Retroactivity and Other Second Looks
The second in a series of essays on the 2024-25 U.S. Sentencing Commission guideline amendment year
If there’s one issue that most clearly divides the current U.S. Sentencing Commission, it’s whether to apply retroactively guideline amendments that lower a sentencing range. For three straight amendment years, commissioners have sharply and publicly disagreed with one another on whether even to have Commission staff prepare a retroactivity analysis of such amendments.
Last summer, in an essay on this Substack, I asked the question, is it time for the U.S. Sentencing Commission to issue a detailed, written, and reasoned opinion on when it applies guideline amendments retroactively? In that essay, I noted that the Commission has expressed no principles for the application of the factors listed in §1B1.10 of the Guidelines Manual for retroactivity and especially how those factors weigh against the interests of finality. “While §1B1.10 has been amended many times to address the mechanics and application issues of amendments that have been applied retroactively, the Commission has not addressed its decision-making on which amendments to apply retroactively.”
The Commission responded to this essay — and other relevant comments on retroactivity — by publishing in the Federal Register a Request for Comment on Criteria for Selecting Guideline Amendments Covered by §1B1.10.
The Background Commentary to §1B1.10 (Reduction in Term of Imprisonment as a Result of Amended Guideline Range (Policy Statement)) provides a non-exhaustive list of criteria the Commission typically considers in selecting the amendments to be included in §1B1.10(d) for retroactive application: “the purpose of the amendment, the magnitude of the change in the guideline range made by the amendment, and the difficulty of applying the amendment retroactively to determine an amended guideline range under subsection (b)(1).” USSG §1B1.10, comment. (backg’d). This non-exhaustive list of criteria has remained substantively unchanged since the Commission originally promulgated the policy statement at §1B1.10 in 1989.
Issues for Comment:
The Commission seeks comment on whether it should provide further guidance on how the existing criteria for determining whether an amendment should apply retroactively are applied. If so, what should that guidance be? Should it revise or expand the criteria? Are there additional criteria that the Commission should consider beyond those listed in the existing Background Commentary to §1B1.10? Are there identifiable sources that the Commission should consult that highlight retroactivity criteria relied upon by other legislative or rulemaking bodies?
If the Commission continues to list criteria relevant to determining whether an amendment should apply retroactively, should it adopt any bright-line rules? Is there a different approach that the Commission should consider for these purposes?
Professors Berman, Chanenson, and I think this is an important issue for the Commission to grapple with, and we submitted the following letter to the Commission two weeks ago in response to the request for comment.
- - -
April 14, 2025
The Honorable Carlton W. Reeves, Chair
United States Sentencing Commission
One Columbus Circle, NE
Suite 2-500, South Lobby
Washington, DC 20002-8002
Dear Judge Reeves:
This letter is in response to the Sentencing Commission’s Request for Comment on Criteria for Selecting Guideline Amendments Covered by §1B1.10, published in the Federal Register on December 30, 2024.[1] Section 1B1.10 (Reduction in Term of Imprisonment as a Result of Amended Guideline Range (Policy Statement)) currently provides a non-exhaustive list of criteria to be used to determine whether a guideline amendment will be included in subsection (d) of §1B1.10 and thus eligible for retroactive application. The Commission seeks comment on whether and how this list, and the guidance otherwise provided around selecting amendments for retroactive application, might be changed.[2]
We appreciate the opportunity to share our thoughts with you on this important issue. Consistent with congressional policies set forth in the Sentencing Reform Act, the First Step Act, and elsewhere, we view the retroactive application of guideline amendments as only one of many good reasons for enabling reconsideration of an otherwise final sentence, and our comments come from this perspective. Recent expansions of post-sentencing mechanisms for review and adjustment of sentences have confused judges, practitioners, defendants, and the public and also risk significant unwarranted disparities. We think it is critical for the Commission to undertake a comprehensive review of these mechanisms in order to propose reforms to make more coherent, transparent, and understandable federal sentencing modification procedures.
- - -
Before the enactment of the Sentencing Reform Act (SRA)[3] in 1984, parole was a fundamental component of federal sentencing law, policy, and practice. It was the mechanism, embodied in law, for reconsideration and reduction of imprisonment sentences imposed by federal district courts. It provided for regular second – and often third and fourth – looks at the length and necessity of the prison term imposed by a judge’s initial sentencing decision, after an offender had served a minimum portion of that prison sentence. Though parole decision-making has traditionally been focused on an offender’s rehabilitative progress and potential for safe reentry into the community, a host of other factors have often influenced parole board decisions.
