SENTENCING: ALLOCATION OF AUTHORITY

In the United States there are now a wide range of legal approaches to the sentencing of criminal offenders. These include indeterminate sentencing systems in many states, statutory determinate systems in a few others, and a growing number of sentencing guideline systems among the states and in federal law. The diversity in legal structure for punishment decisions has brought with it a great deal of experimentation concerning which decision-makers hold meaningful authority—or discretion—over sentencing outcomes. Allocations of sentencing discretion can be remarkably different as one moves from one American jurisdiction to another.

Definition of sentencing discretion

Sentencing discretion, as the term is used in this entry, exists whenever a participant in the design or operation of the criminal justice system can exercise choice in a way that dictates, places limits upon, or contributes to the sentencing outcome of a particular criminal case or whole categories of cases.

This is a broad definition that stretches familiar usages of the word "discretion." For one thing the definition of sentencing discretion embraces both public and private actors who take part in the processes leading up to final punishment outcomes. For another, the definition of sentencing discretion extends to acts of rulemaking by legislatures, sentencing commissions, and appellate courts, just as the definition extends to individualized decisions that affect only one defendant at a time. The expansive definition is necessary, however, to capture the complex and multilayered accumulation of human choices that eventuate, ultimately, in sentencing outcomes.

The term sentencing outcome also bears definition: As used here, it refers to the total of all sanctions as experienced by a criminal offender up until the moment that the legal system relinquishes jurisdiction over the offender.

Most sentencing outcomes build up over time and are the products of cumulative decisions by multiple actors. The sentencing process is "not over until it's over." For example, in most jurisdictions, a trial judge's pronouncement of a five-year sentence on a convicted offender would hardly qualify as a sentencing outcome. At the day of sentencing, we may still be at a relatively early point in the chronology of discretionary decisions affecting the punishment to be experienced by the offender. A parole board, together with officials in the department of corrections, might still hold enormous authority to fix the offender's actual release date from incarceration. (In some states, early release is possible after a small fraction of the judge's pronounced sentence has been served.) In many systems, we cannot guess the sentencing outcome in a prison case until the offender finally has been released on parole, perhaps years after the judge's decision in the courtroom—and additional years may elapse before the offender has successfully completed any term of supervised post-prison release. When this moment finally comes, we can look back and say what the total sentencing outcome has been.

A discretion diagram

A pictorial discretion diagram can be used to visualize many of the important relationships that exist in real-world sentencing systems. Figure 1, "A Discretion Diagram for Sentencing Systems," is a generic diagram that is not tailored to any up-and-running system. It introduces a cast of characters who may or may not possess meaningful sentencing authority in a given punishment structure. These include, at the "systemic level," the legislature (important in all U.S. jurisdictions) and a sentencing commission (chartered in more than twenty U.S. jurisdictions as of 1999). At the "case specific-level," Figure 1 begins with the parties (meant to include the defendant, the defense lawyer, and the prosecutor), adds various judicial actors, and concludes with corrections and parole officials.

From the perspective of a designer of sentencing systems, Figure 1 might be seen as a basic palette of primary colors from which the designer may work to create the fundamental discretionary relationships within a new system. As we shall see below, there are a great variety of design options for the power relationships that can be fashioned among the players shown in Figure 1.

Indeterminate sentencing systems

For most of the twentieth century, the indeterminate sentencing system was the dominant legal structure for punishment decisions in the United States. Through the 1940s, 1950s, and 1960s it was virtually the universal approach in this country, and remains (with adjustments here and there) the basic system followed by nearly half of the American states.

The original philosophy of indeterminate sentencing structures was that most criminals could be rehabilitated, and that this goal required an individualized inquiry into the life history, behavioral deficits, and "treatment" needs of each offender. In order to carry out an individualized program of "correction," legislatures granted judges unconstrained authority to impose punishments within a broad range of statutory options. It was not unusual, for example, for judges to have discretion to select any punishment (or none at all) within statutory maximum penalties as severe as ten, twenty, or thirty years in prison.

