Harvard Law Review Harvard Law Review Harvard Law Review

Criminal Law

United States v. Collins

Sixth Circuit Holds that Imposing a Significantly Below-Guidelines Sentence Informed by a Jury Poll Is Not Substantively Unreasonable.

Community safety is a bedrock justification for the existence of criminal law. Criminal law, the rationale goes, serves the dual purpose of protecting the community from suffering the results of future criminal acts1×1. See, e.g., Michael S. Moore, Placing Blame: A Theory of the Criminal Law 84–85 (2010) (noting traditional theories that justify criminal punishment because it can signal what is wrongful conduct and thus generally deter such conduct in society); Richard A. Posner, An Economic Theory of the Criminal Law, 85 Colum. L. Rev. 1193, 1214–15 (1985) (explaining that a substantive principle of criminal law is preventing crime). while punishing those who have already committed such acts.2×2. See, e.g., Sanford H. Kadish, Foreword: The Criminal Law and the Luck of the Draw, 84 J. Crim. L. & Criminology 679, 684 (1994) (“Convicting offenders serves to identify those who have shown themselves to threaten further breaches of the law, and punishing them constitutes a response to the threat they constitute.”). Understandably, the criminal legal system engages many proxies for the community in the process of adjudicating cases. Juries, for example, can serve as a barometer of the community’s sentiment on the correct level of punishment.3×3. See Ring v. Arizona, 536 U.S. 584, 615 (2002) (Breyer, J., concurring in the judgment) (“In principle, [jurors] are more attuned to ‘the community’s moral sensibility’ . . . .” (quoting Spaziano v. Florida, 468 U.S. 447, 481 (1984) (Stevens, J., concurring in part and dissenting in part))). Recently, in United States v. Collins,4×4. 828 F.3d 386 (6th Cir. 2016). the Sixth Circuit held as a matter of first impression that a federal sentencing judge’s consideration of a jury sentencing poll was a permissible part of determining a defendant’s sentence.5×5. Id. at 390. The court was correct to uphold the sentence on this ground. The jury poll provided pivotal insight into community sentiment regarding fair punishment, exposing a remarkable gap between the jury’s view of an appropriate sentence and the mandatory minimum sentence. However, the Sixth Circuit should have gone further and endorsed the practice. By polling juries, trial courts can leverage their unique institutional ability to measure community sentiment to inform the legislative process regarding mandatory minimums.

On February 10, 2013, a Canton, Ohio, police officer assigned to an FBI task force used peer-to-peer file-sharing software to download child pornography from Ryan Collins’s computer.6×6. See Brief of Plaintiff-Appellant/Cross-Appellee at 7, Collins, 828 F.3d 386 (Nos. 15-3263/3309). The officer subsequently obtained a search warrant for Collins’s residence, where he interviewed Collins.7×7. Id. at 8. According to the officer, Collins admitted to using peer-to-peer software to download and share child pornography.8×8. Id. After confiscating Collins’s computer, the FBI discovered nineteen videos and ninety-three pictures of child pornography.9×9. Collins, 828 F.3d at 388. On October 28, 2014, a jury found Collins guilty of the offenses of receiving and distributing child pornography under 18 U.S.C. § 2252(a)(2) and possessing child pornography under 18 U.S.C. § 2252A(a)(5)(B).10×10. Brief of Plaintiff-Appellant/Cross-Appellee, supra note 6, at 3.

Following the jury’s guilty verdict, Judge Gwin of the U.S. District Court for the Northern District of Ohio conducted a poll of the jury in which he instructed the jurors to “[s]tate what you believe an appropriate sentence is.”11×11. Collins, 828 F.3d at 388. At sentencing, the district judge reported that the average juror response was fourteen and a half months of incarceration and the median was eight months, though the responses ranged from zero to sixty months.12×12. Id. The jury’s average sentencing preference for Collins strayed drastically from each offense’s mandatory minimum — sixty months — and departed dramatically from Collins’s calculated sentence range under the Federal Sentencing Guidelines — 262 to 327 months.13×13. Id. The government repeatedly objected to the court’s consideration of the jury poll as a factor in determining Collins’s sentence.14×14. Id. Nonetheless, the district judge incorporated the poll into his sentencing decision and subsequently sentenced Collins to concurrent mandatory minimum terms of sixty months of incarceration.15×15. See Transcript of Proceedings Had Before the Honorable James Gwin at 26, United States v. Collins, No. 5:14CR279 (N.D. Ohio Feb. 10, 2015) [hereinafter Transcript of Sentencing]. On appeal, the government challenged the district judge’s downward variance from the Federal Guidelines guided by the jury poll.16×16. Collins, 828 F.3d at 388. The government also argued on appeal that the district judge did not properly consider deterrence at the sentencing phase and that the court’s downward variance of 202 months was not adequately supported as required by Gall v. United States, 552 U.S. 38 (2007). See Brief of Plaintiff-Appellant/Cross-Appellee, supra note 6, at 25–27.

