Full Opinion (html_with_citations)

                                RECOMMENDED FOR PUBLICATION
                                Pursuant to Sixth Circuit I.O.P. 32.1(b)
                                       File Name: 21a0277p.06

                    UNITED STATES COURT OF APPEALS
                                  FOR THE SIXTH CIRCUIT



                                                             ┐
 UNITED STATES OF AMERICA,
                                                             │
                                    Plaintiff-Appellee,      │
                                                              >        No. 20-5905
                                                             │
        v.                                                   │
                                                             │
 RODNEY HYMES,                                               │
                                 Defendant-Appellant.        │
                                                             ┘

                          Appeal from the United States District Court
                     for the Eastern District of Tennessee of Chattanooga.
                    No. 1:17-cr-00043-1—Curtis L. Collier, District Judge.

                             Decided and Filed: December 3, 2021

             Before: SUTTON, Chief Judge; SILER and READLER, Circuit Judges.
                                   _________________

                                            COUNSEL

ON BRIEF: Jennifer Niles Coffin, FEDERAL DEFENDER SERVICES OF EASTERN
TENNESSEE, INC., Nashville, Tennessee, for Appellant. Brian Samuelson, UNITED STATES
ATTORNEY’S OFFICE, Knoxville, Tennessee, for Appellee.
                                      _________________

                                             OPINION
                                      _________________

       CHAD A. READLER, Circuit Judge. Rodney Hymes pleaded guilty to possessing crack
cocaine with the intent to distribute. The district court initially sentenced Hymes to 188 months
of imprisonment. But while Hymes’s appeal was pending, we decided United States v. Havis,
927 F.3d 382
 (6th Cir. 2019) (en banc) (per curiam), which cast doubt over the district court’s
reading of the Sentencing Guidelines in Hymes’s case. Accordingly, we vacated Hymes’s
sentence and remanded for resentencing in light of Havis. The district court then resentenced
 No. 20-5905                         United States v. Hymes                                Page 2


Hymes to a within-Guidelines sentence of 124 months of imprisonment. On appeal, Hymes
finds fault with numerous aspects of the district court’s sentencing analysis, including its failure
to defer to national sentencing data accumulated by the Sentencing Commission. But as the
district court was not required to even consider that data (let alone defer to it), and seeing no
other basis to disrupt Hymes’s sentence, we affirm.

                                        BACKGROUND

       Following an investigation into Rodney Hymes’s drug trafficking activities, Chattanooga
police officers confronted Hymes as he was sitting in his car in a grocery store parking lot.
Hymes attempted to escape by ramming a police car. Eventually, officers arrested Hymes and
recovered a wide swath of narcotics from his car. A search of Hymes’s home unearthed even
more drugs. In total, officers seized 43 grams of crack cocaine, among other drugs. A grand
jury indicted Hymes on five different federal drugs and weapons charges. In exchange for
dismissal of the remaining charges, Hymes pleaded guilty to one count of possession with the
intent to distribute crack cocaine in violation of
21 U.S.C. § 841
(a)(1) and (b)(1)(C).

       At Hymes’s sentencing, much of the focus was on his numerous prior state criminal
convictions. They included nearly a dozen motor vehicle offenses, a conviction for evading
arrest and theft under $500, another conviction for failing to appear, and two convictions for
attempting to possess cocaine for resale. Under then-existing circuit precedent, the latter two
convictions put Hymes squarely within the ambit of U.S.S.G. § 4B1.1’s career offender
provision, which automatically establishes a criminal history category of VI and increases the
base offense level for purposes of the Guidelines calculation for offenders with two or more prior
controlled substance convictions. See United States v. Evans,
699 F.3d 858
, 866–68 (6th Cir.
2012) (holding that a “controlled substance offense” includes attempt crimes).