Broad judicial discretion was merely one facet of the case-specific paradigm of indeterminate sentencing. Figure 2 depicts some key features of the allocation of sentencing discretion in a typical indeterminate structure. At the systemic level only the legislature is shown in Figure 2, because traditional indeterminate systems had no sentencing commissions. Moreover, the oval depicting the legislature is shrunken in comparison with most of the other shapes on the diagram. This is meant to illustrate that legislatures play some role in defining punishment outcomes in an indeterminate system, but this role is limited generally to providing the outer boundaries of permissible sanctions. For many if not most serious crime categories in indeterminate systems, the statutorily authorized ceilings on punishment are far in excess of the sanctions most sentencing judges would normally dream of imposing. In this respect, the legislature has opted to make itself a marginal discretionary player as compared with downstream decisionmakers. And because, in traditional indeterminate systems, there is no entity such as a sentencing commission to share systemic discretion with the legislature, such jurisdictions are marked by the relative paucity of systemic oversight by anyone.

The defining feature of indeterminate sentencing structures is their diffusion of meaningful sentencing discretion across numerous actors at the case-specific level. Prosecutors most of the time enjoy unregulated charging discretion; plea bargains between the parties (sometimes including sentence bargains) can have sizable impact on punishment; probation officers, at least in some jurisdictions, make sentence recommendations that are highly influential with judges; the judges themselves usually have a boggling array of choices remaining open to them on the day of sentencing; and, following a judicially pronounced prison sentence, correctional officials and parole boards retain impressive powers to fix actual release dates. (However, in some jurisdictions, parole guidelines have been adopted, in an attempt to at least partially structure release decisions.)

The only case-specific actor without meaningful sentencing discretion in traditional indeterminate systems is the appellate judiciary. For reasons of historical practice, deference to trial courts, and caseload pressures, appellate courts almost universally have resisted responsibility to participate in sentencing outcomes in such systems. This state of affairs persisted despite a vigorous reform movement to inculcate "sentence review" in the 1950s and 1960s, and remains a stable feature of indeterminate systems today.

The abdication of systemic authority in indeterminate sentencing structures, and the concentration of sentencing discretion among case-specific actors, originally reflected a deliberate policy judgment that offender treatment could best be pursued on a case-by-case basis. The correctional process was expected to unfold slowly, as authorities monitored each criminal's progress toward rehabilitation. Parole boards, last in the chronology of case-specific actors, were charged with looking the offender in the eye, examining the full case history, and discerning when the magic moment of rehabilitation had finally arrived.

The most important question to be asked about indeterminate systems today is not whether their configurations of discretion are a fair reflection of an articulable plan. (They are.) Rather, the pertinent question is whether the plan itself is sufficiently sound to justify the bottom-heavy and diffuse allocations of decisionmaking power that characterize indeterminacy. For those who believe that uniformity of punishment is an important goal, or that systemwide policymaking should not be sacrificed wholly to ad hoc individualization, doubts about the wisdom of the indeterminate ideal have become pronounced. This perhaps is inevitable in a world that no longer believes in the widely achievable rehabilitation of offenders, yet continues to operate with the machinery of the damaged theory.

Statutory determinate sentencing

In the 1970s, a handful of states including Arizona, California, Colorado, Illinois, Indiana, and North Carolina modified their former indeterminate sentencing schemes to provide for greater specificity and certainty in authorized punishments as a mailer of statutory command. (Colorado reverted back to indeterminacy in the 1980s, while North Carolina went on to adopt sentencing guidelines in the 1990s.)

The best known statutory determinate sentencing plan is California's. In 1976 the state adopted, and still follows, a "multiple choice" approach under which most serious offenses carry three designated punishments. For example, the current provision concerning first-degree burglary specifies that the sentencing options for the offense include "imprisonment in the state prison for two, four, or six years" (California Penal Code, § 461). In a normal case, the trial judge is expected to impose the middle, or "presumptive," sentence laid out for each crime. As alternatives, the judge may select the "mitigated" or "aggravated" term, provided the judge can cite adequate reasons on the record. Although prison terms are subject to limited reduction at the discretion of corrections authorities, California's determinate sentencing legislation abolished the parole board's authority to decide release dates.