The Sixth Circuit affirmed the sentence. Writing for a unanimous panel, Judge Guy17×17. Judge Guy was joined by Judges Batchelder and Cook. held that the district judge’s explicit use of a jury poll did not render Collins’s sentence substantively unreasonable.18×18. Collins, 828 F.3d at 390–91. In so holding, the court had to confront its inconclusive dicta from United States v. Martin19×19. 390 F. App’x 533 (6th Cir. 2010). — a case involving the same judge’s use of a jury poll — on “whether jury polling provided ‘meaningful data with which to assess the suitability of the applicable Guidelines.’”20×20. Collins, 828 F.3d at 388 (quoting Martin, 390 F. App’x at 538). Reviewing the district judge’s decision for an abuse of discretion, the court set forth several reasons for upholding Collins’s sentence. First, in response to the government’s contention that a jury poll “impermissibly conflates the distinct roles of judge and jury,” the court found crucial the fact that the jury was polled after it had reached a verdict.21×21. Id. at 389. Accordingly, the jury’s factfinding function was not compromised by pondering the sentencing consequences of a guilty verdict.22×22. Id. (citing Shannon v. United States, 512 U.S. 573 (1994), which observed that “[t]he principle that juries are not to consider the consequences of their verdicts is a reflection of the basic division of labor in our legal system between judge and jury,” id. at 579). Next, the court reasoned that “[f]ederal law provides nearly unfettered scope as to the sources from which a district judge may draw in determining a sentence.”23×23. Id. (citing 18 U.S.C. § 3661 (2012)).

Third, the court noted its precedent in United States v. Kamper24×24. 748 F.3d 728 (6th Cir. 2014). establishing that district judges are authorized to reject Federal Sentencing Guidelines on account of policy disagreements.25×25. Collins, 828 F.3d at 389. In Collins, the district judge’s policy objections were clear and longstanding. The Sixth Circuit noted that in an earlier law review article, the district judge criticized the U.S. Sentencing Commission, which issues the Guidelines, for its failure “to determine sentences that would accurately reflect community sentiment.”26×26. Id. at 390 (quoting James S. Gwin, Essay, Juror Sentiment on Just Punishment: Do the Federal Sentencing Guidelines Reflect Community Values?, 4 Harv. L. & Pol’y Rev. 173, 185 (2010)). Accordingly, to address that failure, the district judge considered the jury poll as a sentencing factor for Collins, though he did not treat it as controlling. The Sixth Circuit endorsed the idea that juries “can provide insight into the community’s view of the gravity of an offense,”27×27. Id. (citing Gwin, supra note 26, at 193–94). while a judge can “interpose[]” himself between the jurors’ sentencing preferences and the statutory sentencing factors.28×28. Id.

Finally, the court observed that the district judge faithfully carried out his sentencing function, as evidenced by his noting the applicable Guidelines range and enumerating the factors that compelled a lighter sentence. Such factors included Collins’s “lack of prior convictions, absence of alcohol or drug abuse, possession of a college degree, regular employment, close family ties, and financial responsibility.”29×29. Id. Contrary to the government’s argument, the district judge sufficiently considered deterrence and protection of the public — a requirement under 18 U.S.C. § 3553(a)(2) — when deciding against imposing a longer sentence.30×30. Id. 18 U.S.C. § 3553(a)(2)(B) (2012) instructs a sentencing court to consider a sentence that will “afford adequate deterrence to criminal conduct.” Judge Guy was satisfied with the district court’s reasoning for granting Collins a downward variance.31×31. Collins, 828 F.3d at 391. Thus, Collins’s sentence was not “substantively unreasonable.”32×32. Id.

