The ex post facto clause of the United States Constitution prohibits laws that "increase the punishment for a crime after its commission." Garner v. Jones, 529 U.S. 244, 249 (2000).*fn1 Brennan Sweeney pleaded guilty in May 2009 to possessing and distributing child pornography. The FBI raided his apartment in October 2003, found child pornography on his computers, interviewed him, obtained a confession (Sweeney admitted to possessing over one thousand images of children), seized the computers, and left him in liberty. Four years later, in March 2008, they arrested him. The government offers no explanation for the delay. In the interim, while Sweeney was free, the law concerning punishment of his crimes changed. Under the United States Sentencing Guidelines (the "Guidelines") in effect at the time he committed the crimes, his sentencing range would have been 27 to 33 months. U.S.S.G. § 2G2.2 (2002). But under the current version of the Guidelines, the applicable range is nearly three times higher- 78 to 97 months. The difference results from a five-level increase in the base offense level for distribution of child pornography, see U.S.S.G. § 2G2.2(a) (2009), and a five-level enhancement created in 2003 for offenses involving 600 or more illegal images. Id.§ 2G2.2(b)(7)(D).*fn2 Federal statute requires application of the current Guidelines. 18 U.S.C. § 3553(a)(4)(A)(ii). The issue is whether that statute, as applied to Sweeney, is an unconstitutional ex post facto law.
Five years ago, before the Supreme Court decided United States v. Booker, 543 U.S. 220 (2005), this question had a clear answer. Every Federal Circuit Court of Appeals had held that the ex post factoclause prohibited retrospective application of Guidelines amendments that increased the applicable sentencing range. See United States v. Seacott, 15 F.3d 1380, 1386 (7th Cir. 1994); United States v. Schnell, 982 F.2d 216, 218 (7th Cir. 1992) (collecting cases from every circuit). This unanimity flowed from a Supreme Court decision finding that Florida's sentencing guidelines, which were very similar to their federal counterparts, violated the ex post factoprohibition when applied retrospectively. See Miller v. Florida, 482 U.S. 423 (1987).*fn3
With Booker, some things changed and some things did not. The decision rendered the Guidelines advisory by striking down the portions of the Sentencing Reform Act that required courts to impose sentence within the applicable Guidelines range. Booker, 543 U.S. at 259. In their new advisory guise, however, the Guidelines have influenced sentencing determinations to nearly the same extent as before Booker. Gordon Mehler, et al., Federal Criminal Practice 665 (10th ed. 2010) ("[S]entence lengths and the degree of adherence to the 'advisory' Guidelines have remained essentially what they were when the Guidelines were mandatory."). Thus, while the Guidelines' formal strictures have changed radically, their effect has remained rather constant. The circuits have split on whether these developments change the ex post facto analysis. Compare United States v. Demaree, 459 F.3d 791, 795 (7th Cir. 2006) (Posner, J.) ("We conclude that the ex post facto clause should apply only to laws and regulations that bind rather than advise . . . .); with United States v. Turner, 548 F.3d 1094, 1099-1100 (D.C. Cir. 2008) ("[U]sing the [later version of the] Guidelines created a substantial risk that Turner's sentence was more severe, thus resulting in a violation of the Ex Post Facto Clause."). The Second Circuit has not decided the issue. See United States v. Gilmore, 599 F.3d 160, 166 n.4 (2d Cir. 2010) ("[T]he ex post facto consequences of relying on a later-enacted version of the Guidelines remains an open question.").*fn4