       Hymes asserted that (1) his convictions for attempt crimes did not trigger the career
offender enhancement; and (2) in any event, the district court should vary downward given the
bulk of Hymes’s prior convictions were for traffic offenses. Applying then-existing circuit
precedent, the district court deemed Hymes a career offender under § 4B1.1, subjecting him to a
Guidelines range of 188 to 235 months in prison. Turning to Hymes’s variance argument, the
 No. 20-5905                         United States v. Hymes                                Page 3


district court initially concluded that it “need not” rule on the issue because § 4B1.1 would place
Hymes in the criminal history category of VI regardless of his traffic offenses. But after
Hymes’s attorney pressed the court to impose a sentence “well below the stated [G]uidelines” (as
calculated using the career offender provision), the district court concluded that Hymes had
engaged in “considerable” criminal activity—even discounting the driving offenses—to support
a within-Guidelines sentence. Presented with a Guidelines range of 188 to 235 months, the
district court sentenced Hymes to the very bottom of that range. Hymes appealed.

       While Hymes’s appeal was pending, we issued United States v. Havis. Havis abrogated
prior precedent to hold that attempt crimes did not qualify as controlled substance offenses under
§ 4B1.1.
927 F.3d at 387
. In light of Havis, the government and Hymes agreed that his prior
attempt convictions did not qualify as controlled substance offenses under § 4B1.1.
Accordingly, we granted the parties’ joint motion to remand the case for resentencing in light of
Havis. See United States v. Hymes, 18-6041 (6th Cir. Jan. 10, 2020) (order).

       With the career offender provision off the table, Hymes faced a reduced (but still
significant) Guidelines range of 110 to 137 months imprisonment. Hymes again argued that the
district court should vary downward for numerous reasons. Among them, Hymes reiterated that
his prior driving offenses artificially inflated his criminal history score. He also noted several
post-incarceration developments that, in his view, warranted a variance, including his behavior in
prison and the conditions of confinement for federal prisoners during the COVID-19 pandemic.

       At the resentencing hearing, the district court, after acknowledging Hymes’s right to
preserve his argument that his “criminal history was overstated,” limited its considerations to
variance arguments based on post-sentencing events, deeming any factual or legal conclusions
reached in the previous sentencing “settled and resolved.” And following a lengthy colloquy
regarding Hymes’s COVID-19-based argument, the district court rejected Hymes’s argument
due to concerns of creating unwarranted sentencing disparities. After hearing both Hymes’s and
the government’s arguments for what the appropriate sentence would be within the Guidelines
range, the district court imposed a sentence of 124 months’ imprisonment.
 No. 20-5905                          United States v. Hymes                              Page 4


                                           ANALYSIS

       On appeal, Hymes contends that his sentence was both procedurally and substantively
unreasonable.    See Gall v. United States,
552 U.S. 38, 51
 (2007).            A sentence may be
procedurally unreasonable if, for instance, the district court miscalculates the Guidelines range,
considers an impermissible factor during sentencing, or fails to adequately explain the chosen
sentence. United States v. Rayyan,
885 F.3d 436, 440
 (6th Cir. 2018). In contrast, a substantive
reasonableness challenge focuses on the length of the sentence itself, see United States v.
Clayton,
937 F.3d 630, 643
 (6th Cir. 2019), asking if the sentence is “too long (if a defendant
appeals) or too short (if the government appeals),” Rayyan,
885 F.3d at 442
.

       We generally review a claim of procedural or substantive unreasonableness under the
deferential abuse of discretion standard, meaning we will grant relief “when a ruling is based on
an error of law or a clearly erroneous finding of fact, or when the reviewing court is otherwise
left with the definite and firm conviction that the district court committed a clear error of
judgment.” United States v. Moon,
808 F.3d 1085, 1090
 (6th Cir. 2015) (quoting United States
v. Kerley,
784 F.3d 327, 347
 (6th Cir. 2015)). Where a defendant fails to preserve a procedural
reasonableness argument, the plain error standard further constrains our review. United States v.
Vonner,
516 F.3d 382
, 385–86 (6th Cir. 2008) (en banc). Under that “extremely deferential”
standard, we reverse only in exceptional circumstances to correct obvious errors that would
result in a miscarriage of justice. United States v. Donadeo,
910 F.3d 886, 893
 (6th Cir. 2018).
With this framework in mind, we turn to Hymes’s specific arguments for reversal.

       1. Hymes first argues that the district court made a procedural error by declining to vary
downward in light of the nature of his criminal history. In particular, Hymes emphasizes that his
criminal history score, which was a product of numerous driving offenses, overstated his risk for
reoffending.    Because Hymes preserved this argument below, we review for an abuse of
discretion. Moon,
808 F.3d at 1090
.

