WALKER, Circuit Judge:
Petitioner-appellant Michael Peck appeals from a judgment entered on June 2, 1994 in the United States District Court for the District of Connecticut (Alan H. Nevas, Judge) that granted in part and denied in part Peek’s petition for a writ of habeas corpus pursuant to 28 U.S.C. § 2255. Following a jury trial, Peck was convicted of two counts of structuring cash transactions to evade bank reporting requirements in violation of 31 U.S.C. §§ 5324(3), 5313(a), and 5322. Peck did not file a direct appeal.
In his petition, Peck contended that the Supreme Court’s decision in
Ratzlaf v. United States,
510 U.S. 135, 114 S.Ct. 655, 126 L.Ed.2d 615 (1994), which was decided after the time to appeal .from his judgment of conviction had expired, required that his convictions be vacated because the jury was erroneously instructed with respect to the scienter necessary for conviction. .
In
Peck v. United States,
73 F.3d 1220 (2d Cir.1995)
(“Peck I”),
a panel majority held that Peck had demonstrated, as required by
United States v. Frady,
456 U.S. 152, 167-68, 102 S.Ct. 1584, 1594-95, 71 L.Ed.2d 816 (1982), “cause” for his failure to pursue a direct appeal and “actual prejudice” resulting from the instructional error, and reversed the judgment of the district court. Thereafter, the full court granted the government’s petition to rehear the case in banc and heard oral argument. However, before the in bane court rendered a decision, the Supreme Court decided
California v. Roy,
— U.S. -, 117 S.Ct. 337, 136 L.Ed.2d 266 (1996) (per curiam), which substantially answered the question posed to the in banc court. The in banc court then voted to dissolve and to return the case to the panel for application of
Roy
to the facts of this case.
Peck v. United States,
102 F.3d 1319 (2d Cir.1996) (in banc) (per curiam)
(“Peck II
”).
Because we now conclude, in light of
Roy,
that the error in this case did not have a “ ‘substantial and injurious effect or influence in determining the jury’s verdict,’ ”
Brecht v.
Abrahamson, 507 U.S. 619, 637, 113 S.Ct. 1710, 1722, 123 L.Ed.2d 353 (1993) (quoting
Kotteakos v. United States,
328 U.S. 750, 776, 66 S.Ct. 1239, 1253, 90 L.Ed. 1557 (1946)), we vacate the judgment in
Peck I
and hold that the error in this case was harmless. Accordingly, the district court’s judgment is affirmed.
BACKGROUND
The facts of this ease are described in detail in
Peck I,
familiarity with which is assumed, and only a brief recitation of them will be provided here.
In July 1991, Peck, an attorney practicing in Hartford, Connecticut, was charged with tax evasion' and willful failure to file a tax return. Before his trial on the tax charges, which resulted in his conviction, Peck made a series of cash deposits that formed the basis
of his convictions in this case for structuring cash transactions.
The first count of the indictment in this case involved Peek’s financial transactions between November 7, 1991 and November 27, 1991. During that period, Peck made twelve cash deposits of $7,500 and two'cash deposits of $5,000 at various branches of the Society for Savings, into an account' maintained by his father. On November 7 alone, he made three deposits, each at a separate branch. He declined to make a deposit on the same day at a fourth branch after the bank informed him that the bank would have to file a Currency Transaction Report (“CTR”).
The second count related to similar conduct between March 27, 1992 and April 6, 1992. On March 27, 1992, Peck opened a checking account at Northeast Savings Bank by depositing a $50,000 cheek from a former law associate. Soon thereafter he made the following withdrawals from that account: $6,500 on March 30, $6,000 on March 31, $6,000 on April 1, and $6,500 on April 2. On April 3, he withdrew $7,000 from an account at the Bank of Boston, and $2,000 from an account at the Society for Savings. On April 6, he attempted to withdraw $5,000 from the Bank of Boston account, but when the bank told him that a CTR would have to be filed because of his April 3 withdrawal, he abandoned the transaction.