Figure 3 represents California's version of statutory determinacy. It indicates graphically that the legislature is an important discretionary actor in the California system when compared to legislatures in indeterminate systems. (Compare Figure 2.) Instead of writing capacious punishment options into the criminal code, as in indeterminate structures, the California legislature has chosen to be far more directive in its penalty prescriptions. In one sense, therefore, discretionary actors within the trial court system (the parties, probation officers, and trial judges) enjoy a restricted field of sentencing discretion vis-à-vis the legislature. On the other hand, California's abolition of parole release, plus statutory limitations on the release authority of corrections officials, ensure that the sentencing decisions made by the trial court working group will bear predictable relation to the final sentencing outcomes of particular cases. Much more so than in indeterminate sentencing systems, in other words, the sentences pronounced in courtrooms matter in California. Thus, in Figure 3, courtroom actors are represented as discretionary players with meaningful authority. As in most states, however, the California appellate courts exert modest authority over the propriety of sentences; a trial judge's sentence will not be overturned on appeal unless the appellate court finds an "abuse of discretion."

The theory behind California's determinate sentencing reform was that rehabilitative sentencing had proven a failure, and the legislature elected to pursue a new vision that "the purpose of imprisonment for crime is punishment." Further, the California legislature concluded that "the elimination of disparity and the provision of uniformity in sentences can best be achieved by determinate sentences fixed by statute in proportion to the seriousness of the offense as determined by the Legislature to be imposed by the court with specified discretion" (California Penal Code, § 1170). Consistent with these views, sentencing discretion is now concentrated at the "front end" of the discretionary chronology—in the legislature and court system. There is no perceived need to wait months and years for a determination of when offenders have become rehabilitated, so "back end" authorities are granted limited discretionary roles, or eliminated entirely, from the discretionary equation.

Other statutory determinate sentencing reforms of the 1970s, such as those still surviving in Arizona, Illinois, and Indiana, followed a somewhat different scheme from that in California. The criminal codes in such states were amended to set forth narrowed "ranges" of potential punishment for each offense, as opposed to the fixed integers in California law. Still, the overall plan was similar: Typically such statutes provide a "presumptive" range for ordinary cases, with the bookends of "aggravated" and "mitigated" ranges available for unusual circumstances. In addition, Arizona, Colorado (temporarily), Illinois, and Indiana joined California in the elimination of parole release. Thus, these states effectuated a net shift of sentencing discretion toward the "front end" of the decisionmaking chronology, quite similar to that shown in Figure 3 for the California system.

The defining strength and weakness of statutory determinate sentencing reforms is that they rely on legislatures to choose specific penalties (or narrowed ranges of penalties) for specific crimes. Jurisdiction-wide uniformity in sentencing can be promoted in this way, but state legislatures do not have the time or expertise to ponder exact punishments with care. Nor do legislatures have the attention span needed to monitor their sentencing systems in operation, and to make periodic adjustments in the matrices of presumptive sentences. Indeed, most jurisdictions that enacted statutory determinate laws have found that their legislators tend to pass crazy-quilt amendments over time. In addition, legislative determinacy has proven a weak tool to manage prison population growth. The prison population in states like California, Colorado, and North Carolina grew even more quickly under new determinate laws than before such sentencing reforms were instituted. In part, these developments can be attributed to changes in the political climates of the individual states, but they led many to conclude that legislative determinacy was too blunt an instrument for the administration of a statewide sentencing system. The reform impetus toward statutory determinacy appears to have run its course. No jurisdiction has adopted such a structure since 1980.

Mandatory penalties

An important variation on the theme of statutory determinate sentencing occurs when legislatures fix exact penalties, or exact minimum penalties, for particular crimes. For example, some states have laws stating that the sentence for first-degree murder must be a life term in prison without the possibility of parole. This is an instance of a mandatory penalty. In other states, the authorized punishments for first-degree murder may include a death sentence or, at a minimum, a life term in prison without possibility of parole. This is an instance of a mandatory minimum penalty: Following a conviction of the designated offense, no one at the case-specific level holds sentencing discretion to impose a punishment any less severe than a life term of incarceration. Case-specific sentencing discretion exists to exceed the mandatory minimum penalty, but not to undercut it.

Since the 1970s, federal and state legislatures have enacted large numbers of mandatory penalty provisions, and the public popularity of such measures remains high. Such laws commonly apply to crimes involving serious violence, drugs, or firearms. Another species is keyed to criminal record: "habitual offender" laws have long been used to require heavier-than-normal sentences for criminals with substantial prior convictions. In the 1990s a new incarnation of the habitual-offender approach appeared in the form of "three-strikes" laws. Congress and many states have now adopted such laws, which operate on a similar plan: Upon conviction for a third serious felony (these are defined differently from place to place), the judge must sentence the offender to a life term of imprisonment without parole.