       Hymes is correct that a procedural error occurs when a district court wholly fails to
address a defendant’s nonfrivolous argument. See United States v. Wallace,
597 F.3d 794, 803
(6th Cir. 2010). The record, however, belies Hymes’s claim that the district court did so here. In
 No. 20-5905                         United States v. Hymes                                Page 5


Hymes’s first sentencing, the district court concluded that even “putting [Hymes’s driving
offenses] aside,” he still had engaged in “considerable” criminal activity.        These offenses
included two “relatively recent” convictions for attempting to possess cocaine for resale, which
suggested to the district court that Hymes was unable or unwilling to “live a law-abiding life.”
To the district court, Hymes’s drug crimes showed a risk of recidivism that aligned with the
spirit of the § 4B1.1 enhancement and supported a within-Guidelines sentence. At resentencing,
the district court relied on these observations about Hymes’s risk of reoffending, concluding they
were law of the case. In short, the district court did not ignore Hymes’s argument. And to the
extent Hymes also challenges that court’s views on the import of his attempt convictions, the
deferential abuse-of-discretion standard forbids us from simply second guessing the district
court’s reasoning. See United States v. Phinazee,
515 F.3d 511, 521
 (6th Cir. 2008).

       Hymes counters that the district court’s comments in the first sentencing in fact were
about whether the career offender provision applied. The record, however, does Hymes few
favors: the district court squarely rejected Hymes’s distinct argument for a downward variance.
In fact, Hymes’s counsel conceded during his resentencing that the district court previously had
“ruled on” Hymes’s criminal history argument separate and apart from the career offender issue
and “declined to grant a variance based on that” argument. To be sure, the district court
referenced the career-offender provision in discussing Hymes’s criminal history. But while
evaluating the distinct variance argument, the district court addressed the broader issue of
whether Hymes’s criminal convictions showed a risk of recidivism and concluded they did.
Given that the proposed Guidelines range when the career offender provision was at play did not
trouble the district court due to Hymes’s recent drug crimes, we fail to see why the district court
needed to restate its views during resentencing, when Hymes faced a more lenient Guidelines
range. This is especially so given that our remand in light of Havis was limited in nature, see
United States v. Johnson,
11 F.4th 529, 531
 (6th Cir. 2021), meaning that the district court did
not have license to begin anew, and instead could focus only on the “impact of Havis” on
Hymes’s original sentence, see id.; see also Pepper v. United States,
562 U.S. 476
, 506–07
(2011) (recognizing that only general remands for resentencing allow for de novo resentencing).
In short, the district court did not abuse its discretion in evaluating Hymes’s criminal history
argument.
 No. 20-5905                         United States v. Hymes                                Page 6


       2. Hymes’s second procedural unreasonableness argument is that the district court based
his sentence on improper speculation. Here, Hymes points to the district court’s comment during
resentencing “question[ing]” whether “males . . . age out” of committing crime. As Hymes did
not raise this argument to the district court, to prevail he needs to show that the district court’s
comment was so exceptional as to amount to plain error. Donadeo,
910 F.3d at 893
. Hymes
falls well short of the mark. Nothing in the record suggests that the district court based its
sentence on any speculative ground about the relationship between age and propensity to engage
in drug trafficking.    See United States v. Parrish,
915 F.3d 1043, 1048
 (6th Cir. 2019)
(recognizing that a sentence based on bare speculation amounts to a procedural error). At most,
the district court merely asked a question about age and recidivism during a free-flowing
colloquy with Hymes’s counsel. Perhaps the sentencing court was skeptical that age necessarily
eliminated the risk of future criminal activity, but the record does not show that the question was
the foundation for Hymes’s ultimate sentence.          And in any event, one stray comment at
sentencing cannot establish a reversible sentencing error, let alone a plain error. See United
States v. Cherry,
487 F.3d 366, 373
 (6th Cir. 2007).