At trial, Peck testified to various innocent explanations for the multiple cash transactions. These explanations included that he was repaying an undocumented $100,000 loan from his father, even though he did not receive the loan proceeds until after he had deposited over $50,000 into his father’s account; that separate transactions were made on the same day because they were coincidental to other errands; that he decided not to follow through with some transactions upon being informed that a CTR would have to be filed because he was in too much of a hurry; that he found cash transactions more convenient; and that he never carried more than $7,500 in cash for reasons of safety.
Peck I,
73 F.3d at 1221-23.
The district court declined Peck’s request to charge the jury that the government must prove that Peck knew that structuring was unlawful in order to satisfy the scienter requirement of willfulness contained in 31 U.S.C. § 5322(a). The decision of Judge Nevas was in accord with the settled law of this circuit. This court had twice upheld the scienter instruction given by the district court, which required the government to prove only that the defendant intended to avoid triggering the banks’ reporting re
quirements.
See United States v. Caming,
968 F.2d 282, 238-41 (2d Cir.),
cert. denied,
506 U.S. 956, 113 S.Ct. 416, 121 L.Ed.2d 339 (1992);
United States v. Scanio,
900 F.2d 485, 489-91 (2d Cir.1990). On January 12, 1993, the jury found Peck guilty. On February 26, 1993, the district court sentenced Peck, in principal part, to two years imprisonment and three years of supervised release. Peck did not appeal.
On January 11, 1994, the Supreme Court held in
Ratzlaf
that the word “willfully” in 31 U.S.C. § 5322(a) requires the government to prove that the defendant knew that structuring was unlawful in order to establish a criminal currency structuring violation. 510 U.S. at 149, 114 S.Ct. at -. On April 15, 1994, Peck filed a petition for a writ of habeas corpus pursuant to 28 U.S.C. § 2255. Peck argued that his conviction should be vacated in light of
Ratzlaf
and, unrelated to the
Ratzlaf
issue, that he was entitled to sentencing relief as a result of an amendment to the United States Sentencing Guidelines. The district court granted habeas relief insofar as Peck requested a sentence reduction based upon the Guidelines amendment and denied his petition with respect to the
Rat-zlaf
issue based.on its finding that the result at trial “would [not] have been any different” if the correct scienter charge had been given to the jury.
Peck I,
73 F.3d at 1228. Peck appealed the denial of the petition.
In
Peck I,
a panel majority reversed the judgment of the district court. The majority held that the habeas petition must be granted because the jury that tried Peek did not “substantially answer[ ] the
Ratzlaf
question adversely to him in the course of its actual deliberations.”
Id.
The government requested a rehearing in banc. On May 31, 1996, the full court agreed to rehear the case in banc. As per the court’s order that day, the rehearing was:
limited to the issue of the proper standard for determining, on a collateral attack, whether the failure to give a correct jury instruction on a new element (required by authoritative decisional law announced after the verdict) has resulted in ‘actual prejudice’ within the meaning of
Frady,
456 U.S. at 168, 102 S.Ct. at 1594. Counsel [were directed to] consider whether ‘actual prejudice’ always results whenever the instruction required by
Ratzlaf
is not given, except in those cases where the reviewing court concludes that the jury’s findings resolved in the prosecution’s favor (by necessary implication) the issue that would have been posed by the
Ratzlaf
instruction, or whether ‘actual prejudice’ might not result under other circumstances, including where the reviewing court concludes that a rational jury, instructed in accordance with
Ratzlaf,
after hearing all the evidence, would necessarily have found the defendant guilty beyond a reasonable doubt.