As opposed to statutory determinate sentencing, mandatory penalty provisions do not attempt to rework a jurisdiction's overall sentencing structure. They apply to one offense at a time. Mandatory sentencing statutes have been adopted in jurisdictions that otherwise use indeterminate sentencing for the bulk of offenses, in jurisdictions that follow the statutory determinacy approach for most crimes, and in jurisdictions with sentencing commissions and guidelines. In other words, American legislatures have regarded mandatory penalties as a desirable means to produce zones of hyperdeterminacy within every available structural environment.

Suppose that a legislature has passed a statute providing that anyone convicted of a designated offense (e.g., possession of a weapon in connection with a drug transaction) must be sentenced by the trial judge to five years in prison, and may not be granted parole or good time credit toward release. Figure 4 attempts to capture the most salient discretionary features of such mandatory penalty schemes in operation.

At the systemic level, Figure 4 depicts the legislature as an important discretionary actor. The legislature has exercised an especially direct form of sentencing discretion because it has pronounced that, for all defendants convicted of crime x, the punishment shall be sentence y. Indeed, the legislature means to be the one-and-only sentencer for such cases, and has sought to bring about this result by eliminating other discretionary actors from the decisional process. In contrast, the sentencing commission, even if one exists in the jurisdiction, typically holds no authority to modify or soften the statutory punishments in a mandatory penalty scheme. Thus, the commission is crossed out in Figure 4.

Jumping down to the case-specific level, the probation office and courts are also canceled in the discretionary diagram. Following a conviction for the specified crime, nothing is left for such players to do beyond the rote motions of functionaries. Similarly, any parole authority that may exist in the jurisdiction for other crimes is no longer in force for this offense. And corrections officials, under the imagined statute, have lost their discretion to award or withhold good time credits.

Turning our attention to the enlarged shape in Figure 1 that indicates "the parties," experience has shown that the legislative enactment of mandatory penalties, intentions aside, does not succeed in removing the parties' capabilities to make discretionary choices that carry enormous impact on sentences imposed. For example, in cases where the prosecutor has evidence that an offender has committed the kind of gun-and-drug crime covered by the law, the prosecutor may still decline to bring the charge. If the government charges a lesser offense, out of a sense that five years is too much for this particular case, or from a belief that the required facts will be difficult to prove, or as part of a plea bargain, the outcome is almost certain to be a sentence well below that required by the statute. Thus, charging discretion, while not formally a part of sentencing proceedings, can exert a powerful impact on the sentencing outcome. Similarly, the plea bargaining process, in which both parties participate, can and frequently does operate to avoid the force of a mandatory penalty.

Taking stock of the discretionary configurations of mandatory penalty provisions, it is apparent that legislatures have succeeded in part, and have failed in part, in bringing about the structural adjustments attempted in such laws. Legislatures have surely succeeded in eliminating the decision-making authority of many familiar players in the sentencing structure. This much comports exactly with the theory of mandatory sentencing: the legislature is trying to impose its own judgments about appropriate sentences on everyone else in the system. Getting rid of the input of sentencing commissions, probation officers, judges, corrections officials, and parole boards is consistent with such a purpose.

At the same time, mandatory penalties greatly heighten the importance of discretionary decisions by the parties. Through charging and plea bargaining discretion, the parties have become the only actors at the case-specific level who may pass upon the application of the mandatory sentence. Moreover, their choices are effectively unreviewable. In most other discretionary arrangements, the parties' authority is diluted by the later-in-time discretions of other case-specific players, such as trial courts, departments of corrections, and parole boards. Mandatory sentencing—which at first glance would appear to privilege the discretion of legislatures—instead produces a hegemony of the charging and plea bargaining decision points. It is hardly surprising that empirical studies of mandatory penalties have concluded that such laws actually work to increase arbitrariness and disparities in sentencing.

