Reimagining How the U.S. Sentencing Commission Makes Policy
Moving away from parochial advocacy and towards public interest deliberations
The Sentencing Reform Act of 1984 established the U.S. Sentencing Commission as an independent agency “in the judicial branch of the United States.” More than forty years later, the full implications of that judicial branch placement are still far from clear. But one thing we do know is that the Commission has always been led by a federal judge, and for most of its history, there has been a statutory mandate that no fewer than three of the seven voting commissioners “shall be Federal judges.”
There is a tendency in all of us to believe that whatever expertise we have in one field will translate naturally to others. Our mastery of something gives us a false sense of universal competence. Elon Musk believed it would be so easy to run the federal government because of his success starting and running an electric car company. But I keep checking my bank account for my DOGE dividend check he promised, and nothing ever arrives.
Judges, unfortunately, are not immune from this tendency. As Carol Leonnig and Aaron Davis well documented in their book, Injustice: How Politics and Fear Vanquished America’s Justice Department, Merrick Garland brought a judicial mindset with him to the Department of Justice (DOJ) in 2021, both in the way he ran the Attorney General’s office and in an overly cautious approach to the job he took to avoid the appearance of being political. The consequences of that judicial approach made history.
At the Commission, which has been chaired by seven different judges, a judicial mindset has been the dominant institutional culture, shaping not just who sits on the Commission but how it approaches its work. Judges are trained and experienced in deciding disputes between adversaries. In doing that, they receive written submissions from opposing parties, hold hearings at which those parties make their best arguments, and then retreat to chambers, perhaps to discuss the dispute with their clerks, but certainly to render a judgment of their own. It’s a tried-and-true process for resolving disputes of law and fact between two parties. It can be a very poor process for making sustainable sentencing policy.
The Commission’s overall professional diversity problem compounds this judicial orientation. We have written on the Substack before about Judge Marvin Frankel’s vision for a sentencing commission filled with “lawyers, judges, penologists, and criminologists . . . sociologists, psychologists, business people, artists, and, lastly for emphasis, former or present prison inmates.”
But the U.S. Sentencing Commission’s membership has never matched that vision; it has been dominated by judges and lawyers. Unlike most state sentencing commissions that typically have between one and two dozen members of varying backgrounds and experiences, the U.S. Sentencing Commission has only seven seats, and they’re almost always filled with just lawyers and judges.1
The most tragic illustration of what happens when a sentencing policy making body thinks and acts like a court came in the mid-1990s, when the Commission was confronting the important sentencing question of what to do about the 100-to-1 quantity ratio that once defined the crack and powder cocaine sentencing disparity. The Commission had unanimously concluded that the disparity could not be justified; that the racial consequences were unwarranted; and that the Guidelines could account for differences in aggravating factors associated with individual cases without a broad-brush sentencing scheme based solely on drug type and quantity. But there was disagreement about whether to eliminate the differential ratio entirely or merely reduce it.
The Commission structured its decision-making then as it has done for most of its history. It solicited written briefing from the various parties and held a public hearing resembling an oral argument, at which various interest groups made their most forceful case. Then, Chair Richard Conaboy, a federal judge from the Middle District of Pennsylvania, asked each commissioner to reflect in private on what they had heard and to then vote their conscience. If Congress disagreed, he said without any concern, it could reverse the Commission’s decision as appellate courts reverse lower court decisions. No big deal, he surmised. After virtually no deliberation, the Commission voted four to three in a public meeting to equalize the quantity-based penalties for crack and powder offenses. The vote came as much of a surprise to the commissioners as to the public. The Clinton Administration almost immediately came out against the Commission action. Congress later formally rejected it. A historic opportunity was lost. The Commission would soon lose its quorum and then all of its commissioners. A big deal after all.
- - -
The Sentencing Commission announced last month that it will undertake a comprehensive review of its Rules of Practice and Procedure — the document that in theory governs how the Commission conducts its business, engages the public, develops guideline amendments, and manages its relationships with Congress, the courts, and the Executive Branch. Those Rules were last amended in 2016, and before that in 2007. The last truly comprehensive rethinking dates to the Rules’ original adoption in 1997. This new review, therefore, represents not merely a housekeeping exercise but a genuine opportunity to ask fundamental questions about whether the Commission’s institutional design and policymaking processes are actually adequate to the consequential work the Commission performs.
Fundamental to the task is a recognition that the U.S. Sentencing Commission is not a court. It is a policymaking body. It makes sentencing policy that, at its core, like much policymaking, is a balancing of competing values — proportionality against deterrence, just punishment against rehabilitation, uniformity against the need for individualized justice, granularity against workability. The Commission’s processes have been designed in many ways to look like courtroom adjudication: written briefing from interested parties, public hearings with each witness setting forth its preferred policy alternative, no public deliberation, and then a vote. That architecture is not designed to produce the kind of shared and careful consideration that good sentencing policy requires.
