In the

    United States Court of Appeals
               For the Seventh Circuit
                    ____________________
No. 14-3611
UNITED STATES OF AMERICA,
                                                 Plaintiff-Appellee,

                               v.

TONY A. HURLBURT,
                                             Defendant-Appellant.
                    ____________________

           Appeal from the United States District Court
              for the Western District of Wisconsin.
           No. 14-cr-62-jdp — James D. Peterson, Judge.
                    ____________________

No. 15-1686
UNITED STATES OF AMERICA,
                                                 Plaintiff-Appellee,

                               v.

JOSHUA GILLESPIE,
                                             Defendant-Appellant.
                    ____________________

           Appeal from the United States District Court
               for the Western District of Wisconsin.
       No. 14-cr-106-wmc — William C. Conley, Chief Judge.
2                                         Nos. 14-3611 & 15-1686

                     ____________________

    ARGUED DECEMBER 2, 2015 — DECIDED AUGUST 29, 2016
                ____________________


    BeforeWOOD, Chief Judge, and POSNER, FLAUM,
EASTERBROOK, KANNE, ROVNER, WILLIAMS, SYKES, and
HAMILTON, Circuit Judges.
    SYKES, Circuit Judge. Tony Hurlburt and Joshua Gillespie
pleaded guilty in separate cases to unlawfully possessing a
firearm as a felon. See 18 U.S.C. § 922(g)(1). Their appeals
raise the same legal issue, so we’ve consolidated them for
decision. To calculate the Sentencing Guidelines range in
each case, the district court began with U.S.S.G. § 2K2.1(a),
which assigns progressively higher offense levels if the
defendant has one or more prior convictions for a “crime of
violence.” The term “crime of violence” is defined in the
career-offender guideline and includes “any offense …
that … is burglary of a dwelling, arson, or extortion, in-
volves use of explosives, or otherwise involves conduct that
presents a serious potential risk of physical injury to another.” 

(2014) (emphasis added). The highlighted text is
known as the “residual clause.”
    The residual clause in § 4B1.2(a)(2) mirrors the residual
clause in the Armed Career Criminal Act (“ACCA”), which
steeply increases the minimum and maximum penalties for
§ 922(g) violations. 18 U.S.C. § 924(e)(2)(B). One year ago the
Supreme Court invalidated the ACCA’s residual clause as
unconstitutionally vague. Johnson v. United States, 

, 2563 (2015). The question here is whether Johnson’s
holding applies to the parallel residual clause in the career-
Nos. 14-3611 & 15-1686                                         3

offender guideline. An emerging consensus of the circuits
holds that it does. See infra pp. 16–17.
    In this circuit, however, vagueness challenges to the Sen-
tencing Guidelines are categorically foreclosed. Circuit
precedent—namely, United States v. Tichenor 

,
364–65 (7th Cir. 2012)—holds that the Guidelines are not
susceptible to challenge on vagueness grounds. But Tichenor
was decided before Johnson and Peugh v. United States,

(2013), which have fatally undermined its
reasoning. Accordingly, we now overrule Tichenor. Applying
Johnson, we join the increasing majority of our sister circuits
in holding that the residual clause in § 4B1.2(a)(2) is uncon-
stitutionally vague.
                       I. Background
   Tony Hurlburt was charged in a two-count indictment
with possessing a firearm as a felon, see § 922(g)(1), and
possessing a short-barreled shotgun, see 26 U.S.C. §§ 5841,
5845(a)(2), and 5861(d). He pleaded guilty to the felon-in-
possession count; the second count was dismissed.
    Under the Sentencing Guidelines, the offense level for the
crime of unlawful firearm possession depends in part on the
defendant’s criminal history. For Hurlburt’s crime the base
offense level ordinarily is 18. § 2K2.1(a)(5). But if the defend-
ant has a prior conviction for a “crime of violence or a
controlled substance offense,” the base offense level is 22.
§ 2K2.1(a)(3). For a defendant with two or more prior convic-
tions of either type, the base offense level jumps to 26.
§ 2K2.1(a)(1).
4                                     Nos. 14-3611 & 15-1686

   Application Note 1 to § 2K2.1 incorporates the “crime of
violence” definition in the career-offender guideline, which
reads:
      (a) The term “crime of violence” means any of-
      fense under federal or state law, punishable by
      imprisonment for a term exceeding one year,
      that—
          (1) has as an element the use, attempted
          use, or threatened use of physical force
          against the person of another, or
          (2) is burglary of a dwelling, arson, or ex-
          tortion, involves use of explosives, or other-
          wise involves conduct that presents a serious
          potential risk of physical injury to another.
U.S.S.G. § 4B1.2(a) (emphasis added). The highlighted text is
known as the “residual clause.”
    Hurlburt has a prior conviction for armed burglary, and
at sentencing he conceded that this conviction qualifies as a
predicate crime of violence. The government argued that
another of Hurlburt’s prior convictions—for discharging a
firearm into a building or vehicle, see WIS. STAT. § 941.20—
should also count as a crime of violence. More particularly,
the government argued that this second conviction qualified
under § 4B1.2(a)(2)’s residual clause. Over Hurlburt’s objec-
tion the district judge accepted this argument.
   With two predicate convictions for crimes of violence,
Hurlburt’s base offense level was 26, and the recommended
sentencing range was 84–105 months. The judge imposed a
below-range sentence of 72 months. Without the second
Nos. 14-3611 & 15-1686                                                 5

career-offender predicate, the Guidelines range drops to 57–
71 months.
    In an unrelated case in the same district, Joshua Gillespie
was indicted for unlawfully possessing a firearm as a felon,
and he too pleaded guilty. Gillespie has a prior conviction
for fleeing an officer. See 