       3. Next, Hymes argues that the district court failed to address Hymes’s post-sentencing
rehabilitation. While he did not raise this procedural reasonableness argument at the sentencing
hearing, Hymes did raise it in his briefing before the district court, so we review for an abuse of
discretion. Donadeo,
910 F.3d at 893
. Unlike Hymes’s prior two arguments, his assessment of
the record here is accurate; the district court did not discuss in any detail whether Hymes’s post-
sentencing conduct warranted a more lenient sentence. But Hymes faces a different problem this
time around: his arguments about post-sentencing rehabilitation are inapposite in the case of a
limited remand. See United States v. Richardson,
948 F.3d 733, 743
 (6th Cir. 2020). And even
if Hymes’s conduct in federal prison was fair game during a resentencing, a sentencing judge
need not “explicitly address every mitigation argument” a defendant makes, especially when the
defendant does not press the argument in detail. See United States v. Madden,
515 F.3d 601, 611
(6th Cir. 2008). Doubly so when the argument at issue is meritless. See United States v. Gale,
468 F.3d 929, 940
 (6th Cir. 2006).         That label aptly describes Hymes’s post-sentencing
rehabilitation argument, where he maintains that his sentence should be reduced because he had
“no write-ups” while incarcerated and participated in prison programming. Not only is that
 No. 20-5905                            United States v. Hymes                                Page 7


behavior expected of federal prisoners, but it also is not something a sentencing court should
typically have to grapple with expressly when considering a downward variance. See
id. at 939
(holding that it is “not unreasonable for the district court to place little weight” on the defendant
“simply doing ‘exactly what he has to do’” under the terms of his sentence); cf. Pepper,
562 U.S. at 490
 (holding that a sentencing court is not forbidden from considering a defendant’s extensive
post-sentencing rehabilitation at resentencing after a general remand).

          4. Fourth, Hymes argues that the district court erroneously constrained itself from
imposing a lower sentence. In particular, he argues that the district court—in rejecting his
argument for a downward variance due to the conditions of confinement resulting from the
COVID-19 pandemic—erred by suggesting that the issue was one that only Congress and the
Sentencing Commission could resolve. All parties agree that we review this argument for an
abuse of discretion.

          The district court did not abuse its discretion. For one, with a limited remand, the district
court only needed to determine how Havis affected its original sentence. See Johnson,
11 F.4th at 531
.     The district court had no license to consider post-sentencing developments.            See
Richardson,
948 F.3d at 743
. Even if it could consider such developments, Hymes misreads the
record. Far from “misunderstand[ing] its own authority,” see United States v. Kamper,
748 F.3d 728, 743
 (6th Cir. 2014), the district court instead acknowledged that it had “the authority to
grant” Hymes’s pandemic-based motion for a downward variance. And when the district court
made its discretionary determination under
18 U.S.C. § 3553
(a) to reject Hymes’s argument, it
voiced thoughtful concerns about (1) the sentencing disparities that could result given the
potential for differing views about the nature of the pandemic and (2) the need for criminal
defendants—like the rest of society—to bear the sacrifices of the pandemic. To the extent
Hymes takes issue with this reasoning, we are not left with a “definite and firm conviction” that
the district court’s well-reasoned views are erroneous.          Moon,
808 F.3d at 1090
 (citation
omitted). After all, § 3553(a)(6)’s consideration of unwarranted sentencing disparities affords
the district court discretion as to which disparities are unwarranted. See United States v. Bartlett,
567 F.3d 901
, 908–09 (7th Cir. 2009) (Easterbrook, J.) (“[T]he court is free to have its own
policy about which differences are ‘unwarranted.’”).
 No. 20-5905                            United States v. Hymes                            Page 8