What is missing from the policy development process held in public — structurally absent — is the public and reasoned weighing of competing interests against each other, the reasoning toward a solution that serves not any particular constituency but the public as a whole. Commissioners listen to sequential advocacy, not structured deliberation. As a member of the Commission for ten years, I found this the most excruciating part of being a commissioner. I wrote about in the context of retroactivity decision here —
The Commission receives not a balanced assessment of the relevant considerations but a series of maximally one-sided presentations, each designed to push it in a particular direction. This is the way lawyers are trained. It would have been extraordinary if good policy had consistently emerged from that process.
The Structural Limits of the Current Policy Development Process
Nowhere is this structural problem more apparent than in the Commission’s advisory group structure and, to a significant extent, its representation by the Department of Justice. The Commission maintains standing advisory groups organized around individual constituencies — practitioners, probation officers, tribal issues, victims. Each presents its preferred outcomes to the Commission. And each does so without ever being required to sit across the table from the other groups, hear their arguments, acknowledge the legitimacy of competing interests, or explain how its preferred outcome serves the public interest rather than merely its own.
The single-stakeholder advisory group model is perfectly designed to produce advocacy. It is almost perfectly designed to prevent deliberation. Members don’t hear each other’s arguments until those arguments have been cemented into and presented as written comments or oral hearing testimony. They are never required to confront the costs their preferred positions impose on other constituencies. They are never asked to prioritize, to trade off, to explain how their preferred outcome fully serves the public interest. The result is that the Commission receives a series of maximally opposed position statements and is left entirely on its own to bridge the gap.
What is structurally absent from the Commission’s public policy development process is any mechanism for weighing competing interests against each other — for reasoning toward a solution that serves the public as a whole.
When I served as the Attorney General’s representative to the Commission, we were often part of the problem. I found the most valuable part of the role was its intelligence-gathering function. The Department of Justice encompasses United States Attorney’s offices in ninety-four districts, each with its own caseload, its own prosecutorial culture, and its own ground-level experience of how the Guidelines work and where they fall short. Surveying this vast operational enterprise, as we did about this time each year — asking line prosecutors, supervisors, and also agency officials what was working, what was not, and what they would change — consistently produced insights that no amount of Washington-based consideration could replicate.
The frustrating part came when proposed amendments were published for comment. What happened — at DOJ and across Commission stakeholder groups, with a regularity that came to feel almost mechanical — was for every stakeholder to argue for the outcome that best served its own institutional interests. DOJ argued mostly for positions that favored prosecutorial flexibility and punishment. Victims’ organizations argued for the interests of victims. Defense organizations argued for decreased punishment. Each side marshaled whatever data supported its preferred position, ignored what did not, and presented its case with complete confidence that its preferred outcome was also the just one. Nobody in the room was required to fully confront the competing interests against each other. The task was not structured to find the answer that best served the American people.
- - -
As the Commission takes on its review of its rules and procedures, it need not start from scratch. It is blessed with a remarkable body of knowledge about how rulemaking agencies can do better. This knowledge base includes much information on how the Commission can engage more diverse constituencies, ground policy in stronger empirical foundations, conduct more genuine deliberation, and produce decisions that are more legitimate and more durable.
The Administrative Conference of the United States (ACUS) has produced a library of recommendations on effective rulemaking practice, covering everything from advanced notice procedures to the use of negotiated rulemaking, from digital engagement platforms to plain-language communication. State sentencing commissions — particularly those in Minnesota, North Carolina, Washington State, and Pennsylvania — have developed sophisticated practices for empirical guideline development, racial equity monitoring, post-implementation evaluation, and structured public engagement that the federal Commission has never systematically examined. Federal agencies like the Consumer Financial Protection Bureau, the Environmental Protection Agency, and the Federal Trade Commission have developed stakeholder engagement models and comment-response practices that go far beyond what the Commission currently utilizes. The Model Penal Code’s sentencing provisions (MPC:S), approved by the American Law Institute in 2017, represent the most comprehensive scholarly synthesis of best practices in sentencing commission governance ever produced. Interestingly, it includes a directive that “Commission members should be selected for their wisdom, knowledge, and experience and their ability to adopt a systemwide policymaking orientation. Members should not function as advocates of discrete segments of the criminal-justice system.”
None of this knowledge has been systematically brought to bear on the Commission’s institutional design and its policymaking processes. The Commission should begin its rules review by conducting a formal study — perhaps in partnership with ACUS or with a leading academic institution — that catalogs these best practices, assesses their applicability to the federal sentencing context, and identifies the specific reforms most likely to improve the quality and legitimacy of the Commission’s policymaking. To proceed with rules revision without this foundation would be to navigate without a map that is readily available.