The district judge
counted this conviction as a predicate crime of violence
under the residual clause, which increased Gillespie’s base
offense level to 20. § 2K2.1(a)(4). The resulting Guidelines
range was 92–115 months, and the judge imposed a below-
range sentence of 84 months. Without the career-offender
predicate in the mix, the Guidelines range drops to 51–
63 months.
    Hurlburt and Gillespie appealed and immediately asked
us to suspend briefing to await the Supreme Court’s decision
in Johnson, which raised the question whether the residual
clause in the ACCA’s definition of “violent felony”—a
mirror image of the residual clause in § 4B1.2(a)(2)—is
unconstitutionally vague. We held the cases for Johnson and
reinstated briefing after the Supreme Court issued its opin-
ion. A panel heard argument in both appeals on the same
day. 1 The panel prepared an opinion proposing to overrule
Tichenor and circulated it to the full court in accordance with
Circuit Rule 40(e). An en banc vote followed, and a majority
of the court approved. This, then, is the opinion of the en




1 Another appeal we decide today, United States v. Rollins, No. 13-1731,
also raises the same issue and was argued the same day. Because Rollins
presents an additional issue unique to that case, we have not consolidat-
ed it here.
6                                              Nos. 14-3611 & 15-1686

banc court. 2 See Buchmeier v. United States, 

(7th
Cir. 2009) (using the same procedure).
                           II. Discussion
    In Johnson the Supreme Court held that the ACCA’s re-
sidual clause is too vague to satisfy minimum requirements
of due 

. Hurlburt and Gillespie
argue that Johnson’s holding applies to the identically
phrased residual clause in § 4B1.2(a)(2), which was used in
their cases to increase the base offense level and thus the
recommended sentencing range under the Guidelines. The
Johnson argument is new on appeal, so our review is for
plain error. Henderson v. United States, 

, 1124
(2013); FED. R. CRIM. P. 52(b). Under Rule 52(b) we may
correct a forfeited error if (1) the error is “plain”; (2) affects
the defendant’s “substantial rights”; and (3) “seriously
affects the fairness, integrity, or public reputation of [the]
judicial proceedings.” 

–27 (quo-
tation marks omitted).
    Johnson was not yet decided when the defendants were
sentenced, but plain-error review asks whether the error is
“plain” at the time of appellate review. 

The
defendants maintain that the Johnson error is plain: The two
residual clauses are identical, and because the ACCA’s
residual clause is unconstitutionally vague, it necessarily
follows that the residual clause in § 4B1.2(a)(2) is also uncon-
stitutional.



2District Judge J. Phil Gilbert, of the Southern District of Illinois, served
on the original panel, sitting by designation. We appreciate his willing-
ness to assist the court.
Nos. 14-3611 & 15-1686                                          7

    The logic is compelling, but our decision in Tichenor
stands in the way. Tichenor held that the Guidelines cannot
be challenged on vagueness 

–65. The
defendants maintain that Tichenor has been fatally under-
mined by the Supreme Court’s decisions in Johnson and
Peugh. The government agrees, so the parties join forces in
asking us to overrule Tichenor, apply Johnson, and invalidate
the residual clause in § 4B1.2(a)(2) as unconstitutionally
vague.
    Of course the parties’ agreement doesn’t relieve us of our
obligation to resolve the question ourselves. Sibron v. New
York, 

, 58 (1968). Before proceeding, however, we
pause to note two important recent developments. First, the
Sentencing Commission has amended the Guidelines to
delete § 4B1.2(a)(2)’s residual clause in light of Johnson; the
amendment became effective August 1, 2016. 81 Fed. Reg.
4741, 4742 (2016). Second, the Supreme Court has granted
certiorari in a case on collateral review to address the precise
question presented here: whether Johnson’s holding applies
to the residual clause in § 4B1.2(a)(2). Beckles v. United States,
616 F. App’x 415 (11th Cir. 2015), cert. granted, 

(U.S. June 27, 2016) (No. 15-8544). Beckles will be
heard in the Court’s upcoming term and raises additional
issues unique to its facts and procedural posture. The
Court’s decision is many months away, so we think it best
not to hold these cases for Beckles.
8                                           Nos. 14-3611 & 15-1686

A. Johnson and § 4B1.2(a)(2)’s Residual Clause
    The Due Process Clause 3 prohibits the government from
depriving a person of life, liberty, or property “under a
criminal law so vague that it fails to give ordinary people
fair notice of the conduct it punishes, or so standardless that
it invites arbitrary enforcement.” 

.
Johnson addressed persistent vagueness concerns about the
residual clause in the ACCA’s definition of “violent felony.”
The Act increases the minimum and maximum penalties for
various firearm-possession offenses if the defendant has
three prior convictions “for a violent felony or a serious drug
offense.” 18 U.S.C. § 924(e)(1). The Act defines the term
“violent felony” as follows:
       [A]ny crime punishable by imprisonment for a
       term exceeding one year … that—
           (i) has as an element the use, attempted use,
           or threatened use of physical force against
           the person of another; or
           (ii) is burglary, arson, or extortion, involves
           use of explosives, or otherwise involves con-
           duct that presents a serious potential risk of
           physical injury to another … .
§ 924(e)(2)(B) (emphasis added). The highlighted text is the
residual clause; the residual clause in the career-offender
guideline is a carbon copy.