       5(a). Finally, Hymes contends that his sentence was unreasonable because, in assessing
the “need to avoid unwarranted sentence disparities among [similarly situated] defendants,” the
district court “failed to consider national sentencing data,” which, in Hymes’s view, shows his
within-Guidelines sentence is abnormal. True enough, a district court’s failure to consider the
18 U.S.C. § 3553
(a) factors—including whether a sentence would create unwarranted
disparities—amounts to a procedurally unreasonable sentence. See Gall,
552 U.S. at 53
. But
nowhere have we required a district court to consult the Sentencing Commission’s collected data
before issuing a sentence. Rather, our focus has been on how the district court calculates a
defendant’s Guidelines range. For when a district court correctly does so, it has “necessarily
taken into account the need to avoid unwarranted sentence disparities, viewed nationally.”
United States v. Houston,
529 F.3d 743, 754
 (6th Cir. 2008). With respect to the sentencing data
collected by the Commission in particular, we observed in a footnote in United States v. Stock
that, notwithstanding various “limitations” with that data, it may be helpful as a “starting point
for district judges” in attempting to avoid an unwarranted sentence disparity among similarly
situated defendants.
685 F.3d 621
, 629 n.6 (6th Cir. 2012). In subsequent non-precedential
cases, we have “approved” of district courts using the Commission’s data during sentencing. See
United States v. Clemons,
757 F. App’x 413
, 419 (6th Cir. 2018); see also United States v.
Powell,
852 F. App’x 978
, 980 (6th Cir. 2021); United States v. Patrick,
821 F. App’x 593
, 595
(6th Cir. 2020). And we have even used hortatory language, citing Stock, to suggest that district
courts “should” look to Commission data when considering significant upward or downward
variances. See United States v. Perez-Rodriguez,
960 F.3d 748, 756
 (6th Cir. 2020); United
States v. Lightning,
835 F. App’x 38
, 43 (6th Cir. 2020); United States v. Boucher,
937 F.3d 702, 713
 (6th Cir. 2019). But we have never adopted the view that district courts must consider
national sentencing statistics, whether when entering a within-Guidelines sentence or one that
falls outside the Guidelines range. And to clarify any confusion on the point, we expressly reject
imposing such a requirement on district courts. See United States v. Barcus,
892 F.3d 228, 235
(6th Cir. 2018) (“[C]hanging the Guidelines to correspond to new empirical data is in the hands
of the Commission, not this court.”).

       Any other approach seemingly would elevate the Commission’s statistical data over the
text of the Guidelines themselves. And Hymes, if anyone, should appreciate the dangers that
 No. 20-5905                         United States v. Hymes                                Page 9


loose approach invites.    Recall that Hymes prevailed in his initial appeal due to an error
stemming from courts favoring interpretative guidance over the text of the Guidelines
themselves. Havis,
927 F.3d at 386
. If a district court cannot rely on Guidelines commentary to
add to the plain text of the Guidelines, a district court certainly cannot substitute unvarnished
Commission data in place of a properly calculated Guidelines range. It is the Commission,
remember, “guided by a professional staff with appropriate expertise,” whose “important” role it
is to make national sentencing determinations through the Guidelines, which are themselves
based on “empirical data and national experience.” See Kimbrough v. United States,
552 U.S. 85, 109
 (2007) (quoting United States v. Pruitt,
502 F.3d 1154, 1171
 (10th Cir. 2007)
(McConnell, J., concurring)). It is possible that, in some instances, a snapshot of Commission
data might suggest that a within-Guidelines sentence is out of step with national sentencing
trends. But between those two potential guiding lights, only the Guidelines themselves have
passed “through the gauntlets of congressional review or notice and comment.” Havis,
927 F.3d at 386
. Neither consistency, transparency, nor reliability are aided by effectively allowing
statistical data to override the plain terms of the Guidelines. In other words, the Guidelines, not
statistical reports, are “our barometer for promoting nationwide sentencing uniformity.” United
States v. Freeman,
992 F.3d 268, 289
 (Quattlebaum, J., dissenting), reh’g en banc granted,
847 F. App’x 186
 (4th Cir. 2021) (mem.).

       5(b). Nor do we see an abuse of discretion in the district court’s imposition of Hymes’s
particular sentence. We generally afford a district court—“[w]ith its front row seat at sentencing
proceedings”—“significant discretion in passing sentence,” meaning that a difference in opinion
is “not enough to justify undoing” the district court’s determination as to the length of its
sentence. United States v. Brown,
828 F. App’x 256
, 262 (6th Cir. 2020) (Readler, J., dissenting)
(citing Gall,
552 U.S. at 51
).     In two separate sentencing proceedings, the district court
thoughtfully explained why a sentence near the bottom of the Guidelines range was appropriate.
In particular, the district court, citing Hymes’s recent drug trafficking convictions, concluded
that Hymes was a chronic recidivist and that a sentence below the Guidelines recommendation
would be inappropriate.
 No. 20-5905                         United States v. Hymes                              Page 10