On September 11, 1996, the in banc court heard oral argument. Before the in banc court rendered its decision, the Supreme Court decision in
Roy
clarified the harmless error analysis applicable to an error in a jury instruction on collateral review in the specific circumstances presented in
Peck, see Peck II,
102 F.3d 1319, 1324-26 (Newman, C.J., concurring), . thereby mooting the issue under consideration by the in banc court. On December 27, 1996, the in banc court dissolved itself and returned the case to this panel for reconsideration in light of
Roy. Peck II,
102 F.3d 1319, 1320-21.
DISCUSSION
I.
Harmless Error Jurisprudence
The history of this case illustrates the lack of precision in current harmless error jurisprudence, despite the frequency with which appellate courts must decide if error at trial is harmless. Chief Judge Newman’s insightful concurring opinion in
Peck II
identifies significant uncertainties left in the wake of recent Supreme Court decisions concerning harmless error analysis of federal constitutional error. However, despite these uncertainties, a framework is emerging from recent Supreme Court case-law. From our experience in
Peck I
and
Peck II,
we believe that articulating the framework would be helpful in resolving this ease. Additionally, the framework brings more precision to harmless error analysis, thus reducing (or at least isolat
ing) areas of disagreement—whether the disagreement is over the type of error, the harmless error standard to be applied, the degree of certainty required for finding an error harmless, or the methodology for determining if the constitutional error is harmless.
A.Structural Error versus Trial Error: Is Harmless Error Review Applicable?
The first task for a court in deciding a case involving a federal constitutional error is to determine if the error is “structural error” or “trial error.” Structural errors are fundamental defects in the trial mechanism that affect the entire “framework within which the trial proceeds, rather than simply an error in the trial process itself.”
Arizona v. Fulminante,
499 U.S. 279, 310, 111 S.Ct. 1246, 1264, 113 L.Ed.2d 302 (1991). Examples include the total deprivation of the right to counsel at trial,
see Gideon v. Wainwright,
372 U.S. 335, 83 S.Ct. 792, 9 L.Ed.2d 799 (1963), or the presence of a biased judge,
see Tumey v. Ohio,
273 U.S. 510, 47 S.Ct. 437, 71 L.Ed. 749 (1927).
See Fulminante,
499 U.S. at 309-10, 111 S.Ct. at 1264-65 (discussing five constitutional errors that are structural errors),
and, Sullivan v. Louisiana,
508 U.S. 275, 282, 113 S.Ct. 2078, 2083, 124 L.Ed.2d 182 (1993) (adding a sixth). We made clear in
Yarborough v. Keane,
101 F.3d 894 (2d Cir.1996), that the determination whether an error is structural turns not so much on the particular rule violated, as on whether the error was of sufficient consequence that the criminal process “‘cannot reliably serve its function as a vehicle for determination of guilt or innocence.’ ”
Fulminante,
499 U.S. at 310, 111 S.Ct. at 1265 (quoting
Rose v. Clark,
478 U.S. 570, 577-78, 106 S.Ct. 3101, 3105-06 (1986)).
If a reviewing court determines that the error is of the structural variety, the court’s task is at an end. Structural errors “defy analysis by harmless error standards” and the existence of a structural error “infeet[s] the entire trial process.”
Brecht,
507 U.S. at 629-30, 113 S.Ct. at 1717 (quoting
Fulminante,
499 U.S. at 309-10, 111 S.Ct. at 1264-65) (internal quotations omitted). Therefore, a per se rule of reversal applies when a structural error is present at trial, even if the record contains overwhelming evidence of guilt. If the error is not a structural error, it is a “trial error” and subject to harmless error review.
B.
The Applicable Harmless Error Standard: Direct versus Collateral Review
If the error is the more common “trial error,” the applicable harmless error standard for constitutional error differs depending on the procedural posture of the case. On direct review, the applicable standard for reviewing a constitutional error is whether the error “contribute[d] to the verdict.”
