PUBLISHED
FILED: July 7, 2020
UNITED STATES COURT OF APPEALS
FOR THE FOURTH CIRCUIT
___________________
No. 18-4578
(3:17-cr-00809-JFA-1)
___________________
UNITED STATES OF AMERICA
Plaintiff - Appellee
v.
MICHAEL ANDREW GARY
Defendant - Appellant
___________________
ORDER
___________________
Appellee filed a petition for rehearing en banc, and Appellant filed a response in
opposition. The petition and response were circulated to the full Court. Judge Richardson
recused himself from the case. No member of the Court requested a poll on the petition
for en banc review. Therefore, the petition for rehearing en banc is denied.
Entered at the direction of Chief Judge Gregory.
For the Court
/s/ Patricia S. Connor, Clerk
WILKINSON, Circuit Judge, with whom Judges NIEMEYER, AGEE, QUATTLEBAUM,
and RUSHING join, concurring in the denial of rehearing en banc:
I concur in the denial of rehearing en banc for one reason and one reason only. The
panel’s holding is so incorrect and on an issue of such importance that I think the Supreme
Court should consider it promptly. Any en banc proceedings would only be a detour.
Many, many cases await the resolution of this question.
This court’s decision is far-reaching in its implications. It not only creates a circuit
split of yawning proportions, but also an equally profound schism with the Supreme
Court’s whole approach to error review and remediation. Is it eight—or nine—circuits that
disagree with us? I have lost count, but the ranks are growing. ∗
In Rehaif v. United States, 139 S. Ct. 2191, 2194 (2019), the Supreme Court held
that the government must prove that a defendant knew about his felony status as an element
of an 18 U.S.C. § 922(g) offense. Now our court holds that Rehaif error is a structural error
that is not amenable to harmless or to plain-error review. United States v. Gary, 954 F.3d
∗
Until now, no other circuit has treated a Rehaif error as structural when applying
plain-error review. Rather, the circuits have uniformly held that a defendant cannot show
an effect on his substantial rights where the evidence shows that the defendant knew of his
status as a felon at the time of his gun possession. See United States v. Burghardt, 939
F.3d 397, 403-05 (1st Cir. 2019); United States v. Balde, 943 F.3d 73, 97 (2d Cir. 2019);
United States v. Denson, 774 F. App’x 184, 185 (5th Cir. 2019); United States v. Hobbs,
953 F.3d 853, 857-58 (6th Cir. 2020); United States v. Williams, 946 F.3d 968, 973-75 (7th
Cir. 2020); United States v. Hollingshed, 940 F.3d 410, 415-16 (8th Cir. 2019); United
States v. Fisher, 796 F. App’x 504, 510-11 (10th Cir. 2019); United States v. McLellan,
2020 WL 2188875, at *6-7 (11th Cir. May 6, 2020); see also United States v. Benamor,
937 F.3d 1182, 1189 (9th Cir. 2019). Indeed, the Fifth Circuit has only recently rejected
this court’s structural error holding. United States v. Hicks, – F.3d –, 2020 WL 2301461,
at *2 (5th Cir. May 8, 2020).
2
194, 200 (4th Cir. 2020). Facts are so often the foundation of law. But the panel opinion,
put simply, takes flight from the facts in each and every case.
The retreat from the facts is especially egregious here. There is not a chance that
Gary’s claim would survive the third or fourth prongs of Olano or satisfy the reasonable-
probability test of Dominguez Benitez. Prior to the instant felon-in-possession offenses,
Gary was convicted of second-degree burglary and two counts of assault, each punishable
by more than one year in prison. For these offenses, he spent upwards of nine years in
prison. Moreover, at his sentencing hearing in 2017, Gary admitted that he knew it was
wrong for him to have a firearm. The Rehaif error could thus not have affected his
substantial rights because there is no possibility, not to mention a reasonable probability,
that Gary would not have pled guilty had he been informed of that which the government
could so easily have proven. And as to the fourth prong of Olano, the question simply
answers itself. In other words, considering the facts here, nothing about Gary’s Rehaif
claim has so much as a grain of merit.
The Supreme Court has made clear that structural errors are few and far between.