In the 1970s, as crime rates were rising, academics, advocates, victims, practitioners, and politicians began to express serious concerns about the “truthfulness” of imposed prison sentences. Discretionary release systems, like parole, led to offenders serving only a portion of the announced prison sentence. The average in the federal system, pre-SRA, was around 50%. These systems were seen as undermining both trust and confidence in the criminal justice system, in part because of concerns about inappropriate factors like race and class leading to disparate release decisions. Analysis of prison rehabilitative programs then suggested that “nothing works” to reduce recidivism, which fostered concerns about the effectiveness of parole systems to help reduce crime. These and other forces led to the “truth-in-sentencing” movement, which had as its primary goal to ensure that imposed prison terms were mostly, if not entirely, served.
The SRA embodied the goals of the truth-in-sentencing movement by abolishing federal parole release and adding provisions to Title 18 of the United States Code to ensure that, in most cases, the amount of prison time served by an offender was close to the prison sentence imposed by the sentencing court. Section 3582 of Title 18, for example, makes clear and explicit that a sentence that includes a prison term constitutes a final judgment and cannot be modified except for specific, limited, and delineated circumstances.
The legislative history of the SRA shows that Members of Congress in 1984 were focused on eliminating what they saw as unwarranted disparities in the amount of time served by similarly situated offenders and also counterproductive sentencing uncertainties associated with the federal parole system.[4] Under the then-new law, offenders could only earn limited credits towards early release for good behavior in prison. Those credits could amount to no more than 15% of the imposed prison term and for most offenders was the only way to reduce the time served in prison. And yet, section 3582 still provided four express means for district judges to grant sentence reductions under specified circumstances: (1) for offenders who provided substantial assistance to authorities on motion by prosecutors; (2) for offenders who presented “extraordinary and compelling reasons” (originally) on motion by the Bureau of Prisons; (3) for certain elderly offenders who had served decades in prison; and (4) for offenders who were sentenced based on guideline ranges that had since been reduced and made retroactive.
In 1994, Congress reaffirmed its commitment to truth-in-sentencing when it enacted the Truth-in-Sentencing Incentive Grants program as part of the so-called Clinton Crime Bill. That program provided financial grants to states that adopted policies to ensure that offenders convicted of certain violent crimes served at least 85% of their sentence.[5] About half the states enacted truth-in-sentencing laws either before or as a result of the 1994 Crime Bill.[6]
In the last decade or so, Congress, the American Law Institute (two of us are members of ALI), many states, and the U.S. Sentencing Commission itself have expressed concerns about strict truth-in-sentencing and have embraced new mechanisms for review, reconsideration, and adjustment of imposed sentences, and in particular, especially long sentences. This policy shift has been the result of a new understanding on how prison programming can be effective to reduce recidivism, how incentives to participate in such programs can work to improve public safety at lower costs (both within and outside of prison), how time can lead both to penitence and to self-improvement and reform, and how the values and judgments around sentencing policy can change over time. In addition, the Supreme Court’s Eighth Amendment rulings limiting life without parole sentences (LWOP) for juvenile offenders has required or prompted many jurisdictions to provide new means to reconsider lengthy prison sentences imposed on younger offenders.
Perhaps most consequentially for the federal sentencing system, the First Step Act of 2018, signed into law by President Trump, created a system in which imprisoned offenders can earn time credits for participating in recidivism reduction programming or productive activities. These credits can significantly reduce the portion of an imposed prison sentence actually served in prison.[7] First Step Act credits are in addition to credits for good behavior in prison[8] and any reduction for participating in the Bureau of Prisons’ Residential Drug Treatment program.[9] By some estimates, these new credits can enable certain defendants to be transferred into home confinement after serving as little as half of the prison term announced by the district judge.
In addition, the First Step Act changed the procedures required for judges to be able to consider sentence reductions based on “extraordinary and compelling reasons.” Federal prisoners no longer must depend on the Bureau of Prisons to make a motion for such a reduction; these prisoners are now permitted to petition district courts for such a reduction directly.
Similarly, the American Law Institute, which developed the Model Penal Code and is the leading independent organization working to improve American law, adopted a second look sentencing policy in 2017. Its revision of the Model Penal Code’s sentencing provisions calls for all jurisdictions to “authorize a judicial panel or other judicial decisionmaker to hear and rule upon applications for modification of sentence” from all prisoners who have served 15 years of imprisonment. The commentary to the policy explains the multiple reasons ALI had for providing a means for lengthy prison sentences to be reviewed and potentially modified:
The passage of many years can call forward every dimension of a criminal sentence for possible reevaluation. On proportionality grounds, societal assessments of offense gravity and offender blameworthiness sometimes shift over the course of a generation or comparable periods . . . It would be an error of arrogance and ahistoricism to believe that the criminal codes and sentencing laws of our era have been perfected to reflect only timeless values . . .