3  The Fifth Amendment provides: “No person shall … be deprived of
life, liberty, or property, without due process of law … .” U.S. CONST.
amend. V.
Nos. 14-3611 & 15-1686                                       9

    The Court began its analysis in Johnson by reaffirming the
principle that the Constitution’s prohibition of vague laws
applies “not only to statutes defining elements of crimes, but
also to statutes fixing 

(citing
United States v. Batchelder, 

, 123 (1979)). Before
continuing the doctrinal analysis, the Court first explained
that the residual clause mandates a two-step categorical
approach for classifying crimes as violent felonies. In the
first step, the sentencing court must evaluate the predicate
crime of conviction, hypothesizing “the kind of conduct that
the crime involves in ‘the ordinary case,’” rather than look-
ing to the actual facts of the underlying case; in the second
step, the court must assess whether this hypothesized “ordi-
nary” case of the crime “presents a serious potential risk of
physical injury.” 

features of the residual clause, the Court said,
“conspire to make it unconstitutionally vague.” 

clause “leaves grave uncertainty about how to estimate the
risk posed by a crime” because “[i]t ties the judicial assess-
ment of risk to a judicially imagined ‘ordinary case’ of a
crime, not to real-world facts or statutory elements.” Id.; see
also 

one go about deciding what kind of
conduct the ‘ordinary case’ of the crime involves? ‘A statisti-
cal analysis of the state reporter? A survey? Expert evidence?
Google? Gut instinct?’” (quoting United States v. Mayer,

, 952 (9th Cir. 2009) (Kozinski, C.J., dissenting
from denial of rehearing en banc))). Second, the residual
clause “leaves uncertainty about how much risk it takes for a
crime to qualify as a violent felony.” 

    “By combining indeterminacy about how to measure the
risk posed by a crime with indeterminacy about how much
10                                      Nos. 14-3611 & 15-1686

risk it takes for the crime to qualify as a violent felony, the
residual clause produces more unpredictability and arbitrar-
iness than the Due Process Clause tolerates.” 

noted as well that the residual clause had persistently resist-
ed judicial efforts—by the Justices themselves and the lower
courts—to settle on a stable construction. 

This
interpretive struggle, the Court said, was a “failed enter-
prise,” 

and “the experience of the federal courts
leaves no doubt about the unavoidable uncertainty and
arbitrariness of adjudication under the residual clause,” 

    The clause’s “hopeless indeterminacy,” the Court con-
cluded, “both denies fair notice to defendants and invites
arbitrary enforcement by judges.” 

The Court
called a halt to the long-running interpretive battle and held
that “imposing an increased sentence under the [ACCA’s]
residual clause … violates the Constitution’s guarantee of
due process.” 

2563.
    As we’ve explained, § 4B1.2(a)(2)’s definition of “violent
felony” contains the same residual clause, and we interpret
the two provisions interchangeably, using the same categori-
cal approach that Johnson found impermissibly indetermi-
nate. See, e.g., United States v. Griffin, 

, 802 (7th
Cir. 2011); United States v. Spells, 

, 749 n.1 (7th
Cir. 2008). So unless the Guidelines are immune from chal-
lenge on vagueness grounds, it follows inexorably from
Johnson that the residual clause in § 4B1.2(a)(2) is also uncon-
stitutionally vague. See United States v. Vivas-Ceja, 

, 722–23 (7th Cir. 2015) (applying Johnson to the similarly
phrased residual clause in the “crime of violence” definition
in 18 U.S.C. § 16(b)).
Nos. 14-3611 & 15-1686                                      11

B. Tichenor
    Tichenor held that the Guidelines are immune from
vagueness challenges, but that conclusion is on shaky
ground after Johnson and Peugh. Our decision in Tichenor
rested on two premises. First, we reasoned that vagueness
doctrine doesn’t apply to the Guidelines because they do not
declare any conduct illegal; they’re just “directives to judges
for their guidance in sentencing.” 

(quoting United States v. Wivell, 

, 160 (8th Cir.
1990)). The second premise overlaps the first: We reasoned
that vagueness doctrine doesn’t apply because United States
v. Booker, 

(2005), demoted the Guidelines from
“rules to advice.” 

(quotation marks
omitted). Because the Guidelines are merely advisory, a
defendant has no due-process expectation that he will be
sentenced within the recommended range, 

v. United States, 

, 713–14 (2008)), and he “cannot
rely on them to communicate the sentence that the district
court will impose,” 

    Johnson conclusively refutes Tichenor’s first premise. Cit-
ing 

, the Court confirmed that
vagueness doctrine applies to sentencing provisions as well
as laws declaring conduct illegal. 

.
We touched on Batchelder in Tichenor but dismissed it as
irrelevant, construing its reference to “vague sentencing
provisions” as “mere 

. We now
know the Supreme Court sees things differently; the Consti-
tution’s protection against vague laws is not limited to laws
defining criminal liability. As far as vagueness doctrine is
concerned, it makes no difference that the Guidelines deal
exclusively with sentencing.
12                                      Nos. 14-3611 & 15-1686

    Tichenor’s second rationale has been fatally undermined
by the Court’s decision in Peugh. There the Court held that
the Guidelines, even though advisory, are subject to the
limits imposed by the Ex Post Facto Clause. 

–78. A district judge in Northern Illinois had calculat-
ed Peugh’s sentencing range under the version of the Guide-
lines in effect at the time of sentencing, as required by
18 U.S.C. § 3553(a)(4)(A)(ii), and this newer version yielded a
higher range than the version in effect at the time of the
crime. 

Peugh objected, arguing that applying
the harsher version of the Guidelines promulgated after he
committed the crime violated his rights under the Ex Post
Facto Clause. Based on our circuit precedent, the district
judge rejected the argument, and we affirmed. United States
v. Peugh, 

, 741 (7th Cir. 2012) (citing United States
v. Demaree, 

(7th Cir. 2006)). The Supreme Court
reversed, finding an ex post facto violation.
    The Ex Post Facto Clause prohibits, among other things,
“[e]very law that changes the punishment, and inflicts a
greater punishment, than the law annexed to the crime,
when committed.” 