Community sentiment is a factor intended to be considered in crafting the rules of sentencing, and jury polls — as measures of community sentiment — would be a valuable source of information for those who craft those rules.33×33. Some scholars, however, have condemned the use of juries at sentencing. See, e.g., Charles W. Webster, Jury Sentencing — Grab-Bag Justice, 14 Sw. L.J. 221, 230 (1960) (“Our first advance toward the goal of individualized sentencing should be to remove the power from the jury.”). In Collins, the district judge’s ability to depart from the Guidelines was constrained by a statutory mandatory minimum sentence that was over four times longer than what the average juror opined was an appropriate sentence. The Sixth Circuit was correct in holding that the use of a jury poll was permissible in determining a sentence. However, the court missed an important opportunity to endorse the practice of lower courts conducting such polls to guide sentencing. Trial courts have the institutional capacity to collect uniquely valuable information from juries about the appropriate length of sentencing given specific crimes. This information, especially when aggregated, can inform the legislative process that shapes sentencing laws, including mandatory minimums in child pornography cases.

Community sentiment regarding the appropriate length of incarceration is intended to be of importance in crafting the rules that govern sentencing.34×34. See Henry M. Hart, Jr., The Aims of the Criminal Law, 23 Law & Contemp. Probs. 401, 437 (1958) (arguing that judges at sentencing should consider “the overriding necessity of a sentence which, taken together with the judgment of conviction itself, adequately expresses the community’s view of the gravity of the defendant’s misconduct”). Congress recognized the importance of community sentiment when it instructed the Sentencing Commission to take “into account . . . the community view of the gravity of the offense”35×35. 28 U.S.C. § 994(c)(4) (2012). and “the current incidence of the offense in the community” in creating offense categories for the Federal Sentencing Guidelines.36×36. Id. § 994(c)(7). When a jury is representative of the defendant’s community,37×37. Scholars decry a lack of jury representativeness. Judge Gertner, for example, criticized one paradigmatic case in which prosecutors chose to bring a street crime case to federal court, rather than state court in Massachusetts, where the black defendants’ crimes were alleged to have taken place. In so doing, they reduced the available pool of black jurors from 20% to 7%. Nancy Gertner, 12 Angry Men (and Women) in Federal Court, 82 Chi.-Kent L. Rev. 613, 614 (2007). a jury poll is an ideal way to gauge such sentiment — jurors are thoroughly informed on the facts of the case from hearing both sides, and they are pooled from the surrounding geographic area, making them representatives of the local community. In fact, the district judge in Collins collected information on the jurors’ geographic, occupational, and socioeconomic backgrounds to confirm that the jurors indeed accurately represented the community,38×38. Transcript of Sentencing, supra note 15, at 16 (“[W]e looked at their zip codes, and we looked at the median household income for the zip codes . . . .”); see also id. at 31–32. allowing him to conclude that the Federal Sentencing Guidelines were “off the mark.”39×39. Id. at 26.

The jury poll brought into sharp contrast the Sentencing Commission’s determination of an appropriate minimum sentence for child pornography offenses and community sentiment around the appropriate sentence. Congress prescribed a five-year mandatory minimum sentence for trafficking, receiving, or distributing child pornography with the passage of the Prosecutorial Remedies and Other Tools to end the Exploitation of Children Today Act of 2003.40×40. Pub. L. No. 108-21, 117 Stat. 650 (codified as amended in scattered sections of the U.S. Code). Thus, the district judge’s downward variance confronted a floor of sixty months’ incarceration. So, while the Guidelines suggested a sentence over four times longer than the one arrived at in this case, it was also true that the ultimate sentence was again over four times longer than what the average juror considered appropriate. Actors in the criminal legal system (both practitioners and policymakers) would benefit from knowing if such disparities are common at criminal sentencing.41×41. Judge Bennett of the U.S. District Court for the Northern District of Iowa notes that when he would conduct jury polls, “every time — even here, in one of the most conservative parts of Iowa, . . . they would recommend a sentence way below the guidelines sentence.” Eli Hager, When Juries Help Judge, Marshall Project (Mar. 13, 2015, 12:17 PM), https://www.themarshallproject.org/2015/02/13/when-juries-help-judge [https://perma.cc/4CVQ-JDC4].

