I am testifying next week before the U.S. Sentencing Commission on the topic of retroactivity for pending amendments to the sentencing guidelines dealing with crack cocaine. Amendments to the guidelines do not normally apply to defendants who have already been sentenced. If, however, the Commission chooses to make an amendment retroactive, then imprisoned defendants who were sentenced under the old guideline can apply for sentence modifications based on the amendment. When the Commission reduced crack sentences in 2007, the amendment was made retroactive. The current amendment package, however, presents some new wrinkles because it includes a more diverse mix of changes, some of which will actually increase crack sentences.
A first draft my witness statement appears after the jump. I would welcome suggestions.
I am grateful for the opportunity to address the Commission on the topic of retroactivity for the recently promulgated permanent amendments implementing the Fair Sentencing Act. The FSA brought welcome change to federal sentencing law for crack cocaine, particularly insofar as it replaced the old 100-1 ratio in the quantity thresholds for sentencing the crack and powder forms of cocaine. The pending amendments to the guidelines incorporate the new 18-1 ratio into the drug quantity table (“Part A”), add various new specific offense characteristics to § 2D1.1 of the guidelines (“Part B”), and eliminate a provision of § 2D 2.1 that treats some defendants convicted of simple possession of crack as if they had been convicted of a trafficking offense (“Part C”).
Whether a guidelines amendment should be made retroactive requires a complex balancing of interests. At the end of the day, though, I hope the Commission will be guided by a desire to do justice. While justice in sentencing can be an elusive concept, I believe that the goal is best expressed through the notion of proportionality—the idea that the punishment should fit the crime. Because there are strong arguments that pre-FSA crack sentencing was disproportionately harsh in the federal system, I would urge the Commission to give retroactive effect to Parts A and C. Part B, which includes a mixture of aggravating and mitigating sentencing factors, presents a closer question. If the Commission believes it inadvisable to make distinctions within Part B, then a variety of considerations, including substantial administrative burdens, may point against retroactivity. If distinctions are made, however, then there is a good argument in favor of retroactivity for the minimal role cap.
In the pages that follow, I will try to explain why I favor retroactivity for these provisions. I will conclude with a suggestion relating to the reform of federal drug sentencing more generally.
By way of background, I have studied and written about federal sentencing for more than fifteen years, including time spent as a federal judicial clerk, a litigator in private practice, and a faculty member at Marquette Law School, where I teach Criminal Law, Criminal Procedure, Sentencing, and related courses. I am the author of more forty scholarly articles, most of which relate to sentencing law and policy. I am also an editor of the Federal Sentencing Reporter. Of course, I offer my testimony for the Commission solely as a private individual, and not as a representative of Marquette or any other organization.
I. Retroactivity for Part A
In considering whether any amendment to Chapter Two should be made retroactive, I would suggest a two-step process. First, the Commission should determine whether the amended guideline does a better job than the previous version of achieving just deserts (proportionality). Second, assuming the amended guideline is better in this sense, the Commission should consider whether there are important countervailing considerations that justify keeping the existing, less proportionate sentences in place. Because proportionality is a central objective of the federal criminal-justice system—and arguably should be the overriding concern of any system that deserves the label of criminal justice—the Commission should not forego opportunities to ameliorate substantial numbers of significantly disproportionate sentences absent compelling reasons.
“Proportionality—the notion that the punishment should fit the crime—is inherently a concept tied to the penological goal of retribution.” Contemporary punishment theory particularly emphasizes ordinal, rather than cardinal, proportionality. Ordinal proportionality is oriented to system-wide practices and demands that relatively more serious offenses be punished with greater severity than less serious offenses. In determining relative offense severity, the key variables are intent, harm, dangerousness, and justification. When it comes to differentiating among drug trafficking offenses, harm and dangerousness seem the most important variables.
There seems a very strong case that Part A ameliorates an important instance of disproportionality in the guidelines. It is obvious that the amendment, and the underlying FSA, respond to longstanding assertions by the Commission and many other expert commentators that there is little or no meaningful difference in the harm or dangerousness associated with crack and powder cocaine offenses. Although the 2007 crack amendment helped to soften the disproportionality, the Commission observed even at the time that it “view[ed] the amendment only as a partial remedy to some of the problems associated with” the relative treatment of crack and powder. Consistent with this position, the Commission’s 2010 survey of federal judges found that the overwhelming majority of respondents (70 percent) still felt that the guidelines ranges for crack offenses were too high. Indeed, even the new 18-1 ratio, which seems to have been a result of political compromise with no particular basis in principle, is arguably too high. As originally introduced, the FSA would have equalized the treatment of crack and powder—which is precisely what is done in the great majority of states.