(quoting Calder v.
Bull, 3 U.S. (3 Dall.) 386, 390 (1798) (opinion of Chase, J.)).
The “touchstone of [the Ex Post Facto Clause] inquiry is
whether a given change in law presents a sufficient risk of
increasing the measure of punishment attached to the cov-
ered crimes.” 

(internal quotation marks omitted).
   As the Court explained in Peugh, the Ex Post Facto Clause
“ensures that individuals have fair warning of applicable
laws and guards against vindictive legislative action.” 

And “[e]ven where these concerns are not directly
implicated, … the Clause also safeguards ‘a fundamental
Nos. 14-3611 & 15-1686                                         13

fairness interest … in having the government abide by the
rules of law it establishes to govern the circumstances under
which it can deprive a person of his or her liberty or life.’” 

v. Texas, 

, 533 (2000)); see also
Miller v. Florida, 

, 429 (1987) (explaining that the
Ex Post Facto Clauses “assure that federal and state legisla-
tures [are] restrained from enacting arbitrary or vindictive
legislation” and that “legislative enactments give fair warn-
ing of their effect and permit individuals to rely on their
meaning until explicitly changed”) (internal quotation
omitted)).
    Crucially here, the government argued in Peugh that be-
cause the post-Booker Guidelines are advisory, they “lack
sufficient legal effect” to be considered “law” for purposes of
the Ex Post Facto Clause. 

; 

(“The Government … argues … that the Guidelines are too
much like guideposts and not enough like fences to give rise
to an ex post facto violation.”). The Court emphatically reject-
ed that argument, explaining that “[t]he post-Booker federal
sentencing scheme aims to achieve uniformity by ensuring
that sentencing decisions are anchored by the Guidelines
and that they remain a meaningful benchmark through the
process of appellate review.” 

This “anchoring”
effect, the Court said, is enough to implicate the concerns
underlying the Ex Post Facto Clause; the sentencing court’s
discretion to sentence outside the Guidelines range “[does]
not defeat an ex post facto claim.” 

    The Court explicitly listed the procedural rules and ap-
pellate-review standards that give the post-Booker Guidelines
a degree of “binding legal effect” sufficient to raise ex post
facto concerns. 

District judges must begin their
14                                      Nos. 14-3611 & 15-1686

sentencing analysis with the Guidelines and correctly calcu-
late the applicable sentencing range. 

(citing Gall v.
United States, 

, 50 (2007)). Failure to correctly
calculate the Guidelines range is procedural error. 

552 U.S. at 51). Variances above or below the range
must be specifically justified, and more justification is need-
ed as the variance increases. 

even though “a
district court may ultimately sentence a given defendant
outside the … range,” the Guidelines retain “force” as the
“framework for sentencing.” 

rule that an
incorrect Guidelines calculation is procedural error ensures
that they remain the starting point for every sentencing
calculation … .” 

most importantly, reviewing courts may
presume that a within-Guidelines sentence is reasonable. 

has adopted such a presumption. See United
States v. Mykytiuk, 

, 608 (7th Cir. 2005) (“[A]ny
sentence that is properly calculated under the Guidelines is
entitled to a rebuttable presumption of reasonableness [on
appeal].”).
    With all these formal procedural requirements, the post-
Booker Guidelines, though ultimately advisory, are not
“merely a volume that the district court reads with academic
interest in the course of sentencing.” 

. To the contrary, the Guidelines have real-world conse-
quences despite their demotion to advisory status. Peugh
pointed to “considerable empirical evidence” establishing
that the Guidelines continue to have “the intended effect of
influencing the sentences imposed by judges.” 

Data from the Sentencing Commission “indicate that when a
Guidelines range moves up or down, offenders’ sentences
Nos. 14-3611 & 15-1686                                        15

move with it.” 

Court put it very recently, “[t]hese
sources confirm that the Guidelines are not only the starting
point for most federal sentencing proceedings but also the
lodestar.” Molina-Martinez v. United States, 

,
1346 (2016).
    This combination of formal legal requirements and real-
world effects led the Court to conclude that the Guidelines,
though advisory, are not immune from Ex Post Facto Clause
scrutiny. 

. The Court went on to hold
that “[a] retrospective increase in the Guidelines range …
creates a sufficient risk of a higher sentence to constitute an
ex post facto violation.” 

be clear from this discussion that Tichenor’s sec-
ond premise—that the Guidelines’ advisory status insulates
them from vagueness challenges—did not survive Peugh.
The Court held, after all, that the Guidelines are sufficiently
law-like to trigger Ex Post Facto protection. If the Guidelines
are constraining enough to require compliance with the Ex
Post Facto Clause, it follows that they are constraining
enough to require compliance with the Due Process Clause’s
prohibition against vague laws. We see no principled way to
distinguish Peugh on doctrinal grounds: The two constitu-
tional protections share the same underlying concerns about
fair notice and arbitrary governmental action.
    To the extent that Tichenor relied on Irizarry, Peugh explic-
itly considered and rejected the analogy. Irizarry held that
district judges are not required to give notice before impos-
ing an above-Guidelines sentence based on the sentencing
factors in 18 U.S.C. § 

. Peugh distin-
guished Irizarry this way:
16                                      Nos. 14-3611 & 15-1686

       It is true that we held, in Irizarry v. United
       States, 

, 713–714 … , that a defend-
       ant does not have an “expectation subject to
       due process protection” that he will be sen-
       tenced within the Guidelines range. But … the
       Ex Post Facto Clause does not merely protect re-
       liance interests. It also reflects principles of
       “fundamental justice.” 

,
       

.