Sentencing commissions and guidelines

Since the 1980s, the most popular form of "sentencing reform," among jurisdictions that have wished to move away from the traditional approach of indeterminate sentencing, has been the creation of a sentencing commission empowered to promulgate sentencing guidelines. The first sentencing commissions were chartered in Minnesota and Pennsylvania in the late 1970s; their guidelines went into effect in the early 1980s. Today, there are more than twenty sentencing commissions across the United States, and sixteen jurisdictions with some form of sentencing guidelines in operation. Most of the remaining commissions are at work formulating guideline proposals that have not yet been adopted by their state legislatures.

The amount of sentencing discretion held by sentencing commissions varies widely across American guideline jurisdictions. In some jurisdictions, such as Washington and North Carolina, the guidelines and guideline amendments authored by the commission do not take effect unless affirmatively enacted into law by the state legislature. This might be called a "legislative adoption" model. In other jurisdictions, such as Minnesota and in the federal system, the commission's guidelines and amendments automatically become effective within a specified time period after their promulgation by the commission—unless the legislature acts affirmatively to disapprove the commission's proposals. This might be called a "legislative veto" model. All other things being equal, the share of discretion held by a sentencing commission vis-à-vis the legislature is greater in the legislative-veto model than in the legislative-adoption scheme.

In addition to the commission's relationship to the legislature, the discretionary interrelationships of all actors within guidelines systems have differed dramatically from jurisdiction to jurisdiction since guidelines were first created in 1980. Some sentencing commissions are very powerful players when compared with trial judges in their systems and, in some other places, trial court discretion eclipses the authority of the commission. In a handful of U.S. guideline jurisdictions, the appellate bench exerts meaningful discretion in the realm of sentencing decisions, but in the majority of guidelines states, the appellate courts remain marginal discretionary players. Most guideline jurisdictions have abolished the release authority of the parole board and have limited the release discretion of corrections officials, but in some guideline systems these "back-end" release discretions remain fully intact.

Discretion diagrams can be useful tools to capture these and other permutations of decision-making authority across guideline jurisdictions. The diagrams allow the major discretionary features of different systems to be summarized visually, so that policy makers may compare the advantages and disadvantages of alternative structures.

The federal system

The current federal guidelines system, in effect since 1987, attempts to concentrate sentencing discretion at the systemic level and constrain such discretion at the case-specific level. In both respects, the federal system differs sharply from most guideline structures that have been adopted at the state level.

Although there is a great deal of variation across the country in the operation of the federal sentencing laws, Figure 5 attempts to capture a number of general observations that have been made about the system. First, at the systemic level, both Congress and the U.S. Sentencing Commission are pictured in Figure 5 as highly efficacious discretionary players. Through the adoption of many mandatory penalties, and through close oversight of the commission, Congress has assumed a major role in the determination of sentencing outcomes. The U.S. Sentencing Commission has also exerted great authority over sentences at the case-specific level through the creation of highly detailed guidelines and provisions that give federal district court judges only narrow authority to "depart" from penalties prescribed by the guidelines.

At the case-specific level, the most obvious discretionary effects of the new federal system are at the back end of the decision-making chronology. The release authority of the parole board has been abrogated entirely (indicated by an "X" in Figure 5), and the good-time credits that may be granted by federal corrections officials have been cut back to a maximum of 15 percent of the pronounced prison term (thus only a small shape for "corrections officials" in Figure 5). At the case-specific level, most sentencing authority has been removed from the back end of the system and is clustered among the actors who participate in the litigation process.

Among litigation players, the parties have probably assumed the lion's share of sentencing discretion in the federal system. Under either mandatory penalties or the tightly restrictive federal guidelines, the charge(s) of conviction can have forceful if not determinative impact upon what the punishment in a given case will be. Thus, the charging decisions of prosecutors, plus the plea agreements consummated by both parties, will often deliver cases to the judiciary with very little room to move punishments up or down from that prescribed by statute or guideline.

As between the two parties in federal litigation, it has generally been said that the prosecutor holds considerably greater power over sentencing outcomes than the defense attorney. While it is true that roughly 90 percent of federal cases are resolved through a guilty plea—and both sides must agree to such resolutions—prosecutors are perceived as enjoying a much stronger bargaining position than defense counsel.

The heightened authority of federal prosecutors might be ascribed to three sources. First, given the large number of mandatory penalties in federal law, the prosecutor's charging authority can determine by itself whether a heavy mandatory punishment will be at issue in a given case. Given the prosecutor's sole discretion in such charging decisions, the government may also use the threat of bringing such a charge, or the promise to dismiss such a charge, as bargaining leverage in plea negotiations.