- - -
When it does this review, the Commission will discover that one best practice that it should consider utilizing is negotiated rulemaking. The Negotiated Rulemaking Act of 1990 provides a framework, regularly used by other federal regulatory agencies, in which all significantly affected interests are convened at the same table, under the guidance of a trained neutral facilitator, to negotiate the content of a proposed rule by consensus. The core insight is that bringing parties together — requiring them to hear each other’s arguments, to acknowledge the legitimacy of competing interests, and to understand that the alternative to agreement is a rule they have no hand in shaping — transforms the dynamic from advocacy to problem-solving.
Applied to the Commission’s processes, the model might work like this. The existing single-stakeholder advisory groups — and any new groups added to fill current representational gaps — could continue to serve a vital function: fact-finding. They would gather and transmit to the Commission the knowledge, experience, and practical wisdom of their constituencies. They would tell the Commission what the Guidelines look like from the ground. That is genuinely valuable work, and those groups should likely be preserved and strengthened for that purpose.
But the deliberative function — the work of weighing those inputs against each other and reasoning toward balanced policy recommendations — would be taken up by a separate, structured process: a negotiated rulemaking panel, convened for each significant proposed amendment, comprising representatives of all relevant constituencies, working under the guidance of a professional facilitator, with a mandate to produce either a consensus recommendation or a structured report documenting the points of agreement, disagreement, and the tradeoffs at stake.
The facilitator’s role is critical and often underappreciated. A skilled facilitator does not merely manage process, she helps parties move from positional bargaining to interest-based negotiation, distinguishing what each party is demanding from what each party actually cares about and why. She creates conditions in which stakeholders can acknowledge privately held uncertainties, explore creative options that no party would propose unilaterally, and build the trust necessary for genuine compromise. She holds participants accountable — to each other and to the public — for the quality of their reasoning.
The goal of this process is not consensus for its own sake. It is to find, through structured deliberation among those who best understand the relevant considerations, the public interest compromise that best serves the American people — not the solution that best serves victims’ organizations, not the solution that best serves DOJ, not the solution that best serves the defense bar, but the solution that most faithfully reflects what justice actually requires when all of those legitimate interests are weighed honestly against each other.
- - -
There is already a model in the sentencing field that captures the spirit of this approach. The Villanova Sentencing Workshop (VSW), developed by Professor Steven Chanenson, now in its twenty-fourth year, is described in Steve’s essay here —
The VSW, like the Yale, Minnesota, and other sentencing workshops, brings together sitting judges, prosecutors, public defenders, parole board members, corrections experts, and law students to work through real sentencing cases in an immersive, multi-day setting. The key to its success is precisely what the Commission’s advisory process lacks: participants from across the system, with genuinely different perspectives and institutional roles, are required to engage directly with each other’s reasoning — not to present their positions to a passive audience, but to deliberate together. As one federal judge who participated in a predecessor workshop reflected, issues that had seemed simple suddenly revealed many complex layers, and the conversations proved more sophisticated and nuanced than any she had encountered in her own courtroom. The VSW has already served as a policy sandbox for the Pennsylvania Commission on Sentencing, allowing it to test drive ideas in a confidential deliberative setting before they are ready for formal consideration. The federal Commission could develop analogous structured deliberative processes — not as educational exercises, but as genuine inputs to its policymaking.
The goal of all of this is to find, through structured deliberation among those who best understand the relevant considerations, the public interest compromise that best serves the American people.
The Stakes Are High; the Moment is Now
The stakes are high. The Commission’s Sentencing Guidelines govern the fate of approximately sixty thousand federal defendants each year. The decisions embedded in those Guidelines — about the relative severity of punishment for different offenses, the weight given to criminal history, the proper balance of sentencing purposes, and much more — are consequential policy choices that impact, directly or indirectly, all of us. They determine who goes to prison and for how long, who returns to their families and communities, and how the vast coercive power of the federal government is distributed across the American population in service of public safety. And yet the processes by which those decisions are made remain, in important respects, remarkably thin.
The Commission should approach this review with both humility and ambition: humility because the problems with the current model are deep and structural, and ambition because there is an extraordinary wealth of experience — from other federal regulatory agencies, from the Administrative Conference of the United States, from state sentencing commissions, and from the academic literature on deliberative governance — that the Commission has never systematically drawn upon. As mentioned, the first step is to develop a comprehensive report cataloging best practices and the hard-won lessons before drafting or amending a single rule.
The Commission’s rules review is an opportunity that should not be wasted on modest adjustments to notice periods and comment deadlines. The deeper questions — about who has voice in the Commission’s deliberations, about how competing interests are weighed, about whether the Commission’s processes are designed to produce deliberation or merely advocacy — are questions that the current rules have never adequately engaged. There is now enough experience, from enough different sources, to know that better models exist. The Commission should begin its review by learning from them and then build a new procedural architecture worthy of the consequential work it is called to perform.
In other essays, we have imagined what the Commission might be like if economists ran it –
and we contemplated whether machine learning could bring new thinking and insights to the Commission –