. The Court’s reference to “reliance
interests” is shorthand for the fair-notice principle, and its
reference to “fundamental justice” captures the Ex Post Facto
Clause’s concern about arbitrary governmental action.
Vagueness doctrine reflects the same concerns. Indeed, in
Johnson the Court concluded that the ACCA’s residual clause
both denies fair notice and invites arbitrary enforcement. But
the fair-notice principle was mostly in the background; the
Court’s chief concern was that the radical indeterminacy of
the residual clause made judicial enforcement essentially ad
hoc and arbitrary.
   Irizarry is also distinguishable for another reason. It ad-
dressed a question about procedural notice: Must the sentenc-
ing court give the defendant notice and an opportunity to
respond before imposing an above-Guidelines sentence
under § 

–13. Vagueness doctrine and
the Ex Post Facto Clause enforce a different notice principle:
the substantive requirement that the law must give clear
notice of the conduct that it prohibits and the consequences
that attach to a violation. See United States v. Pawlak, 

, 909 (6th Cir. 2016) (explaining the difference between
Nos. 14-3611 & 15-1686                                      17

“adversarial notice” and “ex ante notice”). For this addition-
al reason, Irizarry doesn’t inform the analysis here.
    Finally, our dissenting colleagues warn that if vagueness
doctrine extends to the advisory Guidelines, then other
broad and open-ended provisions are vulnerable—for
example, the “sophisticated means” enhancement,
§ 2B1.1(b)(10); the “vulnerable victim” enhancement,
§ 3A1.1(b); the “abuse of trust” enhancement, § 3B1.3; and
even the foundational concept of “relevant conduct,”
§ 1B1.3, which applies to all crimes. See Dissent at pp. 27–28.
Johnson itself specially addressed this kind of objection and
rejected it. The Court explained at length that the vagueness
defect in the ACCA’s residual clause is not just its use of
indeterminate language; it’s that the clause uses indetermi-
nate language and must be applied categorically, without
regard to real-world facts. 

–58. The
Court could not have been clearer on this point: It’s the
combination of indeterminate language and categorical
application that makes the clause fatally vague. The same is
true of the residual clause in the Guidelines. But the categor-
ical feature is unique to the residual clause; other Guidelines
provisions are applied to actual facts on the ground. That
distinction makes a difference under Johnson.
    Simply put, after Peugh we can no longer say, as we did
in Tichenor, that because the Guidelines are “advice” rather
than “rules,” they are immune from challenge on vagueness
grounds. Because Tichenor has lost its analytical foundation,
we now overrule it. Applying Johnson, we hold that the
residual clause in § 4B1.2(a)(1) is unconstitutionally vague.
   With this holding, we join a growing consensus among
the circuits. See 

(applying Johnson to the residual
18                                            Nos. 14-3611 & 15-1686

clause in § 4B1.2(a)(2) and finding it unconstitutionally
vague); United States v. Madrid, 

, 1211 (10th Cir.
2015) (same); United States v. Taylor, 

, 933 (8th
Cir. 2015) (holding that Johnson applies to the career-offender
guideline but remanding for determination of the vagueness
question). Several other circuits have accepted the govern-
ment’s concession without further discussion or assumed
without deciding that Johnson applies to the career-offender
guideline. See United States v. Soto-Rivera, 

, 59 (1st
Cir. 2016) (accepting the government’s concession that the
§ 4B1.2(a)(2) residual clause is unconstitutionally vague
without deciding the issue); United States v. Maldonado, 636 F.
App’x 807, 810 & n.1 (2d. Cir. 2016) (assuming without
deciding that “the due process concerns that led Johnson to
invalidate the ACCA’s residual clause as void for vagueness
are equally applicable to the Sentencing Guidelines”); United
States v. Townsend, 638 F. App’x 172, 178 & n.14 (3d. Cir.
2015) (invalidating the Guidelines’ residual clause after
Johnson without extended discussion of whether the vague-
ness doctrine applies). One circuit has declined to apply
Johnson to the Guidelines. See United States v. Matchett,

, 1194–95 (11th Cir. 2015). 4
C. Remedy
    For both Hurlburt and Gillespie, the Johnson error pro-
duced a Guidelines range that was too high. That’s ordinari-
ly enough to satisfy the prejudice requirement of plain-error

4 As we’ve noted, the Supreme Court has granted certiorari in a case
from the Eleventh Circuit. Beckles v. United States, 616 F. App’x 415 (11th
Cir. 2015), cert. granted, 

(U.S. June 27, 2016)
(No. 15-8544). The Beckles panel followed Matchett, the Eleventh Circuit’s
precedent on this question.
Nos. 14-3611 & 15-1686                                        19

review. To establish that the error affected their substantial
rights, the defendants must show “a reasonable probability
that, but for the error, the outcome of the proceeding would
have been different.” 

,
(internal quotation marks omitted). “When a defendant is
sentenced under an incorrect Guidelines range[,] … the error
itself can, and most often will, be sufficient to show a reasona-
ble probability of a different outcome absent the error.” 

(emphasis added). This is because the Guidelines
“inform and instruct the district court’s determination of an
appropriate sentence. In the usual case, then, the systemic
function of the selected Guidelines range will affect the
sentence.” 

    Hurlburt’s 72-month sentence fell below the original
Guidelines range but is above the correctly calculated range
once the Johnson error is removed. The same is true of
Gillespie’s 84-month sentence. The defendants request full
remand for resentencing.
    The government argues for a limited remand akin to the
procedure we adopted in United States v. Paladino, 

(7th Cir. 2005), for Booker errors. Paladino fashioned a
“limited remand to permit the sentencing judge to determine
whether he would (if required to resentence) reimpose his
original sentence.” 