Second, it is widely acknowledged that the severity of penalties in the federal system has increased substantially since before guidelines were enacted in 1987, in part because the guidelines themselves are more punitive than the preguidelines sentencing practices of federal judges, and in part because of the proliferation of mandatory penalties. As a result, many federal defendants who go to trial today face heavier sentences following a conviction than would have been the case before 1987. This unappetizing prospect can compel defendants to accept plea bargains that are less favorable than those to which they would have assented before the guidelines. This is clearly a net gain to the prosecution—although it is also a direct outgrowth of the sizable sentencing discretions exercised by Congress and the sentencing commission. If those systemic decision-makers had not decided that heavier sanctions were wanted as a matter of public policy, the bargaining position of federal prosecutors would not have eclipsed that of defense lawyers to the extent we have seen.

Third, prosecutors in the federal system have gained a unique boost to their sentencing discretion under the "relevant conduct" provision of the federal guidelines (United States Sentencing Guidelines, § lB1.3). The provision requires that sentences be increased at the sentencing hearing when the government proves, by a preponderance of the evidence, that the defendant committed certain offenses in addition to those of which the defendant was convicted. Thus, for offenses within the scope of the relevant conduct provision, a federal prosecutor has two opportunities to secure punishment: The prosecutor may seek a formal conviction by trial or guilty plea, or the prosecutor may wait to prove the offense at the sentencing hearing, under a reduced standard of proof, and where the defendant may not claim such rights as trial by jury, the exclusionary rule, double jeopardy, or the protections of the federal rules of evidence. No state guideline system has yet created an analogue to the relevant conduct provision, so it may be viewed as a source of prosecutorial sentencing authority that is entirely unique to the federal guidelines.

At least once in a while, in some but not all federal districts, the relevant conduct provision has also worked as a source of sentencing discretion enjoyed by probation officers. Under federal law, probation officers are required to bring evidence of all offenses included within the provision to the attention of the sentencing court—whether or not the prosecutor has chosen to do so. In theory, this requirement was intended to cut back on prosecutorial discretion by ensuring uniform sentencing for all relevant conduct even when the United States attorney would prefer to procure a plea bargain by agreeing to drop some or all relevant conduct allegations. In practice, however, federal probation officers do not routinely "upset the applecart" of negotiated guilty pleas by exposing facts at sentencing that the parties had agreed to keep out. But anecdotal accounts tell us that this has happened in some federal cases, perhaps depending on the individual probation officer's talents and predilections, and perhaps occurring in some districts more than others. Even though this form of probation officer sentencing discretion does not appear to be a stable feature of guideline sentencing in the federal system, it does suggest that the probation officer in Figure 5 should be depicted with more than minimal power to influence sentences.

Last, but by no means least important in Figure 5, are judicial actors. Here, the most remarkable feature of the federal system is how little sentencing discretion is left in the hands of district court judges. Under mandatory penalties, as we have seen, trial judges often have no discretion at all. In addition, district court judges are permitted small latitude to "depart" from the narrow sentencing ranges calculated under the federal guidelines. When district courts have chosen to deviate from the guidelines, the federal appellate courts have been vigilant in upholding the literal terms of the guidelines. Indeed, one recent study found that the chances of a trial court sentence being reversed on appeal in the federal system were ten times weaker than under the guidelines in Minnesota, and nearly fifty times greater than in Pennsylvania's guidelines system. Thus, federal district court judges are often heard to complain that their function has been reduced to "sentencing by computer."

In contrast to the shrunken sentencing discretion of district courts, the federal appeals bench has assumed a vigorous role in the new system. It is important to emphasize, however, that the discretion exercised by the U.S. Courts of Appeals has not been in the direction of creating new judge-made principles for the punishment of offenders. Instead, the powers of the appellate bench have been devoted heavily toward the strict enforcement of the commission's guidelines. This behavior on the part of the appeals courts is in fact a linchpin of the entire discretionary structure of federal sentencing. If the courts of appeals were more relaxed in their review of district court sentences, then the sentencing discretion of district court judges would swell immediately. As a consequence, the systemic powers of Congress and the U.S. Sentencing Commission would deflate (because their prescriptions were less enforceable). And, very likely, if the guidelines were to lose some of their constraining power over trial courts, the importance of the parties' charging and plea bargaining discretions would drop off as well. The appellate courts in the federal system are like a finger in the dike, preventing the trial courts from exercising meaningful discretion over sentencing outcomes, while at the same time allowing great quantities of discretion to pool among actors earlier in the decision-making chronology.

