 and 14 is similar, the purposes and scope of these Rules are different.
Kotteakos rejected the argument that variance between the indictment and proof at trial should be per se reversible because such errors “naturally” result in prejudice. Relying on the legislative history of the harmless-error rule, the Court concluded that such presumptions should not lightly be inferred. “The only permissible presumption,” the Court said, “would seem to be particular, arising from the nature of the error and ‘its natural effect’ for or against prejudice in the particular setting.” 328 U. S., at 765-766.