But we’ve generally rejected
the Paladino-style limited-remand procedure when the
sentencing error involves a miscalculation of the defendant’s
Guidelines range. See United States v. Williams, 

(7th Cir. 2014). “When a district court incorrectly calculates
the [G]uideline[s] range, we normally presume the improp-
erly calculated [G]uideline[s] range influenced the judge’s
choice of sentence, unless he says otherwise.” United States v.
20                                   Nos. 14-3611 & 15-1686

Adams, 

, 743 (7th Cir. 2014). Neither judge said
otherwise here.
  Accordingly, we VACATE the defendants’ sentences and
REMAND for resentencing.
Nos. 14‐3611 & 15‐1686                                                    21 

   HAMILTON,  Circuit  Judge,  joined  by  POSNER,  FLAUM,  and 
EASTERBROOK, Circuit Judges, dissenting. 
    By now the Supreme Court’s decision in Johnson v. United 
States,  135  S.  Ct.  2551  (2015),  is  familiar  to  all  in  the  federal 
criminal justice system. Johnson held that the “residual clause” 
in the Armed Career Criminal Act definition of a violent fel‐
ony, see 18 U.S.C. § 924(e)(2)(B), is unconstitutionally vague. 
The en banc majority now holds that the reasoning of Johnson 
extends to a similar residual clause in an advisory Sentencing 
Guideline  on  career  offenders,  holding  the  guideline  provi‐
sion  unconstitutionally  vague.  The  majority  also  overrules 
United  States  v.  Tichenor,  683  F.3d  358  (7th Cir.  2012), which 
held, correctly in my view, that the advisory Guidelines are 
not susceptible to vagueness challenges. 
    The majority’s holding is both premature and erroneous. 
There is no need for us to decide this now. There is already a 
circuit split,  and the Supreme Court is  likely to  rule on this 
question in the coming term. See United States v. Beckles, 616 
Fed. Appx. 415 (11th Cir. 2016), cert. granted, 136 S. Ct. 2510 
(2016),  which  should  be  argued  in  the  autumn  of  2016.  We 
should not hurry to send many cases back to district courts 
for re‐sentencings that may well prove unnecessary. I respect‐
fully dissent. 
    The best course at this point would be for this court simply 
to wait for the Supreme Court to decide the issue in Beckles. If 
we must reach the merits now, we should stick with Tichenor 
and agree with the Eleventh Circuit, which decided in United 
States v. Matchett, 802 F.3d 1185, 1193–96 (11th Cir. 2015), that 
the so‐called “residual clause” in the advisory Guideline def‐
inition of a crime of violence, see U.S.S.G. § 4B1.2(a)(2), is not 
unconstitutionally  vague  because  it  is  only  advisory.  Judge 
22                                          Nos. 14‐3611 & 15‐1686 

Pryor’s opinion in Matchett is careful and persuasive. Doctri‐
nal and practical considerations support that view. After all, 
how can non‐binding advice be unconstitutionally vague? 
    To  begin  with  the  doctrine,  the  residual  clauses  in  the 
Armed  Career  Criminal  Act  and  the  advisory  Sentencing 
Guidelines have identical language, but their legal effects dif‐
fer in a fundamental way. That difference should lead to dif‐
ferent answers on the issue of constitutional vagueness. The 
vague  definition  in  the  statute  led  directly  to  higher,  often 
much higher, mandatory minimum and maximum sentences. 
The most common effect of the Armed Career Criminal Act 
provision was to require a fifteen‐year mandatory minimum 
sentence for a felon in possession of a firearm when the statu‐
tory maximum was otherwise just ten years. See Johnson, 135 
S. Ct. at 2560 (“Invoking so shapeless a provision to condemn 
someone to prison for 15 years to life does not comport with 
the Constitution’s guarantee of due process.”). 
    In  contrast,  the  definition  in  the  advisory  Sentencing 
Guidelines  leads  to  no  direct  consequences  of  any  kind.  It 
simply gives the sentencing judge advice about an appropri‐
ate sentence. Unlike in statutory cases, the parties are free to 
argue that the Guidelines’ advice about the defendant’s crim‐
inal history is either too harsh or too lenient. The judge may 
accept the Guidelines’ advice or reject it. In fact, the law re‐
quires  the  judge  to  treat  the  advice  as  only  advice.  A  judge 
who presumes the Guidelines’ advice produces a reasonable 
sentence  commits  reversible  error.  Gall  v.  United  States,  552 
U.S. 38, 50 (2007); Rita v. United States, 551 U.S. 338, 351 (2007). 
    The doctrinal foundation of the majority opinion is incon‐
sistent with the overall sweep of Supreme Court decisions fol‐
lowing United States v. Booker, 543 U.S. 220 (2005), which held 
Nos. 14‐3611 & 15‐1686                                           23 

the Guidelines advisory as the remedy for the Sixth Amend‐
ment problems with mandatory sentencing rules that require 
judicial  fact‐finding.  Since  Booker,  the  Supreme  Court  has 
been  trying  to  maintain  a  delicate  balance,  recognizing  that 
the difference between “binding law” and “advice” depends 
on the different standards of appellate review. See Gall, 552 
U.S. at 50–51. 
    Since Booker, the Court has treated the Guidelines essen‐
tially as advice for almost all purposes, but as closer to bind‐
ing  law  for  just  one.  For  purposes  of  the  Sixth  Amendment 
rights to jury trial, to proof beyond a reasonable doubt, and to 
grand jury indictment, the Guidelines are now advice. Booker, 
543 U.S. at 245. For purposes of due process notice, they are 
advice. Irizarry v. United States, 553 U.S. 708, 714 (2008). For 
purposes of sentencing policy, they are also advice. Kimbrough 
v. United States, 552 U.S. 85, 110–11 (2007) (sentencing court 
could reject guideline advice on ratio of crack and powder co‐
caine sentences). 
    In one sense, though, the Court has treated the Guidelines 
as more law‐like. For purposes of the Ex Post Facto Clause, 
the Guidelines are closer to the binding law end of the spec‐
trum. Peugh v. United States, 569 U.S. —, 133 S. Ct. 2072 (2013). 
The analysis in Peugh was based on the persuasive effect the 
Guidelines have, including “anchoring” effects on sentencing 
judges. It is not easy to reconcile Peugh with the cases treating 
the  Guidelines  as  advisory,  particularly  since  constitutional 
doctrine allows completely unguided sentencing discretion, at 
least apart from capital cases. Perhaps Peugh is the first sign 
of a sea‐change in this area of the law, but given the extensive 
24                                                  Nos. 14‐3611 & 15‐1686 