Chapman v. California,
386 U.S. 18, 24, 87 S.Ct. 824, 828, 17 L.Ed.2d 705 (1967).
On collateral review, as the Court explained in
Brecht,
considerations of finality, federalism, and comity warrant the application of a less-onerous harmless error standard. 507 U.S. at 635-37, 113 S.Ct. at 1720-22. In
Brecht,
the Court held that the appropriate standard applied on collateral review of federal constitutional error is the same as the standard applied on direct review of non-constitutional error, namely, whether the error “ ‘had substantial and injurious effect or influence in determining the jury’s verdict.’ ” 507 U.S. at 637, 113 S.Ct. at 1722 (quoting
Kotteakos,
328 U.S. at 776, 66 S.Ct. at 1253).
C.
The Degree of Certainty
The degree of certainty required— that is, how convinced a reviewing court must be before it can declare a federal constitutional error harmless—also differs depending on whether the review is direct or collateral. On direct review, the reviewing court must be convinced that the error is “harmless beyond a reasonable doubt.”
Chapman,
386 U.S. at 24, 87 S.Ct. at 828. On collateral review, in determining whether the constitutional error substantially influenced the jury’s decision, a reviewing court should grant relief if it is in “grave doubt as to the harmlessness” of a constitutional error.
O’Neal v. McAninch,
513 U.S. 432, -, 115 S.Ct. 992, 999, 130 L.Ed.2d 947 (1995). A judge is in “grave doubt” when “the matter is so evenly balanced that [the judge] feels
himself in virtual equipoise as to the harmlessness of the error.”
Id.
at-, 115 S.Ct. at 994.
D.
Methodology
The usual methodology for determining whether the harmlessness of a constitutional trial error is established with the requisite degree of certainty is to examine the record as a whole to determine if a rational jury, absent the error, would have arrived at the same verdict whether the standard is that of
Chapman
(direct review) or
Brecht
(collateral review). However, prior to the Supreme Court’s decision in
Roy,
courts were unsure as to the appropriate harmless error methodology to apply on habeas review of error in instructing the jury on the elements of the offense. It was this uncertainty that gave rise to our disagreement in
Peck I.
Supreme Court cases could be read to support the view of the panel majority in
Peck I,
and the view of the Ninth Circuit in
Roy v. Gomez,
81 F.3d 863, 867 (9th Cir.1996) (en banc), that the proper methodology for applying
Brecht,
based on the Court’s earlier decision in
Carella v. California,
491 U.S. 263, 109 S.Ct. 2419, 105 L.Ed.2d 218 (1989), is to determine if an actual jury finding resolved the missing or misdescribed instructional element.
See also United States v. Marder,
48 F.3d 564, 573 (1st Cir.1995);
United States v. Whiting,
28 F.3d 1296, 1309 & n. 12 (1st Cir.1994). There was also support for the view, based on the Court’s decision in
Brecht,
that a reviewing court should examine the record to determine whether a properly instructed, rational jury would have found the missing element.
See Peck I,
73 F.3d at 1230-33 (Walker, J., dissenting);
Roy,
81 F.3d at 869-70 (Wallace, J., dissenting);
Whiting,
28 F.3d at 1309 & n. 12;
Marder,
48 F.3d at 573. The Supreme Court’s decision in
Roy
resolved the issue for the circumstances presented in this case.
The respondent in
Roy
was a habeas petitioner pursuing a collateral attack upon a state court conviction for first-degree murder and robbery. At the trial in
Roy,
the jury charge correctly stated, under then-governing California law, that a person could be convicted of first-degree murder if,
inter aha,
a defendant had
knowledge
of his accomplice’s unlawful purpose and
assisted
in the accomplishment of the crime. — U.S. at -, 117 S.Ct. at 337-38. A California Supreme Court decision after Roy’s trial heightened the government’s scienter burden by requiring that a jury find that a defendant had the
intent or purpose
of assisting in the accomplice’s crime in order to convict a person of first-degree murder.