Although powder cocaine sentencing provides the most natural benchmark to assess the proportionality of crack sentencing, it is not the only benchmark that should be borne in mind. Serious violent crime is normally regarded as the most aggravated form of criminal conduct and is appropriately treated as the anchor of the severity scale for any sentencing system. It is an especially suitable benchmark for crack offenses to the extent that historical associations between crack distribution and street violence have driven crack sentencing policy. Here, too, we may have concerns regarding the proportionality of even the improved system implemented by the 2007 amendment. Consider a crack offender found responsible for 50-150 grams, a common quantity for federal defendants. This produces a base offense level of 30. Notably, this is considerably higher than the offense level of 18 for reckless homicide (§ 2A1.4(a)(1)(A)). If one takes the view that a primary reason to punish crack offenders is to respond the risk that crack distribution will result in serious physical injury, it seems odd to punish crack offenses more severely than other risky conduct that actually results in the death of another human being.
To be sure, some crack offenders may also be responsible for actual deaths, but there is an extensive network of specific offense characteristics and mandatory minimums to respond to cases involving actual injury or violence. The base offense level for a drug offense should not assume actual injury, but only a risk of injury. Viewed in that light, there should be significant proportionality concerns whenever a drug base offense level exceeds the offense levels for violent crimes involving actual death or serious bodily injury. And not only do crack offenses commonly produce a higher base offense level than reckless manslaughter, they do the same in comparison to voluntary manslaughter (base offense level 29, § 2A1.3) and aggravated assault—even when the assault results in permanent or life-threatening bodily injuries (offense level 21, § 2A2.2(a)(3)(C)).
All of this suggests that even the new drug-quantity table may treat crack offenses with disproportionate severity. But that, of course, is not the question of immediate interest. Rather, the question is whether the amendment moves in the direction of greater proportionality. The answer is almost certainly yes.
B. Countervailing Considerations
If we assume that Part A results in greater proportionality—and hence greater justice—there would have to be some compelling reason to justify withholding the amendment’s benefits from the thousands of imprisoned individuals who were sentenced under the previous, less proportionate regime. There are, I take it, chiefly two contenders: (1) the possibility that earlier-released defendants will commit more crimes than if they had been held to the full extent of their original sentences, and (2) the administrative burdens on the lawyers, judges, and others who must process thousands of requests for sentence modification.
Although these concerns cannot be dismissed as trivial, it does not seem that they are any more substantial in connection with the present amendment than they were as to the 2007 amendment. Unless there is some reason to believe that the 2007 retroactivity decision was mistaken, the Commission would do well to repeat that decision in 2011.
As to the risk of new crimes, the vast majority of defendants who could potentially seek sentence modification will be released at some time—the only question is when. As Professor Anne Morrison Piehl pointed out in her testimony regarding retroactivity for the 2007 crack amendment, the timing does matter insofar as recidivism risks are related to an offender’s age—the longer an offender is held, the older he will be at the time of release, and hence the less likely to offend again. However, Piehl estimated that “the impact of releasing younger inmates will impose minimal additional crime costs on society, especially after recognizing that recidivism rates are much higher for property offenses than for subsequent violent offenses.” I am not aware of any empirical research on whether Piehl’s estimation has proven correct, but it remains at least a plausible assessment on its face.
Whatever the recidivism risks, they should be evaluated in light of two additional considerations. First, authorizing retroactive application of Part A does not require the amendment to be applied in any given case. Indeed, the guidelines specifically authorize district judges to take public safety into account in deciding whether and to what extent to grant a sentence reduction. The Commission’s most recent analysis of sentence reduction requests in response to the 2007 amendment found that at least 206 such requests were denied in order to protect the public. Thus, district judges can be expected to take recidivism risks into account on a case-by-case basis if the Commission decides in favor of retroactivity. Moreover, the Commission could also consider providing nonbinding guidance to district judges to aid their assessment of recidivism risks; for instance, analysis could be conducted of the post-prison performance of the first wave of those released after the 2007 amendment in order to identify the most useful predictors of recidivism risk.