case  law  treating  the  Guidelines  as  truly  advisory,  that  re‐
mains to be seen.1 
    If the Supreme Court extends the rationale of Peugh, as the 
majority does here, and embraces wholeheartedly the concept 
that the Guidelines are like laws, that result would be difficult 
to reconcile with the Booker remedy, which spared the Guide‐
lines from Sixth Amendment challenges by making them ad‐
visory. The delicate doctrinal balance the Court has tried to 
maintain  since  Booker  would  be  threatened  by  extending 
vagueness jurisprudence to the advisory Guidelines. 
   In Welch v. United States, 578 U.S. —, 136 S. Ct. 1257 (2016), 
the Supreme Court held that Johnson is retroactive, applicable 
on  collateral  review  of  federal  sentences  under  28  U.S.C. 
§ 2255. If Johnson is extended to the advisory Guidelines, the 
argument will be powerful for applying that new holding ret‐

                                                 
1 In extending Johnson to the Guidelines, the Sixth Circuit has tried to split 

hairs even more finely, saying that for one due process notice purpose—
“adversarial notice”—the Guidelines are advisory, as in Irizarry, while for 
a  supposedly  different  due  process  notice  purpose—“ex  ante  notice”—
they are closer to binding laws. United States v. Pawlak, 822 F.3d 902, 909–
10 (6th Cir. 2016). This distinction cannot withstand scrutiny. The hypoth‐
esis  of  “ex  ante  notice”  is  that  a  person  deciding  whether  to  commit  a 
crime is entitled to know what punishment the law prescribes. The answer 
is the statutory sentencing range, regardless of the Guidelines. The hypo‐
thetically rational candidate for prosecution should know already there is 
no guarantee of a guideline sentence. Under Booker, Gall, Kimbrough, and 
§ 3553(a), the sentencing court will be free to impose a non‐guideline sen‐
tence for many reasons. That is exactly the same knowledge imputed to a 
defendant  and  defense  counsel  on  the  eve  of  sentencing.  Given  that 
knowledge, the Court held in Irizarry, due process does not require further 
advance  notice  to  the  defense  about  reasons  why  the  sentencing  judge 
may be considering not imposing a guideline sentence. 
Nos. 14‐3611 & 15‐1686                                             25 

roactively. If that is the just result, so be it. But we should rec‐
ognize the likely consequences. Federal courts would need to 
revisit thousands of previous guideline sentences that relied 
on the residual clause in the career criminal guidelines. 
    And  what  would  be  the  point?  The  difference  between 
statutory mandates and advisory Guidelines means that the 
practical consequences of the vagueness holdings differ dra‐
matically. In the Armed Career Criminal Act cases affected by 
Johnson, many cases must result in lower sentences, and in vir‐
tually all cases, lighter sentences are reasonably likely on re‐
sentencing.  But  if  guideline  sentences  are  remanded  on  the 
theory that Johnson should apply to the advisory Guidelines, 
in every case that is remanded the district court will be free to 
impose  exactly  the  same  sentence  again.  In  fact,  the  district 
courts probably should do so. 
    To understand why, consider the intellectual gymnastics 
required by the “categorical” approach to recidivist enhance‐
ments.  Applying  that  approach,  courts  must  focus  on  ele‐
ments of a prior offense of conviction and must ignore what 
the  defendant  actually  did.  The  results  are  often  arbitrary. 
See, e.g., Mathis v. United States,  579 U.S. —, 136  S.  Ct. 2243 
(2016) (breadth of state burglary statute, which included bur‐
glarizing  vehicles,  meant  that  defendant’s  prior  convictions 
for  actually  burglarizing  occupied  houses  did  not  count  as 
convictions  of  “violent  felonies”).  Yet  both  before  and  after 
Booker, the Guidelines have allowed sentencing judges to ex‐
amine  what  the  defendant  actually  did.  See  18  U.S.C. 
§ 3553(a)(1) (requiring consideration of “history and charac‐
teristics  of  the  defendant”);  § 3661  (“No  limitation  shall  be 
placed on the information concerning the background, char‐
acter, and conduct of a person convicted of an offense which 
26                                                  Nos. 14‐3611 & 15‐1686 

a court of the United States may receive and consider for the 
purpose  of  imposing  an  appropriate  sentence.”);  U.S.S.G. 
§ 4A1.3 cmt. (encouraging departures where guideline crimi‐
nal history is over‐ or under‐representative); cf. United States 
v. Sonnenberg, 628 F.3d 361, 367–68 (7th Cir. 2010) (applying 
categorical approach  to vacate  career criminal enhancement 
but encouraging district judge on remand to consider defend‐
ant’s actual conduct in case of child sex abuse). 
    As a result, we should expect little gain in terms of fairness 
for  defendants  by  telling  sentencing  judges  (a)  they  cannot 
use the residual clause in the career offender Guideline, but 
(b) they remain free to consider available information about 
the defendant’s actual conduct in the earlier crime and to sen‐
tence  accordingly.  In  fact,  they  should  be  doing  so  already. 
The  judge’s  job  is  first  to  calculate  the  guideline  range  but 
then to exercise judgment and discretion under § 3553(a) in 
light of all the available information. 
    If  Johnson  is  extended  to  the  Guidelines,  and  if  a  judge 
were to reduce the sentence because of such a reversal, that 
decision might be evidence that the judge did not do his/her 
job at the initial sentencing. No facts would be different, only 
the advice that the judge was supposed to evaluate critically 
the first time around. Absent new material facts, a different 
sentence on remand would tend to show that the judge fol‐
lowed  the  Guidelines  too  mechanically,  perhaps  presuming 
they were reasonable without assessing the defendant’s indi‐
vidual history and characteristics and the particular circum‐
stances of the case.2 

