Sentencing and interbranch dialogue.

Author:Fish, Eric S.
Position:II. Delegation and Feedback Loops B. Combining Individual and Systemic Justice through Dialogue through Conclusion, with footnotes, p. 581-608

    The current academic and policy discussion about sentencing assumes that the legislature must choose between judges and a sentencing commission. (111) Either judges are given the authority to sentence with the commission providing nonbinding advice, or the commission is tasked with setting sentencing policy and judges are only allowed to fill in the gaps. In short, each actor must be delegated its own zone of authority where its respective advantages and disadvantages will flourish. The dialogic model of sentencing explodes this premise. It presents a strategy for combining the advantages of judges and sentencing commissions that limits the defects of each, and that causes the sentencing system to rationally evolve over time.

    In a dialogic sentencing system, judges and the sentencing commission approach their respective roles not as independent authors of sentencing policy, but as collaborators in elaborating a law of sentencing that serves the ends of both uniformity and individualization. (112) This means that judges do more than just follow the guidelines or fill in gaps where the guidelines are incomplete. Rather, judges systematically influence the process of guideline writing by having their departure decisions incorporated into the guidelines formula. The commission, in turn, does more than just provide a set of policy suggestions that judges can take or leave at their discretion. The guidelines actually influence judges, either through the force of persuasion, the force of peer effects, or the force of appellate review. Thus judges' sentences help determine the content of the guidelines, while the guidelines in turn help determine judges' sentences. And a practice of reciprocal reason giving, wherein the commission provides reasons for its guideline choices and judges provide reasons for departures, ensures that the signal is clear in both directions. Such a system depends on judges and sentencing commissions adopting a cooperative attitude towards each other, rather than fighting with one another to control the sentencing process. If these actors have radically different sentencing views or regard one another as an enemy to be dominated, the dialogic model will fail. In the federal system, for instance, Congress and the Sentencing Commission have often viewed their role as to control judges rather than to cooperate with them, (113) although this dynamic has changed somewhat recently with respect to the Commission. (114) The success of dialogue, then, is contingent on politics. But if the various actors adopt a cooperative orientation, this model is superior to the alternatives. There are several different institutional frameworks in which such dialogic feedback loops can work: with advisory guidelines or presumptive guidelines, with or without appellate review, with a sentencing commission controlled by the judiciary or the political branches, and so on. But the basic logic of the system is that each actor systematically influences the decisions of the other.

    A feedback loop system reduces the problems associated with rigid guidelines, because it permits a judge to depart in many circumstances where the guidelines produce a sentence that the judge finds unjust. Thus the sentencing hearing retains its meaning, and the individual judge's practical wisdom and moral authority are retained, because the judge can exercise discretion to sentence outside the guideline range. Further, circumstances where the guidelines fail to incorporate the relevant facts (or incorporate them but give them arbitrary and illogical weight) can also be avoided, and indeed fixed, because judges' departures will systematically influence the process of guideline writing. The guidelines will thus change over time, as judicial departures are incorporated into the guidelines in a variety of ways to account for a greater diversity of fact patterns. This change might lead to more complexity, as guidelines are added to incorporate new factors into the calculation, or it might lead to less complexity, as illogical guidelines are deleted to reflect judges' rejection. And this change can happen in a variety of ways: through appellate case law that is incorporated into the guidelines, through commission-collected data on departure practices, or through formal amendments to the guideline formula. The guidelines will adapt over time to reflect judges' views about proper sentences in particular cases, and so will avoid the problem of rigorously applying commission-generated sentencing policies in cases where they are unjust or nonsensical. A feedback loop system thus preserves the main benefits of judicial discretion--it maintains the role of the judge in sentencing and ensures that sentences will be individualized, not the product of an incomplete and rigid formula.

    A dialogic sentencing system also limits the pathologies of judicial discretion. In a presumptive sentencing system where departures are appealable, interjudge disparity will obviously be limited by appellate case law. But even in dialogic systems that are purely advisory, feedback loops reduce interjudge disparity by rendering the guidelines more likely to earn judges' compliance. If judges understand the sentencing commission's role as helping judges to develop a comprehensive law of sentencing, and thus recognize the guidelines as stemming in part from judicial practice rather than solely the policy decisions of a separate agency, they will likely be more willing to follow the guidelines. (115) Moreover, responsive guidelines are also more likely to accord with judges' views of what a proper sentence should be in a particular case. While judges will still certainly bring their different substantive moral viewpoints to bear on their sentencing decisions, the framework provided by the guidelines can gain judges' adherence through a process akin to John Rawls' idea of "reflective equilibrium." (116) According to Rawls, "[w]hen a person is presented with an intuitively appealing account of his sense of justice, he may well revise his judgments to conform to its principles even though the theory does not fit his existing judgments exactly." (117) Similarly, when judges are confronted with a guidelines formula that tracks and partly explains their own intuitions about sentencing, they may well revise those intuitions to accommodate the guidelines. This is especially so if the sentencing commission persuades judges by including reasoned explanations for its recommendations, or if the commission can use peer effects by showing that other judges generally follow the guidelines. (118)

    Finally, a dialogic guidelines system is more morally rational than its alternatives, because the incorporation of judicial departures corrects for oversights, omissions, and errors in the guidelines formula that are made clear in concrete cases. A guidelines system governed by feedback loops follows a very different institutional logic from a guidelines system that is imposed from without by an agency or legislature. In the latter case, the external agency or legislature generates its own abstract moral judgments to cover wide categories of cases and imposes these moral judgments on the sentencing process. By contrast, a system governed by feedback loops generates moral judgments from concrete fact patterns--specific cases where a judge must impose a sentence. These concrete judgments are reported back to the commission, which evaluates them and uses them to update the guidelines in an iterative process. Over time, the system absorbs new data from judges' decisions and incorporates it into an ever more sophisticated and complete framework. Thus, a morally rational system is generated through an experimental process involving many little decisions, not imposed all at once.

    Such a system generates sentencing policy through a process of what Charles Lindblom calls "muddling through." (119) Rather than undertaking the impossible task of specifying in advance every relevant sentencing principle and how it should be weighed in every kind of case, the system integrates many individual decisions applying different principles in particular cases over time. (120) Judges choose principles at the same time as they decide sentences, and these individual decisions are aggregated over time into a self-consistent sentencing system with a coherent set of values. A dialogic system also embodies the contrast between top-down and experimentalist bureaucracy that is developed in Charles Sabel's and William Simon's work on the post-New Deal administrative state. Mandatory guidelines follow an old-line command-and-control model, emphasizing rule-bound bureaucracy and deference to ineffable expertise. (121) Feedback loops, by contrast, are a form of experimental administration. The central authority establishes certain framework goals, and then provides decentralized power to autonomous local units that craft their own methods of achieving the goals. The local units, in turn, must frequently report on their performance to the central authority and provide reasons for their decisions, while the central authority facilitates comparisons between local units. The framework goals and evaluative procedures are updated on the basis of this review process. (122) A feedback loop system for sentencing follows an analogous logic, with the legislature establishing the framework goals (such as increased severity, more cost savings, a focus on rehabilitation, or whatever else), the sentencing commission functioning as the central authority coordinating and comparing the various local units, and judges working as local units making sentencing decisions. Over time, these institutions develop sentencing policies through an iterative learning process that corrects problems in the initial set of guidelines by testing them in concrete cases, departing where they are...

To continue reading