This point has been made not once but repeatedly. The Court has found structural error
only in a “very limited class of cases,” Johnson v. United States, 520 U.S. 461, 468 (1997),
and has instead “adopted the general rule that a constitutional error does not automatically
require reversal of a conviction,” Arizona v. Fulminante, 499 U.S. 279, 306 (1991).
The narrow band of structural errors is distinct because they inherently taint the
integrity of a trial from beginning to end. See, e.g., United States v. Gonzalez-Lopez, 548
U.S. 140, 150 (2006) (denial of right to counsel); Sullivan v. Louisiana, 508 U.S. 275, 281-
3
82 (1993) (defective reasonable-doubt instruction); Vasquez v. Hillery, 474 U.S. 254, 263-
64 (1986) (racial discrimination in grand jury selection); Tumey v. Ohio, 273 U.S. 510, 535
(1927) (lack of impartial judge).
Structural errors are to be limited, in other words, to the kind of error that by itself
invalidates the criminal proceeding. See Weaver v. Massachusetts, 137 S. Ct. 1899, 1907
(2017) (defining structural error as one that “affects the framework within which the trial
proceeds, rather than being simply an error in the trial process itself”) (internal markings
and quotation omitted). A denial of the right to counsel, racial bias in criminal justice
proceedings, and an infirm reasonable doubt instruction are easily identified as the category
of error that sweeps across any particular offense, and speaks overarchingly to the kind of
flaws that any citizen would instinctively know to be both unlawful and unfair. Put
otherwise, structural errors are innately infectious, necessarily impugning each part of a
trial, rather than potentially consequential, depending on the facts and circumstances of a
given case. Because such errors lack a ready way to quantify their impact, they defy
analysis by harmless or plain-error review.
A Rehaif error comes nowhere near this level. It is not even close. Rather, it belongs
with the large category of errors that the Court has deemed non-structural, in recognition
of the fact that the illusory search for perfection in the criminal justice process can so easily,
as the saying goes, become the enemy of the good. See, e.g., Delaware v. Van Arsdall,
475 U.S. 673, 681-82 (1986) (holding that a restriction on defendant’s ability to cross-
examine witness in violation of Sixth Amendment was non-structural error); United States
v. Hasting, 461 U.S. 499, 509 (1983) (same for improper remark regarding defendant’s
4
silence at trial in violation of Fifth Amendment); Chambers v. Maroney, 399 U.S. 42, 52-
53 (1970) (same for admission of evidence taken in violation of Fourth Amendment); see
also Fulminante, 499 U.S. at 306-07 (collecting cases). To borrow from the Van Arsdall
Court, these cases stand for “the principle that the central purpose of a criminal trial is to
decide the factual question of the defendant’s guilt or innocence, and promote[] public
respect for the criminal process by focusing on the underlying fairness of the trial rather
than on the virtually inevitable presence of immaterial error.” Van Arsdall, 475 U.S. at 681.
Nor does Rehaif error gain admittance into the gallery of structural error because it
pertains to an element of a charge. I intend no disrespect to the Supreme Court’s fine
decision in Rehaif to note that the appearance of what our court now terms a structural
defect has come rather late in the day, after many decades of prior practice to the contrary—
a fact that should weigh not only against retroactive review, but against recognition of the
error as a structural one as well.
To that end, the Supreme Court has plainly resisted the linkage between elements
errors and structural error. Take Neder v. United States, for example. 527 U.S. 1 (1999).
There, the trial court incorrectly omitted from the jury instructions an element of the
offense charged against Neder. The Court squarely rejected the argument that such an error
was structural, and instead reviewed for harmless error because “an instruction that omits
an element of the offense does not necessarily render a criminal trial fundamentally unfair
or an unreliable vehicle for determining guilt or innocence.” Id. at 9. Indeed, time and
again, the Court has applied harmless error review to elements errors. See, e.g., Johnson,
520 U.S. at 468-69; California v. Roy, 519 U.S. 2, 5 (1996); Yates v. Evatt, 500 U.S. 391,
5
393 (1991); Carella v. California, 491 U.S. 263, 266 (1989); Pope v. Illinois, 481 U.S. 497,
502-503 (1987).