The Miller case-the foundation for the circuits' pre-Booker consensus-is the logical place to start an analysis of Booker's ex post facto implications. At the time the defendant in Miller committed his crime, the Florida guidelines prescribed a sentencing range of 3 1/2 to 4 1/2 years. By the time he was sentenced, however, changes to the guidelines had increased that range to 5 1/2 to 7 years. Miller, 482 U.S. at 424. The Court took a simple approach to the constitutional analysis. First, it explained that the purpose of theex post facto prohibitionis to guarantee "fair notice" of the punishment a crime will draw. Id. at 430 ("[A]lmost from the outset, we have recognized that central to the ex post facto prohibition is a concern for 'the lack of fair notice and governmental restraint when the legislature increases punishment beyond what was prescribed when the crime was consummated.'") (quoting Weaver v. Graham, 450 U.S. 24, 30 (1981)). Next, the Court applied a two-part test "derived" from the notice principle: "to fall within the ex post facto prohibition . . . the law 'must be retrospective, that is, it must apply to events occurring before its enactment;' and second, 'it must disadvantage the offender affected by it.'" Id. (quoting Weaver, 450 U.S. at 29.). The Florida guidelines clearly met both elements. As applied to Miller, they were retrospective, because the 5 1/2 to 7 year range did not apply to his crime at the time he committed it. Id. at 430-31. And the change plainly disadvantaged him: working from the new guidelines, the sentencing court had imposed a 7-year sentence-the top of the applicable range. Under the Florida system, sentences within the guidelines range were presumed reasonable and were not reviewable on appeal; non-guidelines sentences, on the other hand, required a written statement of "clear and convincing" reasons justifying departure. Thus, under the retrospective amendment, the 7-year sentence was presumptively reasonable and unreviewable, whereas under the guidelines in effect at the time of the crime, any sentence over 4 1/2 years would have required clear and convincing written justification. Id. at 432-33.
Judging by these features of the written law, the Court held that the retrospective amendment "substantially disadvantaged" Miller and therefore violated the ex post facto clause. Id.
This case is different than Miller in some respects. The federal Guidelines after Booker do not on their face demonstrate that a retrospective enhancement works to a defendant's "substantial disadvantage," at least not to the same extent as the Florida guidelines. To be sure, the federal Guidelines remain the "starting point and the initial benchmark" for a federal sentence, and district courts are required to calculate and consider the applicable sentencing range. Gall v. United States, 552 U.S. 38, 50 (2007). Nonetheless, the Guidelines are but one of many sentencing factors that must be considered, and district courts may not presume that a Guidelines sentence is reasonable. Id.; Nelson v. United States, 129 S.Ct. 890 (2009) (reversing sentence where the district court concluded "unless there's a good reason in the [statutory sentencing] factors . . . the Guidelines sentence is the reasonable sentence."). As the Second Circuit recently observed, "[i]t is . . . emphatically clear that the Guidelines are guidelines-that is, they are truly advisory . . . . District Judges are, as a result, generally free to impose sentences outside the recommended range." United States v. Cavera, 550 F.3d 180, 189 (2d Cir. 2008) (en banc). Facially, then, the Guidelines do not control sentencing results to the extent that the Florida guidelines at issue in Miller did.
As it turns out, however, "starting points and initial benchmarks" exert a gravitational pull that, though not legally prescribed, becomes obvious in application. At a negotiating table, the first offer frames the exchange, and in federal district court, sentences hew to the advisory Guidelines. This phenomenon is well-documented,*fn5 and the posture of this case bears further proof of its force. Under the current Guidelines, Sweeney's sentencing determination would start from a benchmark of 78-97 months. Under the 2002 version, the point of reference is 27-33 months. It is not controversial that these different benchmarks will, in all probability, produce different sentences; the government does not contest this proposition, and even those who opine that Booker cures the ex post factoproblem would seem to acknowledge its truth. See Demaree, 459 F.3d at 794 ("Most federal sentences . . . continue after Booker to be within the guidelines' sentencing ranges."). This case turns, therefore, on whether a retrospective law falls within the ex post factoprohibition if it "substantially disadvantages" a defendant-or poses a "significant risk of increased punishment," in the language of later Supreme Court cases*fn6 -in application, even if that effect does not flow ineluctably from the face of the law.