The discretionary relationships surveyed here are complex to grasp, despite the visual aid of Figure 5. Most of the allocations of discretion in the federal system make sense, however, in light of the policies the system was designed to achieve. As recounted by Kate Stith and José A. Cabranes, one major goal of federal sentencing reform was to stiffen punishments for a host of crimes, and to take away the discretion of district court judges to mete out penalties that Congress and the U.S. Sentencing Commission perceived were too lenient. Certainly, to date, these objectives have largely been achieved. Strict enforcement of the federal guidelines was also supposed to promote uniformity in sentencing and, once again, the planned reduction of trial court discretion is consistent with this purpose. However, just as we saw earlier in the case of mandatory penalties (see Figure 4), the federal guidelines system has not solved the riddle of prosecutorial and plea bargaining discretion at the case-specific level.

The Delaware system

The Delaware guidelines system, as it has been in operation since 1990, is illustrated in Figure 6. The Delaware system utilizes the same basic array of institutions as those at work in the federal system (legislature, commission, guidelines, courtroom actors, correctional officials, the abolition of parole release), but to very different effect. A comparison of Delaware with the federal structure is a vivid way to contrast the allocations of discretion that are possible in different guidelines systems.

The greatest distinction between Delaware's guidelines and those in federal law is that, while the federal guidelines are tightly binding on federal judges, Delaware's guidelines are merely "voluntary"–to be used, or not used, by trial judges as they wish. To indicate the weak legal force of the Delaware guidelines, both the legislature and sentencing commission are pictured in Figure 6 as diminutive discretionary players. When these systemic actors speak through guidelines, they utter no more than advisory statements concerning what punishments should be.

At the case-specific level, the current Delaware approach to back-end decision makers is quite similar to federal law. As of 1990, parole release was abolished in Delaware and the release authority of corrections officials was placed under new limitations. Also like the federal system, the sharp confinement of discretion at the back end of the system tends to shift authority toward actors at the adjudication stage.

Among the litigation players, however, the Delaware structure could hardly be more different from the federal arrangement. Because guidelines are only advisory, trial courts may impose any punishment authorized in the criminal code—and Delaware's statutory penalties are expressed in broad statutory ranges left over from the days of indeterminate sentencing. Also like indeterminate schemes, the appellate courts in Delaware exercise virtually no review of the sentencing decisions of trial judges. The result of advisory guidelines and the near absence of sentence review, however, is the creation of far more sentencing discretion in Delaware's trial courts than that possessed by trial judges in indeterminate systems. This is because, under indeterminacy, the trial court's discretion is diluted by the later-in-time discretions of corrections officials and parole authorities. (Compare Figure 6 with Figure 2.) In the Delaware system, however, the trial judge is the last decision-maker to hold meaningful power over punishment outcomes.

It is especially interesting to compare the discretionary power of trial courts under the Delaware guidelines with that in the federal structure. Although both jurisdictions purport to be "sentencing guidelines" jurisdictions, and both systems operate with a similar cast of institutional players, the legal structure of the federal system brings all pressure to bear on choking off the sentencing discretion of trial judges, while Delaware has oriented its system to removing all checks upon the sentencing authority of its trial bench. The comparison of Delaware and the federal system illustrates the following point: As the sentencing discretion of the legislature or sentencing commission is inflated (as in federal law), the sentencing discretion belonging to trial courts tends to deflate—and vice versa. In Delaware, the weak and advisory powers exercised by systemic decision-makers are important complements to the very appreciable powers held by trial judges.

There have been no formal studies of prosecutorial discretion or plea bargaining practices under the Delaware guidelines, but it is clear that the case resolution presented by the parties cannot "tie the judge's hands" to the same extent as in the federal system. Significantly in Delaware, one does not hear the complaints so often voiced in the federal system that prosecutors have gained huge reservoirs of sentencing discretion under guidelines. It is difficult structurally to have a powerful trial bench and, at the same time, a system in which prosecutors hold determinative authority over sentencing outcomes. In Delaware, if the guideline sentence attending a plea agreement is not to a trial judge's satisfaction, the judge may depart from the voluntary guidelines at will.