Moreover, there are many reasons to think that the guilty plea context is an
especially poor one for recognizing an elements error as structural. Indeed, the Court has
already decided as much. See Henderson v. Morgan, 426 U.S. 637, 645-46 (1976)
(surveying factual record before vacating defendant’s plea on the ground that he was
misinformed as to a key element of the charge against him); Bousley v. United States, 523
U.S. 614, 622-23 (1998) (applying fact-bound exceptions for excusing procedural default
to defendant’s claim that his guilty plea was invalid because he was misinformed as to the
government’s burden of proof at time of his plea).
In fact, the guilty plea context is one in which the Court has assiduously resisted
automatic vacatur of a plea. The standard for vacatur is quite fact-dependent. It asks
whether there is a “reasonable probability” that absent the error the defendant would not
have entered the plea. United States v. Dominguez Benitez, 542 U.S. 74, 83 (2004); see
also id. at 81 n.6 (noting single Rule 11 error “not colorably structural”). I see no reason
whatsoever why this standard, so widely adopted and so pervasively relied upon, should
be rendered nugatory on the basis of the kind or character of the error asserted rather than
its impact on the particular proceeding. The Supreme Court’s standard stands in stark
contrast to the novel standard adopted in this case. The reasonable probability standard is
conscientiously attentive to facts. Our court’s opinion is wholly oblivious to them.
Furthermore, the guilty plea context is one where the Supreme Court, again in
contrast to this court, has been especially attentive to finality. United States v. Davila, 569
6
U.S. 597, 608 (2013) (stressing the “particular importance of the finality of guilty pleas”)
(internal markings and quotation omitted). The reason for the finality of guilty pleas could
not be more evident or obvious. Not only are they uniquely susceptible to buyer’s remorse;
they rest on a well-understood tradeoff—one where the defendant receives the certainty of
present benefits in exchange for forgoing the possibility of some future benefit down the
road. Rehaif is exactly the sort of future decisional benefit a routine guilty plea ordinarily
waives. The panel opinion now seeks to undo all this. For if Rehaif is deemed structural,
it could be used to overturn a wealth of pleas at some future point in time, given that the
courts are more likely to give structural errors retroactive application.
The costs to criminal justice of the panel’s ruling are immense. If they were offset
by some gain in the administration of justice, that would be one thing. But, as Gary’s case
shows, the vast majority of defendants who will seek to take advantage of a structural
Rehaif error are perfectly aware of their felony status. Felony status is simply not the kind
of thing that one forgets. It is, after all, a § 922(g) offense that one has pled guilty to, a
plea that would wholly lack a factual basis before or after Rehaif if the defendant were not
a felon. For those very few who claim plausibly to be unaware of their felony status, the
reasonable probability standard in Dominguez-Benitez stands ready to pick them up.
I hesitate to raise such a mundane consideration as the resources available to judges
and litigants in discussing this issue, but it remains the fact that the resources of our system
are finite. The erosion of finality in the context of such basic criminal offenses as § 922(g)
and 18 U.S.C. § 924(c) will strain the resources of the lower federal courts in no small
measure. Not only that, but prosecutorial resources will be tested to the limits by
7
multitudes of defendants seeking to withdraw and renegotiate their pleas. Not only that,
but the resources of public defenders will be tested as well, as they try to balance their
obligations to existing clients who face serious charges and the legions of defendants who
seek to redo past bargains. We are adding not just one, but two major burdens to our
system. The first arises from the sheer volume of guilty pleas. The second arises from the
fact that § 922(g) is at or near the top of our most frequently charged criminal offenses.
In many instances, the reform of the criminal justice system is salutary. The First
Step Act is an example of a long-overdue reform, see Pub. L. No. 115-391, 132 Stat. 5194
(2018), notwithstanding the fact that it adds to the business of our justice system. But there
are limits to the extent that each new twist and turn in decisional law should spark a crisis
of volume in criminal justice administration.
Volume diminishes those qualities of conscientious deliberation for which, I
suspect, each of us who plays some part in criminal justice proceedings would like to be
known. This court’s ruling is unfortunate in so many ways. I respect the decisions of my
colleagues, but I do hope that the Supreme Court will undo the error here and align us with
the other circuits in our country.
8