       Again emphasizing Commission statistical data, Hymes laments the district court’s
purported failure to fairly account for the risk of a sentence disparity under
18 U.S.C. § 3553
(a)(6). But § 3553(a)(6) ordinarily is “an improper vehicle” for challenging a within-
Guidelines sentence because the Guidelines, as already explained, address the statutory purpose
of combatting disparity. United States v. Volkman,
797 F.3d 377, 400
 (6th Cir. 2015). Nor, for
that matter, is avoiding unwarranted disparities the “summum bonum in sentencing.” Bartlett,
567 F.3d at 908–09; see also United States v. Reyes-Hernandez,
624 F.3d 405, 420
 (7th Cir.
2010) (“A district court should consider all relevant § 3553(a) factors collectively, not just what
is in § 3553(a)(6).”). The district court, to be sure, relied upon § 3553(a) factors other than the
need to avoid unwarranted sentence disparities. But that is perfectly appropriate because the
latter factor is simply one among many to consider when imposing a sentence. Kimbrough,
552 U.S. at 108
.

       And even on its own terms, Hymes’s argument is not convincing.               He reads the
Commission’s Interactive Data Analyzer database to show that crack offenders placed in
criminal history category VI in 2019 received a sentence within the applicable range just over a
quarter of the time. From that assessment, Hymes maintains that his within-Guidelines sentence
is too long relative to similarly situated federal prisoners. But the data relied upon by Hymes is
not as forceful as he contends. For starters, as the government observes, by including all
offenders with a criminal history score of VI, the statistics Hymes points to necessarily include
those with a career offender status that automatically results in such a score. When the Data
Analyzer discounts for such a status and removes other non-similarly situated offenders (such as
those who offer substantial assistance to the government), the results show that in the last half
decade, roughly half of defendants similarly situated to Hymes were sentenced within the
Guidelines range. See U.S. Sent’g Comm’n, Interactive Data Analyzer, Sentences Relative to
Guideline Range – Crack Cocaine – CHC VI (excluding Career Offender Status) (FY 2015-
2020), https://ida.ussc.gov/ (last accessed Dec. 2, 2021).

       What is more, Hymes’s argument fails to account for differences among offenders within
the same criminal history category. The mere existence of a colorable difference between
Hymes’s sentence and the national average for category-VI-crack defendants falls well short of
 No. 20-5905                         United States v. Hymes                                Page 11


showing an “unwarranted disparity” under § 3553(a)(6).            Even among category-VI-crack
defendants, there are vast differences in defendants—from drug quantity at issue to acceptance of
responsibility to offense conduct to the likelihood of rehabilitation. For instance, it seems
unlikely that many other category-VI-crack defendants received an enhancement for obstruction
of justice, as Hymes did as a result of his decision to ram his vehicle into a police cruiser during
his initial arrest. And a criminal history category is no monolith. Especially so, it seems, for
category VI, which presently includes any defendant with 13 or more criminal history points.
We therefore harbor serious skepticism that a defendant could ever establish an unwarranted
sentence disparity under § 3553(a)(6) merely by showing he received an above-average sentence
for a particular criminal history category and a particular drug. Such comparisons, as Judge
Quattlebaum vividly explained, do not “compare apples to apples”; they “compare[] an apple to
all fruits, and decr[y] that, because the apple looks different than oranges or bananas, something
must be wrong with it.” Freeman,
992 F.3d at 286
 (Quattlebaum, J., dissenting).

       All of this is to underscore that a district court that chooses to consult statistical data as
part of its sentencing analysis should measure the data with “extreme care.”
Id. at 289
. Article
III courts, keep in mind, are composed of lawyers, not statisticians. Here, Hymes’s central beef
with the district court ultimately is how it weighed the various § 3553 factors. Given the record
before us, we appropriately decline to second-guess that weighing process here. United States v.
Trevino,
7 F.4th 414, 433
 (6th Cir. 2021).

                                         CONCLUSION

       For the forgoing reasons, we affirm the judgment of the district court.