Second, important ethical considerations militate against prolonging the service of disproportionate sentences purely on the basis of recidivism concerns. In effect, this would constitute a system of preventive detention masked as punishment. Professors Paul Robinson and John Darley have voiced important objections to such conflations of penal incarceration and preventive detention:
The conditions of confinement under the criminal justice system are conditions of punishment. Yet, the justification for confinement under an incapacitation strategy is not punishment but prevention, akin to the system of preventive detention we use for those with infectious diseases or mental illness that is likely to lead to violent behavior. Systems of preventive detention are morally ambiguous, but certainly we are most comfortable with them when they involve detention conditions that are not punitive in nature, involve “treatment” efforts that attempt to remove the elements in the individual that cause the presumed dangerousness, and continually reassess the dangerousness of the individual who is incarcerated. This is not a good description of the workings of the prison system in the United States.
In light of such ethical considerations, the Commission should insist on truly compelling evidence of heightened danger to the public before precluding retroactivity on a categorical basis. Otherwise, it would be preferable to leave public-safety considerations to district judges to assess on a case-by-case basis, denying or reducing the size of sentence modifications only where there is strong, individualized evidence of risk.
In addition to recidivism risks, the second major objection to retroactivity lies in the administrative burdens of processing thousands of requests for sentence modifications. The burdens would not be insubstantial and could conceivably even impair the court system’s capacity to do justice in other cases. Still, it is hard to imagine any category of cases in which it is more important—to defendants, to their families, to the taxpaying public—to “get it right” than cases involving long terms of imprisonment.
The significance of the administrative burdens should be assessed in light of at least three additional considerations. First, the experience of implementing the 2007 amendment remains fresh in the minds of key personnel in most districts; retroactivity for the similarly structured Part A would likely be handled even more efficiently than the 2007 amendment because of the ability to replicate established procedures and avoid a steep learning curve. Second, fewer than half of the potential sentence reductions would result in a release date within two years, and more than forty percent would involve a release date in year four or later, which means that district judges would have a great deal of flexibility in prioritizing reduction requests and fitting them as conveniently as possible within court calendars. Finally, the Supreme Court’s recent decision in Dillon v. United States, 130 S. Ct. 2683 (2010), narrows the scope of the sentence modification proceeding by clarifying that the new sentence may not, in general, go below that amended guidelines range. The relatively modest scope of the decision to be made by district judges will lighten administrative burdens and facilitate agreements between the government and defendants so as to avoid formal, adversarial processes.
In sum, Part A represents a substantial advance in proportionality; public-safety concerns may more appropriately be addressed through case-by-case adjudication than through a categorical rejection of retroactivity; and the administrative burdens of retroactivity seem unlikely to exceed the burdens of the 2007 amendment.
II. Other Mitigating Changes: Parts B and C
In addition to Part A, other aspects of the new crack amendment package also offer potential reductions in sentence length. These mitigating amendments do not present quite so compelling a case for retroactivity because they are not responsive to the sort of longstanding, widely shared proportionality concerns that motivated Part A. However, retroactivity would still be appropriate for these additional mitigating changes.
First, the changes do reflect traditional proportionality-based distinctions made in drug sentencing. Deleting the cross-reference in § 2D2.1(b)(1) restores the common distinction between simple possession and trafficking offenses. This presumably reflects the view that trafficking in drugs is more harmful than personal use insofar as the offender is feeding addiction and self-destructive behavior in others. Similarly, the new mitigating role cap (§ 2D1.1(a)(5)) responds to the concern that a largely quantity-driven approach to drug sentencing can attribute an unfair proportion of the harm and danger of large-scale, organized drug trafficking to bit players in the organization who have little real control over the organization’s business. Likewise, a new “mitigating motive” specific offense characteristic (§ 2D1.1(b)(15)) provides an additional benefit to certain minimal participants who were motivated by an intimate or familial relationship or by threats or fear. This speaks to justification, which is a traditional consideration in the culpability calculus.