Two recent Supreme Court cases shed light on the issue. In California Dep't of Corr. v. Morales, 514 U.S. 499 (1995), and Garner v. Jones, 529 U.S. 244 (2000), the Court rejected ex post facto challenges to amendments to state parole rules that decreased the required frequency of parole hearings for certain inmates. The law at issue in Morales allowed the California parole board to hold hearings at intervals as wide as three years for inmates convicted of more than one crime involving loss of life, whereas the law at the time defendant committed his crime required a hearing every year. 514 U.S. at 503. Similarly, in Garner, the Georgia Parole Board issued rules lengthening the permissible intervals between hearings from 3 to 8 years for inmates serving life sentences. 529 U.S. at 247. In addressing the ex post factoissue in both cases, the Court reviewed the face of the parole rules and evidence of how the rules operated in application. See Morales, 514 U.S. at 510 n. 7 ("The ex post facto standard we apply today . . . looks to whether a given legislative change has the prohibited effect of increasing the measure of punishment") (emphasis added); id. at 510-11 (considering parole statistics). Garner embraced the "as applied" analysis with particular clarity: "When the rule does not by its own terms show a significant risk [of increased punishment], the respondent must demonstrate, by evidence drawn from the rule's practical implementation . . . that its retroactive application will result in a longer period of incarceration than under the earlier rule." 529 U.S. at 255. Applying this framework, the Garner court rejected the prisoner's ex post factochallenge because it found the "requisite risk [wa]s not . . . demonstrated on the record." Id. at 251. But the Court remanded the case for a determination of whether the defendant had been permitted sufficient discovery on the question of "whether the [parole] amendment . . . in its operation, created a significant risk of increased punishment." Id. at 257. Though there was no question that the language of the amendment itself did not evince such a "significant risk," the Court nonetheless sought to ensure that defendant had received the opportunity to prove the risk through evidence of how the rule operated in application. Id. Thus, even though the Garner court denied the constitutional challenge before it, it put to rest any notion that ex post factoreview should stop at the face of the law in question.
Most federal courts that have considered Booker's impact on the ex post facto analysis read Garner the same way. The D.C. Circuit adopted the Supreme Court's "as applied"approach in holding retrospective application of the advisory Guidelines unconstitutional. See United States v. Turner, 548 F.3d 1094, 1100 (D.C. Cir. 2008) ("The proper approach is . . . to conduct an 'as applied' constitutional analysis . . . .") (citing Garner, 529 U.S. at 251.). Numerous district courts, writing before and after Turner, have applied the same reasoning and reached the same conclusion. United States v. Kladek, 651 F. Supp. 2d 992, 994-97 (D. Minn. 2009); United States v. Doyle, 621 F. Supp. 2d 345, 351 (W.D. Va. 2009); United States v. Restrepo-Suares, 516 F. Supp. 2d 112, 116-17 (D.D.C. 2007); United States v. Safavian, 461 F. Supp. 2d 76, 82-83 (D.D.C. 2006); see also Thomas v. Yates, 637 F. Supp. 2d 837, 847 (E.D. Cal. 2009) ("Under Garner, ex post facto challenges to a facially neutral parole statute require an as-applied analysis."). And since Turner, no court outside the Seventh Circuit has held to the contrary, at least according to the Court's research.*fn7
The Seventh Circuit's decision in Demaree is the only major outlier. 459 F.3d at 795. That case espouses a different sort of ex post facto analysis, one that limits the scope of the constitutional prohibition to laws that evince a significant risk of increased punishment on their face. Interestingly, there was no dispute in Demaree that a retrospective Guidelines amendment had, in fact, increased the defendant's sentence. Id. at 792-93 ("The [sentencing] judge applied the 2004 guidelines . . . and sentenced Demaree to 30 months. But he added that if the 2000 guidelines were applicable to her case instead, he would have sentenced her to only 27 months."). The court held, however, that because an amendment increasing the Guidelines range no longer requires a concomitant sentencing increase, retrospective application of such an amendment does not contravene the ex post facto clause, despite the probability-manifest from the facts of the caseitself-that a Guidelines increase will nonetheless cause an increase in the eventual punishment. Id. at 795.*fn8