On utilitarian premises, lengthy sentences may also fail to age gracefully. Advancements in empirical knowledge may demonstrate that sentences thought to be well founded in one era were in fact misconceived . . . For example, research into risk assessment methods over the last two decades has yielded significant (and largely unforeseen) improvements. Projecting this trend forward, an individualized prediction of recidivism risk made today may not be congruent with the best prediction science 20 years from now. Similarly, with ongoing research and investment, new and effective rehabilitative or reintegrative interventions may be discovered for long-term inmates who previously were thought resistant to change.
Several states have also adopted second-look mechanisms and others are actively considering adopting them. A prosecutor-initiated resentencing law was enacted in California in 2018, and similar laws have been enacted in Illinois, Maryland, Minnesota, Oregon, and Washington, and have been proposed or introduced in Georgia, Massachusetts, New York, Texas, and Utah, among other states. Of course, many states still maintain parole release and other discretionary release mechanisms.
And finally, the Commission itself has increasingly embraced second looks. It has applied many important guideline amendments retroactively in recent years, including several amendments that have impacted thousands of imprisoned persons. It has also made significant changes to what it considers “extraordinary and compelling reasons,” potentially justifying a second look and a sentence reduction under 18 U.S.C. § 3582(c).
- - -
As the Commission itself has documented, the changes in federal law and policy around the reexamination and reduction of otherwise final sentences that have occurred over the last decade or so, and especially the First Step Act reforms, have created confusion – for offenders, victims, judges, probation officers, and the public – around the portion of imposed federal sentences that will actually be served. The Commission has done a valiant job trying to explain on its website how the various mechanisms work. Nonetheless, predicting on the day of sentencing the amount of prison time likely to be served by the sentenced individual is quite challenging. This is understandable as the current mechanisms for reconsideration of sentences are the result of different policy determinations made over time by different Congresses, Administrations, and Commissions. For example, the Biden Administration enacted regulations for First Step Act credits that were quite different than the first Trump Administration. The various mechanisms for reconsideration of otherwise final sentences do not have a singular philosophical foundation nor are they applied identically by all federal judges.
The piecemeal approach to reform has not only led to confusion but also to the risk of a wide range of possible disparities. We urge the Commission to enhance its research focus to better asses the risks and reasons for possible unwarranted disparity in the application of post-sentencing imprisonment term reductions. The Commission’s retroactivity decisions, which have themselves arguably been haphazard, inconsistent, and thinly explained over the years, have contributed to confusion and potential disparity. The eligibility requirements for earning and then applying First Step Act credits have as well. The disparate application of compassionate release criteria by judges across the country is yet another factor.
We think rather than risking more confusion with any complicated new criteria for retroactivity, the Commission should first undertake a broad review of all the various existing second-look and sentence adjustment mechanisms and seek to develop a set of coherent policies that can better embody congressional and Commission values and our better understanding of what works to promote successful reentry. We think this review will be a great asset for just describing all of the existing sentence review mechanisms and their operation currently, and we would hope the Commission could then develop a set of legislative proposals to harmonize and improve the current system. The Commission’s work would be greatly advanced, and would contribute to nationwide reform efforts, by studying the various state second-look and compassionate release, including modern discretionary parole release, models as well as the federal experience with the First Step Act.
There is much to learn from the states and the ALI and other reform proposals. To that end, the Federal Sentencing Reporter is partnering with Stanford Law School to convene a small second look conference this fall, where we will gather academic and practitioner experts to review the various models and experiences. We would love for commissioners to join us there. We think this type of convening can be part of a comprehensive Commission review of all second-look mechanisms. It could help the Commission develop a set of coherent principles and policies for sentencing reductions in order to recommend sound legislative proposals and other reforms that may be needed.
We believe an improved, more coherent and coordinated federal post-sentencing reconsideration system can be developed. A reformed second-look system would seek to better and more clearly balance the various values at stake, including: sentencing honesty and fairness, appropriate finality in sentencing, victim interests, incentives for good behavior and self-improvement, recognition of changed circumstances, costs, and efficiency. It could make into a coherent whole a set of policies and procedures that may be contributing to haphazard and disparate results.
A reformed approach might, for example, mandate judicial reconsideration of all (or nearly all) sentences that include imprisonment terms of 10 years or more after the service of half of the imposed prison term. (Prior to the enactment of the Sentencing Reform Act, federal parole consideration occurred after one-third of the announced sentence and actual release occurred, on average, after about half the sentence was served.) That reconsideration would allow courts to review the First Step Act and good behavior credits earned by the offender, any changes to the guidelines made over the intervening years, any relevant court decisions or other changes in law, among many other considerations. It would give all offenders the incentive to participate in recidivism reducing programs and to make amends with any victims in their case. It would also provide greater predictability, at least of procedure, for those victims. It would make far more transparent and understandable both the sentence imposed and the process for its review and implementation and would be cost-effective and not a significant burden on the federal courts.