Through regular access to juries composed of members of the local community, trial courts can efficiently elicit context-specific community preferences. Such courts are “deeply intertwined with and embedded within the local criminal justice systems in which they reside” and “have the potential to gain a broader understanding of the very criminal justice systems that . . . they are called upon to oversee.”42×42. Andrew Manuel Crespo, Systemic Facts: Toward Institutional Awareness in Criminal Courts, 129 Harv. L. Rev. 2049, 2068 (2016). In a recent twelve-month period, district courts in the Sixth Circuit held 165 criminal jury trials.43×43. U.S. Courts, Table T-1. U.S. District Courts — Civil and Criminal Trials Completed, by District, During the 12-Month Period Ending September 30, 2015, at 2, http://www.uscourts.gov/file/19576/download [https://perma.cc/6Q5H-E582]. The aggregation of jury recommendations across even a fraction of these cases would provide informative estimates of systematic discrepancies between jury recommendations and imposed sentences. Professor Andrew Crespo has called on courts to collect such data, termed “systemic facts,” about the administration of the criminal legal system, noting that “contemporary criminal courts have at their disposal far more information about the systemic and institutional workings of their local justice systems than we — or they — have thus far come close to realizing.”44×44. Crespo, supra note 42, at 2066. The results of jury polls are one concrete example of such crucial facts that have significant valence for the administration of criminal law.

Given their capacity to collect such data, courts are uniquely suited to influence the legislative process.45×45. Cf. David S. Law, A Theory of Judicial Power and Judicial Review, 97 Geo. L.J. 723, 774 (2009) (explaining that in the context of judicial review, the courts convert “information supplied by litigants into clear signals about the constitutionality of government conduct”). This influencing process is a dynamic game of incomplete information.46×46. See James R. Rogers, Information and Judicial Review: A Signaling Game of Legislative-Judicial Interaction, 45 Am. J. Pol. Sci. 84, 88 (2001). That is, it is an evolutionary process that develops over time as the product of the actions of interdependent actors, the legislature and the judiciary, each with their own perceptions of the “true state of the world” and the effects of policy choices.47×47. Id. at 90; see also id. at 88–89. The most prominent example of this legislative-judicial interaction is the standard practice of judicial review. See id. at 85. The most important feature of this process is that, ex ante, the legislature has both imperfect knowledge of the true state of the world and imperfect ability to forecast the extent to which the laws it enacts will achieve the policy goals it desires.48×48. Id. at 84–85. By contrast, the judiciary has the ability to gather valuable post hoc information about the impacts of laws the legislature has enacted.49×49. Id. at 87–88. Therefore, the judiciary has an advantage over the legislature with regard to assessing the impacts of laws.

Judicial data collection for the purpose of informing legislation generates minimal separation of powers concerns50×50. Cf. Abner J. Mikva, A Reply to Judge Starr’s Observations, 1987 Duke L.J. 380, 382 (“It is true, of course, that if Congress has not done its job, the courts should not do it for them. Courts should not legislate; it is not our role.”). — courts acting as fact collectors are merely serving as means through which community sentiment is measured and conveyed. In other words, specific indications of community sentiment measured by the courts can be used to inform the legislature’s understanding of general community sentiment. Even if legislators act in accordance with what they believe is public opinion, they cannot know how such opinion might be modulated in the face of a specific criminal case. While members of the public could become more punitive,51×51. Cf. Nancy J. King & Rosevelt L. Noble, Felony Jury Sentencing in Practice: A Three-State Study, 57 Vand. L. Rev. 885, 923 (2004) (finding that in Virginia, average jury sentences in noncapital felony cases were “more severe than average sentences after bench trial or guilty plea”). research illustrates that when they are given a factual scenario — akin to the level of detail available to juries at trial — they often favor drastically lower sentencing than what is advised by the Guidelines or required by the mandatory minimum.52×52. See Adriaan Lanni, The Future of Community Justice, 40 Harv. C.R.-C.L. L. Rev. 359, 388 (2005) (“[W]hen given detailed descriptions of specific cases, studies show that respondents often suggest sentences that are more lenient than the mandatory minimum in their jurisdiction. This discrepancy appears to result from the lay tendency to assume that the typical fact pattern for a particular offense is far more serious than it actually is . . . .” (footnote omitted)).