Id.
at -, 117 S.Ct. at 338. Upon habeas review, the Ninth Circuit held that the instructional error at the respondent’s trial was not harmless.
See Roy,
81 F.3d at 868.
The methodology used by the Ninth Circuit for determining if the error was harmless was substantially similar to the methodology employed in
Peck I.
The Ninth Circuit held that “the omission is harmless only if review of the facts found by the jury establishes that the jury
necessarily
found the omitted element.”
Roy,
81 F.3d at 867. The Ninth Circuit based its opinion “primarily [on] a concurring opinion in
Carella,
a case that dealt with legal presumptions.”
Roy,
— U.S. at -, 117 S.Ct. at 338.
In
Roy,
the Supreme Court reversed the Ninth Circuit, rejecting the
Carella
concurrence’s methodology for instructional error of the type present in
Roy.
The Court stated that, subsequent to its decision in
Carella, Brecht
held that in habeas proceedings, courts “ordinarily should apply the harmless error standard that the Court had previously enunciated in
Kotteakos.” Roy,
— U.S. at -, 117 S.Ct. at 338 (internal quotations omitted). The Court noted that although the standard erroneously applied by the Ninth Circuit was, “certainly consistent with the concurring opinion in
Carella,”
the decision in
Brecht,
reaffirmed in
O’Neal,
controlled harmless error review of trial errors on collateral attack.
Roy,
— U.S. at -, 117 S.Ct. at 389.
Although the Court discusses the issue in terms of “standards,” we think the dispute in
Roy,
like the dispute in
Peck I,
can be more precisely described as concerning the appropriate methodology for determining the harmlessness of an error. Indeed, the Ninth Circuit correctly applied the
Brecht
standard and the
O’Neal
“grave doubt” test,
see Roy,
81 F.3d at 868 (stating that “we must determine whether reversal is required under the
Brecht /O’Neal
line of cases”), but erred by applying the methodology discussed in the
Carella
concurrence as opposed to that discussed in
Brecht. See Roy,
81 F.3d at 868 (holding that because the court is “unable to say the jury
necessarily found
the required intent[, ujnder
Carella, Brecht,
and
O’Neal
” the error is not harmless).
In
Brecht,
the Court examined the “record as whole” to determine if the error was harmless. 507 U.S. at 638, 113 S.Ct. at 1722;
see id.
at 642, 113 S.Ct. at 1724 (Stevens, J., concurring) (noting that “[t]o apply the
Kotteakos
standard properly, the reviewing court must ... make a
de novo
examination of the trial record”).
Roy’s
rejection of the
Carella
concurrence’s methodology clarified that as to “trial error” in a jury instruction on an offense element,
Brecht
mandates that, on habeas review, a court examine the record as a whole.
Thus,
Roy
established that on collateral review,
Brecht
sets forth the correct methodology for determining if an instructional error of the type present in
Roy
is harmless. The correct methodology for determining, on direct review, whether trial error in the form of an omitted or misdescribed element of the offense in the jury instructions is harmless is still an open question. An answer to this question may be forthcoming.
See Johnson v. United States,
— U.S. -, 117 S.Ct. 451, 136 L.Ed.2d 346 (1996) (granting certio-rari on question of whether reversal of a perjury conviction is required when judge, not jury, decided element of materiality,
see
65 U.S.L.W. 3364 (Nov. 19, 1996)).
II..
The Error in Peck
In this case, if we conclude first that the instructional error at Peck’s trial is a “trial error,”
Roy
mandates that the proper methodology is to examine the entire • record to determine whether we are persuaded (with the degree of certainty set forth in
O’Neal)
that the error did not have a “substantial and injurious effect or influence in determining the jury’s verdict.”
Before proceeding to the harmless error analysis in this case, we note certain areas of agreement with
Peck I.
We agree with
Peck I
that
Ratzlaf
applies retroactively. 73 F.3d at 1224-25.- In addition,
Peck I
correctly held that in order to obtain habeas relief, Peck must satisfy the “cause and actual prejudice” standard of
Frady.