Under such a system, most guideline changes could and would be considered as one of a number of factors in the judicial reconsideration process, and formal and blanket retroactivity of guideline amendments would be reserved only for amendments to change patently unjust or unlawful guidelines. The Supreme Court recognized in Montgomery v. Louisiana, that “[a] conviction or sentence imposed in violation of a substantive rule is not just erroneous but contrary to law and, as a result, void. [Citations omitted.] It follows, as a general principle, that a court has no authority to leave in place a conviction or sentence that violates a substantive rule, regardless of whether the conviction or sentence became final before the rule was announced.” 577 U.S. 190, 203 (2016). When an unlawful or patently unjust guideline is changed, it would be wrong to leave in place sentences based on such a guideline.
Judges have expressed a similar view that blanket retroactivity for guideline amendments should be circumscribed along these lines. The Criminal Law Committee, for example, recognized that retroactive application of guideline amendments should be limited to instances “when an amendment would rectify an inequity.” Letter from the Honorable Edmond E. Chang Chair, Committee on Criminal Law of the Judicial Conference of the United States, to the Honorable Carleton W. Reeves, pp. 4-5, June 21, 2024. Under a reformed system along the lines we are suggesting, this would be all that would remain for guideline retroactivity determinations.
Similarly, under this kind of reform, “compassionate release” could be limited to significant health or family circumstances that amount to a kind of exercise in structured compassion rather than an opaque mechanism for reconsidering sentences generally based on more ubiquitous arguments revolving around such things as legal changes or offender rehabilitation. In addition, because all sorts of non-legal considerations are often the focus of compassionate release motions, decision-making regarding this form of “compassionate release” might be soundly allocated to a special (independent) body with expertise in prison administration, medicine, and social work, which could make more informed and consistent release decisions than district judges. Congress, through the First Step Act, remedied one problem with the operation of the statutory mechanism for compassionate release by allowing prisoners to file motions directly, but the Commission’s review and analysis of judicial decision-making on these motions might lead to recommendations for further reform of this process.
With these comments, we are just seeking to set forth a few tentative ideas regarding what comprehensive sentence reduction reform might look like in the federal system. Were the Commission to study and analyze the operation of existing methods in depth, the current experiences, good and bad, could and should further inform reform perspectives.
- - -
Retroactive application of guideline amendments is only one of many good reasons for sometimes reconsidering an otherwise final sentence. We think making more coherent, transparent, and understandable federal second-look mechanisms is an important goal for the future of federal sentencing. We believe the Commission should comprehensively review all the existing and proposed post-sentencing review and adjustment mechanisms, should encourage robust and public discussion among relevant policymakers and stakeholders about important sentencing reconsideration principles and practices, and ultimately propose and advance reforms to create that better second-look policy.
We hope these comments – and the fall conference – can help the Commission.
Sincerely,
/s/ Jonathan J. Wroblewski
Director, Semester in Washington Program
and Lecturer on Law
Harvard Law School
/s/ Douglas A. Berman
Newton D. Baker-Baker & Hostetler Chair in Law
Executive Director, Drug Enforcement and Policy Center
The Ohio State University Moritz School of Law
/s/ Steven L. Chanenson
Professor of Law
Faculty Director, David F. and Constance B. Girard DiCarlo
Center for Ethics, Integrity and Compliance
Villanova University Charles Widger School of Law
[1] U.S. Sent’g Comm’n, Sentencing Guidelines for the United States, 89 Fed. Reg. 106761 (December 30, 2024), https://www.govinfo.gov/content/pkg/FR-2024-12-30/pdf/2024-31278.pdf.
[2] Id.
[3] Sentencing Reform Act of 1984, Pub. L. 98-473, 98 Stat. 1837 (1984).
[4] S. Rep. No. 98-223 at 45 (1983) (“[S]entencing in the Federal courts is characterized by unwarranted disparity and by uncertainty about the length of time offenders will serve in prison.”).
[5] Pub. L. 103-322, 108 Stat. 1796 (1994).
[6] U.S. Gov’t Accountability Off., GAO/GGC-98-42, Truth in Sentencing Availability of Federal Grants Influenced Laws in Some States 6 (1998).
[7] Pub. L. 115-391, 122 Stat. 657 (2018).
[8] 18 U.S.C. § 3624.
[9] 18 U.S.C. § 3621.