Of course, any mitigating amendment must be assessed against the backdrop of a generally quite tough quantity-driven approach to federal drug sentencing that, as noted above, can easily lead to longer terms for nonviolent drug offenders than for those who commit very serious violent crimes. In light of this general background disproportionality, nearly any new culpability-based ground for reducing crack sentences would at least arguably advance the cause of just punishment.
As for countervailing considerations, the most important distinction between Part A and the other mitigating amendments would be with respect to administrative burdens. Although the amendments restoring the possession/trafficking distinction and imposing the minimal role cap seem to require fairly straightforward offense-level recalculations, the mitigating motive amendment may require some extensive new fact-finding. Moreover, in some of the older cases that could potentially be revisited if the amendment were made retroactive, it might be quite difficult to perform the fact-finding in a reliable fashion, given fading memories, lost evidence, deceased or disabled witnesses, and the like. For these reasons, the Commission might be justified in distinguishing the mitigating motive amendment from the other mitigating amendments for purposes of retroactivity.
III. Aggravating Changes: Part B
In addition to the new mitigators, Part B of the pending amendment package introduces several new aggravating specific offense characteristics. In light of these provisions, it is possible that some crack defendants in the future will actually be assigned a higher offense level under the amendments than they would have received before. However, if the aggravating amendments were made retroactive, this would presumably be for the limited purpose of canceling out mitigating provisions of the amendment package and thereby disqualifying some defendants from sentence modification (§ 1B1.10(a)(2)(B)) and raising the floor for others (§ 1B1.10(b)(2)(A)).
Even for this limited purposes, there does not seem to be any good reason to make the aggravating amendments retroactive. In cases where the aggravators are present, the district judge may exercise his or her discretion to take them into account in deciding whether and to what extent to grant a sentence reduction. Under Booker and Kimbrough, the new aggravators will be merely advisory in future cases; they ought not to be given any greater weight in sentence modification proceedings. Although this will inevitably result in some disparity in the way the aggravators are weighed in sentence modification proceedings, there are always trade-offs in selecting any level of judicial discretion, and there is no apparent reason for selecting a different level in sentence-modification proceedings than in original sentencings.
The analysis might be different if the pending amendments were poised to reduce a set of sentences that already seemed to be on the low side. But quite the contrary is true—even the new 18-1 ratio is arguably on the high side. It seems unjust then to categorically rule out the 18-1 ratio for any class of otherwise-eligible crack defendants. Where aggravating circumstances really do make the 18-1 ranges inadequate in particular cases, district judges can exercise their discretion to reject or diminish the magnitude of the requested sentence reductions.
Likewise, the analysis might be different if there were some reason to think that sentencing judges were not already taking into account the “new” aggravators in selecting a sentence within the guidelines range or in choosing whether and by how much to deviate from the guidelines. But it seems more likely that judges have been weighing serious relevant conduct like threats of violence and bribery of law-enforcement officers all along. To deny the benefits of new mitigating amendments on the basis of these same considerations might amount to de facto double-counting of aggravators in some cases. Again, it is better to allow district judges—who are in the best position to know why they did what they did the first time around—to sort these matters out on a case-by-case basis.
It is especially important to retain the discretion of district judges in light of the complexity of the drug-sentencing guidelines. The new aggravators are layered on top of an intricate preexisting system of aggravators. As the sheer number of aggravators goes up, the risk increases that they will interact in complex and unforeseen ways to result in enormous sentences for defendants whose culpability is really not that great—the forest gets lost for the trees. The Commission has already anticipated some of the potential difficulties and attempted to address them through new application notes (e.g., in notes (3)(B), (27), and 29(B) to § 2D1.1), but it is probably overly optimistic to think that these difficulties exhaust the possibilities for unfairness. Of course, the ability to compensate for unforeseen interactions of specific offense characteristics is one of the most compelling justifications for having judicial sentencing discretion in the first place.
IV. Other Considerations in Drug-Sentencing Reform
The Commission has recently indicated an interest in undertaking a broader reconsideration of the drug-trafficking guidelines. Such a reconsideration could accomplish many useful changes in an extraordinarily important and controversial area of federal sentencing. I would like briefly to note one consideration that I believe the Commission would do well to bear in mind as part of any such reconsideration: federal-state sentencing disparities.