73 F.3d at 1225. Under
Frady,
Peck must demonstrate “cause” excusing his failure to appeal, and “actual prejudice” resulting from the instructional error. 456 U.S. at 167-68, 102 S.Ct. at 1594-95.
We adhere to our view in
Peck I
that Peck demonstrated “cause” for his failure to pursue on direct appeal his contention of instructional error because this court had definitively resolved the question as to the proper jury instruction regarding scienter and the Supreme Court had not agreed to review the issue prior to the deadline for Peck to file an appeal. 73 F.3d at 1225-26. Prior to
Rat-zlaf,
it was the law of this circuit that a criminal structuring violation “may be established without proof that the defendant knew that structuring is unlawful.”
Scanio,
900 F.2d at 491;
see Caming,
968 F.2d at 239. Therefore, at the time judgment was entered against Peck, an appeal based on Peek’s challenge to the jury instructions on willfullness given by Judge Nevas would have been frivolous.
Although we agree with the conclusions in
Peck I
that
Ratzlaf
applies retroactively and that Peek demonstrated “cause” for his failure to appeal, upon reconsideration we vacate the holding of
Peck I
that “actual prejudice” existed based upon the methodology, now rejected by
Roy,
of looking to the jury’s finding to see if the missing instructional element was also found.
At Peek’s trial, the jury was instructed as to Peck’s intent and willfulness that the government must prove that Peck had knowledge of the reporting requirements of 31 U.S.C. § 5313(a) and acted for the specific purpose of evading those reporting requirements.
Peck I,
73 F.3d at 1223. The omission of the heightened scienter requirement subsequently imposed by
Ratzlaf,
which required the knowledge that structuring is unlawful, was, like the error in
Roy,
an “error
in the instruction that defined the crime.”
Roy,
— U.S. at -, 117 S.Ct. at 339. In
Roy,
the Court stated that the instructional error in that case could be characterized as either an omission or a misdescription of an element of the crime.
Id.
The Court found that this type of error was a “trial error,” and therefore subject to harmless error review.
Id.
Like the error in
Roy,
the error at Peek’s trial is as easily characterized as a “misde-scription of an element” of the crime (the description of the knowledge requirement), as an error of “omission” of a component of an element (the omission of the heightened knowledge requirement). Thus, we conclude that, as in
Roy,
the error in this case is a “trial error.”
Because the error is a “trial error” in the context of collateral review, we examine the record as a whole to determine if, with the level of certainty set forth in
O’Neal,
the error is harmless under the
Brecht
standard. We agree with
Peck I’s
characterization of the evidence presented at Peek’s trial. In
Peck I,
the majority and dissent were in accord that Peek’s testimony “oscillated between the implausible and the preposterous.” 73 F.3d at 1228. We concluded that the jury’s verdict demonstrated “an evident and well founded disbelief of Peck’s testimony” and stated that the district court was correct in concluding that a properly instructed jury would have found Peek guilty of structuring.
Id.
In
Peck I,
the majority reversed the district court after concluding, consistent with the methodology of the
Carella
concurrence and the Ninth Circuit in
Roy,
that the jury that tried Peek did not make an actual finding on the heightened scienter element required by
Ratzlaf. Id.
However, the Supreme Court’s rejection of this methodology in
Roy
leads us to vacate our judgment in
Peck I.
Upon a careful review of the record as a whole, we conclude that a properly instructed, rational jury would have found the heightened scienter element required by
Ratzlaf
beyond a reasonable doubt. Thus, the instructional error did not have a “substantial and injurious effect or influence in determining the jury’s verdict” and we affirm Peck’s conviction.
Conclusion
Upon rehearing, we vacate our judgment in
Peck I
and affirm the district court’s denial in part of Peck’s petition for a writ of habeas corpus.