                                                 
2 Federal prisoners who were sentenced as career offenders based on the 

residual clause before Booker, while the Guidelines were still considered 
Nos. 14‐3611 & 15‐1686                                                           27 

    In considering the consequences here, the scope of the ma‐
jority’s vagueness holding under the Guidelines will also be 
difficult to limit. As the Eleventh Circuit noted in Matchett, if 
we  extend  due  process  vagueness  doctrine  to  the  advisory 
Guidelines, many guideline sentences will be subject to chal‐
lenge. 802 F.3d at 1196. I do not see a principled way to avoid 
extending  such  a  holding  beyond  the  residual  clause  in  the 
crime of violence definition to other provisions in the Guide‐
lines that are at least as vague if not more so. 
     Consider,  for  example,  the  enhancements  for  “sophisti‐
cated means” in fraud crimes, § 2B1.1(b)(10); the “vulnerable 
victim” in § 3A1.1(b) (defined as someone who is “unusually 
vulnerable due to age, physical or mental condition, or who 
is otherwise particularly susceptible to the criminal conduct”—
which sounds a lot like a residual clause); aggravating roles 
in § 3B1.1 (based on “otherwise extensive” criminal activity); 
or “abuse of trust” under § 3B1.3. Many departure provisions 
in Part 5K of the Guidelines are quite vague, as is the provi‐
sion in § 4A1.3 for over‐ or under‐representative criminal his‐
tory. Even the fundamental concept of “relevant conduct” in 
§ 1B1.3 could easily be challenged as vague if we are worried 
about  whether  defendants  have  fair  notice  of  the  conse‐
quences  of  their  crimes.  Yet  this  pervasive  vagueness  in 
Guideline provisions is not a bug in the system. It is a feature. 
It is intended to provide sentencing judges with needed flex‐
ibility. 



                                                 
binding, might have a much stronger argument for extending Johnson to 
their  sentences.  These  cases  do  not  present  that  issue,  and  I  express  no 
views on it. 
28                                         Nos. 14‐3611 & 15‐1686 

    Perhaps one might draw a line between the residual clause 
and  every  other  provision  of  the  advisory  Guidelines,  by 
simply declaring that the result is limited to categorical deter‐
minations rather than application of vague standards to spe‐
cific facts. But it is difficult to see a principled basis for such a 
limited  rule,  particularly  since  § 3553(a)  already  calls  upon 
judges to take into account the real‐world facts of prior con‐
victions. The majority has opened the door to vagueness chal‐
lenges  to  any  advisory  Guidelines.  As  a  matter  of  broader 
constitutional  doctrine,  including  the  difference  between 
binding  and  advisory  Guidelines  that  is  essential  to  the 
Court’s  Sixth  Amendment  jurisprudence,  it  would  be  more 
sound to maintain instead the distinction between vague sen‐
tencing  advice  (permitted)  and  a  vague  sentencing  statute 
with mandatory consequences (not permitted). 
    After  all,  judges  can  find  vague  sentencing  advice  from 
many sources. Section 3553(a)(2) tells judges in a vague and 
contradictory  way  to  follow  several  conflicting  theories  of 
punishment at once, so that a sentence should reflect the seri‐
ousness of the offense, promote respect for the law, provide 
just  punishment, afford adequate deterrence to crimes, pro‐
tect the public from further crimes of the defendant, and reha‐
bilitate the defendant. Judges can find further vague or inde‐
terminate advice about sentencing in law review articles, phil‐
osophical reflections on crime and punishment, advice from 
probation  officers  and  law  clerks,  and  even  from  appellate 
opinions.  The  fact  that  some  of  the  advice  may  be  vague 
should not render the sentence unconstitutional. 
    I recognize that the Guidelines have a special, elevated sta‐
tus among those other available sources of advice, but they do 
Nos. 14‐3611 & 15‐1686                                         29 

remain advisory. And as we and later the Supreme Court con‐
sider the vagueness issue here, it is worth remembering that 
one simple remedy to a regime of somewhat vague advisory 
Guidelines would be to eliminate some or all of the advice and 
to leave sentencing judges to their own devices. The permis‐
sibility of such discretion has been consistent in all of the Su‐
preme  Court’s  recent  sentencing  decisions  under  the  Sixth 
Amendment,  from  Apprendi  v.  New  Jersey,  530  U.S.  466,  481 
(2000); id. at 544–45 (O’Connor, J., dissenting), through Booker, 
543 U.S. at 233, and Alleyne v. United States, 570 U.S. —, —, 133 
S. Ct. 2151, 2163 (2013). 
     Such unguided discretion would be the vaguest regime of 
all. Defendants would face even greater uncertainty about po‐
tential sentences and even greater risk of arbitrary variation 
in sentences. Yet that is all perfectly constitutional. Why not 
allow some vagueness in the Guidelines, whose advisory sta‐
tus is essential to avoid Sixth Amendment violations? 
    Another permissible remedy would be to impose manda‐
tory sentences by statute, denying judges any flexibility. But 
stripping sentencing judges of discretion leads to other unfor‐
tunate  results,  including  delegating  sentencing  decisions  to 
prosecutors’  charging  decisions.  Better  to  leave  the  Guide‐
lines as true guidelines, despite their vagueness and flexibil‐
ity.