Depending on the statistical patterns, data gathered by the courts through jury polls may also help finally convince Congress to lessen the punishments for child pornography offenses. The history of child pornography sentencing legislation illustrates that federal courts have been unsuccessful in convincing Congress to do so.53×53. See Carol S. Steiker, Lessons from Two Failures: Sentencing for Cocaine and Child Pornography Under the Federal Sentencing Guidelines in the United States, 76 Law & Contemp. Probs., no. 1, 2013, at 27, 37. With regard to mandatory minimums more broadly, the judiciary has pled with Congress for reform. See, e.g., United States v. Harris, 154 F.3d 1082, 1085 (9th Cir. 1998) (“We urge Congress to reconsider mandatory minimum sentences.”); Mandatory Minimum Sentencing Laws — The Issues: Hearing Before the Subcomm. on Crime, Terrorism, and Homeland Sec. of the H. Comm. on the Judiciary, 110th Cong. 2–3 (2007) (statement of Rep. Robert C. “Bobby” Scott, Chairman, Subcomm. on Crime, Terrorism, and Homeland Sec. of the H. Comm. on the Judiciary) (“The Federal Judicial Conference . . . has written the House Judiciary Committee over a dozen times in the last 10 years urging us not to adopt the mandatory minimum sentences . . . .”). In 1990, when the Sentencing Commission fashioned a new Guideline lowering the base offense level for transporting and receiving child pornography in response to a “pattern of judicial downward departures,” Congress passed legislation reversing the Commission’s downward adjustments.54×54. Steiker, supra note 53, at 38; see also id. at 38–39. The Commission established forty-three offense levels, increasing with the severity of the crime, under authority delegated by the Sentencing Reform Act of 1984. Eric P. Berlin, Comment, The Federal Sentencing Guidelines’ Failure to Eliminate Sentencing Disparity: Governmental Manipulations Before Arrest, 1993 Wis. L. Rev. 187, 190–93. In 1995, 1998, and 2003, Congress would again pass legislation directing the Commission to increase the penalty associated with a child pornography conviction.55×55. Steiker, supra note 53, at 39–40. In fact, in 2003, “for the first time in the history of the Guidelines, Congress . . . made direct amendments to the Guidelines.” Id. at 40. And despite the wave of courts of appeals undermining the child pornography Sentencing Guidelines,56×56. See id. at 42–43 (highlighting such cases in the First, Second, Third, and Ninth Circuits). the Guidelines remain excessively punitive, as seen in the facts from Collins. Accumulated jury polls would be especially potent vehicles for delivering the message that statutory punishments for child pornography convictions are excessive since juries are themselves composed of voters.57×57. See Andrew G. Deiss, Comment, Negotiating Justice: The Criminal Trial Jury in a Pluralist America, 3 U. Chi. L. Sch. Roundtable 323, 349 (1996) (“In the United States jury service has long been linked to the right to vote: from at least the late nineteenth century, state officials have used lists of registered voters to identify qualified jurors . . . .” (footnotes omitted)). Therefore, any poll of jurors is also a poll of voters. This fact links the outcome of the poll to both the personal interests of legislators seeking reelection and the public interest of legislators hoping to accurately represent the views of their constituents in the legislature.58×58. See William J. Stuntz, The Pathological Politics of Criminal Law, 100 Mich. L. Rev. 505, 529–30 (2001) (“[C]rime is one of those matters about which most voters care a great deal. . . . If there is any sphere in which politicians would have an incentive simply to please the majority of voters, it’s criminal law.”).

Opponents of the use of jury polls argue that conducting such polls could bias a jury’s judgment of guilt and displace the role of the judge at sentencing.59×59. See Hager, supra note 41. It is impossible to observe the effects of the implementation of jury polls as a regular practice ex ante. If jury polls are used by trial judges as a single factor among many, however, there seems to be no reason to think that the salience of such polls to jurors will be great enough to bias jurors against adjudicating guilt objectively. Similarly, because a jury poll is not controlling, sentencing judges’ depth of experience will retain overwhelming influence over sentencing.60×60. Cf. id. (quoting Judge Kane of the U.S. District Court for the District of Colorado as saying, “I can’t think of any responsibility a judge has greater than sentencing. It is a complex matter requiring a depth of experience and knowledge that no jury could have . . . .”). While community sentiment is far from the only concern that matters in crafting sentencing rules, it is a highly relevant input.

Had it encouraged lower courts to adopt the practice of conducting jury polls, the Sixth Circuit would have unlocked those courts’ potential to illuminate systemic facts. Trial courts are uniquely positioned to gather data on community sentiment around sentencing policies. If these courts were to systematically gather such data, they could empower other institutional actors to contribute meaningfully to the legislative process.61×61. See Crespo, supra note 42, at 2106 (arguing that “opening a court’s internal systemic facts to litigants, researchers, [and] other institutional actors” releases “the simple power of transparency”). Specifically, trial courts could make concrete the notion that sentencing laws for crimes involving child pornography are far more punitive than the average juror would recommend by aggregating data from a wide variety of cases. This data could induce legislators to update their beliefs about what punishments are appropriate — and, perhaps more importantly, their understanding of what punishments their constituents believe are appropriate — and could ultimately change sentencing laws. The Sixth Circuit therefore missed an important opportunity to further the social ideal that in a democracy, laws should be reflective of social values.