Federal-state sentencing disparities are often quite dramatic in the drug area, and it can be difficult to see principled reasons why some cases are prosecuted in federal court and other seemingly quite similar cases are prosecuted in state court. Such disparities, if seen as based on invidious discrimination or even simply random chance, hardly build respect for law, one of the central purposes of the federal guidelines system. For this reason, I would urge the Commission to undertake a study of major areas in which federal and state drug sentencing diverge and to seek to minimize the extent to which the guidelines contribute to such disparities. The Commission’s analysis of federal-state disparities in crack sentencing could provide a model. To be sure, the guidelines cannot hope to achieve conformity with fifty quite different state sentencing systems. Yet, it should be possible to identify ways in which the federal system is out of step with a substantial majority of states—as it was with the 100-1 ratio—and to move the federal system in the direction of the states’ center of gravity. Areas of concern may relate to the relative treatment of different types of drugs, specific offense characteristics, availability of alternatives to incarceration (e.g., drug treatment courts), and overall severity.
 Ewing v. California, 538 U.S. 11 (2003) (Scalia, J., concurring in judgment).
 See Andrew von Hirsch, Penal Theories, in The Handbook of Crime and Punishment 659, 668 (Michael Tonry ed. 1998).
 Michael M. O’Hear, Sentencing the Green-Collar Offender: Punishment, Culpability, and Environmental Crime, 95 J. Crim. L. & Criminology 133, 156-59 (2004).
 U.S. Sentencing Comm’n, Report to Congress: Cocaine and Federal Sentencing Policy 10 (2007).
 Results of Survey of United States District Judges, January 2010 Through March 2010, 23 Fed. Sent. Rep. 296, 304 (2011).
 S. 1789, 111th Cong. (2009) (as introduced on Oct, 15, 2009).
 U.S. Sentencing Comm’n, supra note 5, at 98.
 See id. at 25.
 The Supreme Court has recently noted the critical difference between offenses that result in loss of life and those that do not for purposes of constitutional proportionality analysis. See Graham v. Florida, 130 S. Ct. 2011, 2034 (2010) (“The Constitution prohibits the imposition of a life without parole sentence on a juvenile offender who did not commit homicide.”).
 See, e.g., 18 U.S.C. § 924(c)(1)(A)(iii) (imposing consecutive ten-year prison term for discharge of firearm during and in relation to drug trafficking offense); 21 U.S.C. § 841(b)(1) (imposing twenty-year mandatory minimum if death results from certain drug-trafficking offenses).
 Anne Morrison Piehl, Testimony Before the United States Sentencing Commission, Public Hearing on Retroactivity: Benefits and Costs of Retroactivity 3 (Nov. 13, 2007).
 U.S. Sentencing Guidelines Manual § 1B1.10, cmt. n.1(B)(ii) (2010).
 U.S. Sentencing Commission, Preliminary Crack Cocaine Retroactivity Data Report tbl. 9 (April 2011 data). I say “at least” because no reason was provided for an additional 443 denials, some of which may have been motivated in whole or in part by public-safety concerns, while another 238 were denied based on the § 3553(a) factors, which include “protect[ing] the public from further crimes of the defendant.” In addition to the outright denials, it seems likely that at least some sentence reductions were diminished in magnitude based on public-safety considerations.
 Paul H. Robinson & John M. Darley, The Utility of Desert, 91 Nw. U. L. Rev. 453, 467 (1997).
 Memorandum to Sentencing Commissioners from Offices of Research and Data and General Counsel on Retroactivity, May 20, 2011, at 30.
 O’Hear, supra note 4, at 158-59.
 Proposed Amendments to Sentencing Guidelines, 76 Fed Reg. 3193-02, 3196 (Jan. 19, 2011).
 See Michael M. O’Hear, National Uniformity/Local Uniformity: Reconsidering the Use of Departures to Reduce Federal-State Sentencing Disparities, 87 Iowa L. Rev. 721, 730-35 (2002).
 Michael M. O’Hear, The Original Intent of Uniformity in Federal Sentencing, 74 U. Cin. L. Rev. 749, 763-64, 772-73 (2006). See also 18 U.S.C. § 3553(a)(2)(A) (noting that sentences should “promote respect for the law”).
 U.S. Sentencing Comm’n, supra note 5, at 98-114.