This Court finds Demaree unpersuasive for two reasons. First, in refusing to engage in anything beyond facial analysis, Demaree contravenes Supreme Court precedent. See Garner, 529 U.S. at 251-53 (examining parole rule's "practical implementation").*fn9 Second, by drawing a constitutionally dispositive, bright-line distinction between laws that "bind" and "advise," Demaree grafts an arbitrary boundary upon the ex post facto prohibition. The Supreme Court has pointedly eschewed bright-line distinctions in the ex post factocontext, see Morales, 514 U.S. at 509 ("[W]e have long held that the question of what legislative adjustments 'will be held to be of sufficient moment to transgress the constitutional prohibition' must be a matter of 'degree.' . . . We have previously declined to articulate a single 'formula' for identifying those legislative changes that have a sufficient effect on substantive crimes or punishments to fall within the constitutional prohibition . . . .") (emphasis in original) (citations omitted), and Demaree does not justify departure from that well-settled approach. AsSweeney's case demonstrates-perhaps more convincingly than Demaree's-even advisory laws may generate such clear and universally acknowledged sentencing consequences as to fall solidly within the constitutional concern for "fair warning." See Miller, 482 U.S. at 430. So it makes little sense to cast aside Sweeney's ex post factochallenge simply because the Guidelines are advisory in form.
The Court does not disagree with everything about Demaree; the opiniondoes make an apt observation about the scope of the ex post factoprohibition: "If any regulation traceable to Congress that disadvantages a criminal defendant is therefore an ex post facto law, even if it is purely advisory, the constitutional prohibition will be unmoored from both its purpose and the circumstances in which statutes and regulations have heretofore been deemed to be ex post facto laws." 459 F.3d at 794. Morales voiced a similar concern: "[C]ountless [legislative] changes might create some speculative, attenuated risk of affecting a prisoner's actual term of confinement . . . but that fact alone cannot end the matter for ex post facto purposes." 514 U.S. at 508-09. In essence, both Demaree and Morales identify a line-drawing problem: ex post facto jurisprudence must distinguish laws with only "speculative, attenuated" links to increased punishment-like "restrictions on the hours that prisoners may use the prison law library," id. at 508, or appropriations of "more money for federal prisons on the theory that prison crowding induces judges to give shorter prison sentences," Demaree, 459 F.3d at 794-from laws that create a truly "significant" risk of increased punishment. The question is how Courts should go about separating the "speculative" from the "significant." Morales embraced a test of degree, under which the punishment implications of a law are assessed without resort to rote distinctions between different categories of retrospective laws. 514 U.S. at 509. That approach, though indeterminate, allows courts to call fouls where they really exist-where, for example, an advisory law subjects an offender to punishment under a "benchmark" three times more severe than what existed at the time of the crime. Demaree's effort to resolve the same problem by creating a new distinction between binding and advisory laws is unsupported by the case law and incongruous with the purposes of the ex post facto prohibition.
The Court holds that application of the current Guidelines to Sweeney would violate the ex post facto clause. Accordingly, at Sweeney's May 27, 2010 sentencing hearing, the Court applied the 2002 Guidelines and, working from the applicable range of 27 to 33 months, imposed a 24-month sentence. That the Court ultimately issued a non-Guidelines sentence while also holding that the Guidelines fall within the scope of the ex post facto prohibition is undeniably ironic. The basis of the Court's holding, however, is not that sentencing courts are constrained by the applicable sentencing ranges (though, most often, courts do impose Guidelines sentences); rather, the reasoning is that the Guidelines range shapes the eventual sentence, whether the final number falls within the range or not. Here, for example, if the Court had started from a benchmark of 78 to 97 months, Sweeney's sentence likely would have been higher than 24 months. Similarly, the Demaree sentencing court stated that it would have departed upwards if it had applied the older, less onerous Guidelines, but it noted that the lower Guidelines range would nonetheless have resulted in a lower sentence. 459 ...