Given the realities of courtroom operations, it is probable that plea agreements in Delaware are honored in spirit by trial courts much of the time. (Otherwise the courts, along with everyone else, would have to bear the costs of more frequent trials.) Therefore, the parties surely have substantial impact on sentencing decisions in Delaware, just as the recommendations of trusted probation officers may influence some judicial decisions. But the overall structure of Delaware law places the greatest and most definitive sentencing discretion directly in the trial judges. The discretionary allocations under Delaware's guidelines are so different from those under the federal guidelines, that the Delaware system might be described as the federal system "insideout."

The Minnesota system

Minnesota, the oldest of all guideline systems in the United states, has taken a middle course between the polar discretionary arrangements illustrated in the federal system and the Delaware system. Where the federal system attempts to concentrate most sentencing discretion at the systemic level, and where the Delaware system attempts to collect very little authority at the systemic level, the approach in Minnesota has been to balance the discretion exercised at the systemic level (by legislature and commission) with discretion discharged at the case-specific level (by the judiciary). Figure 7 attempts to capture this equilibrium in the apportionment of discretion with shapes of similar size for the major discretionary players at both levels.

Less concerned with enforced uniformity than the architects of the federal system, the designers in Minnesota saw the case-specific discretion of trial judges as a desirable feature in the overall system. The original set of guidelines promulgated by the Minnesota Sentencing Commission purported to govern only "ordinary" or "typical" cases. The theory was that trial judges would follow the guidelines prescriptions in the majority of cases, but would retain discretion to depart from the guidelines in cases that were sufficiently atypical. Such a departure decision, however, would have to be explained on the record, and the trial judge's reasoning made subject to appellate review.

As in the federal system, the actual operation of the Minnesota guidelines has been greatly dependent upon the behavior of the state's appellate courts. If Minnesota's appellate judiciary had chosen to enforce the guidelines to their strict letter, as the federal appeals bench often has done, then trial judges in Minnesota might have become marginal discretionary players in the new system. Instead, however, the Minnesota Supreme Court and Court of Appeals have taken a measured approach to guideline enforcement. Soon after guidelines were enacted, the appellate courts declared that the guidelines were legally binding on trial judges (not merely voluntary, as in Delaware), yet the appellate courts said they would allow sentencing judges to select non-guideline punishments in cases where "substantial and compelling reasons" could be set forth on the record. In practice, trial court sentences below the guideline range have almost always been upheld on appeal in Minnesota and sentences somewhat above the range are usually affirmed as well, but sentences greatly above the range are hard-pressed to survive on appeal in the absence of compelling justification. One study of sentencing appeals rendered in 1995 found that, overall, a sentence handed down by a trial judge in Minnesota had only one-tenth the chances of being reversed on appeal as in the federal system.

The appellate bench in Minnesota has enforced a kind of give-and-take arrangement between the sentencing commission and trial judges. The commission's guidelines are expected to govern most of the time, but a sentencing court sufficiently moved by the circumstances of an individual case can work around the guidelines, subject to review. As among Minnesota's sentencing commission, the trial bench, and the appellate bench, it would be hard to say which has been the dominant player within the sentencing structure since guidelines were enacted in 1980. Indeed, the three institutions are constantly jostling within a framework of shared powers.

To fill out the description of discretionary relationships shown on Figure 7, it should be noted that the Minnesota legislature has not played a consistently powerful role in the determination of punishment outcomes. When the Minnesota Sentencing Commission was first created in the 1970s, the legislature gave it few specific instructions about what its mission should be. Thus, the initial shape of Minnesota's guidelines was more the result of commission discretion than legislative directive. Over time, however, state legislators have learned that they can exert immediate influence over sentencing policy by instructing the commission to amend its guidelines in specific ways. Such legislative interventions have not been an annual event in Minnesota, but they testify to the power of systemic governance available to the legislature should the legislature choose to use it.

On the case-specific level, Minnesota, like many other guidelines jurisdictions, chose to eliminate parole release when sentencing guidelines went into operation, and also cut back on the early release authority of corrections officials. As we have seen in both the federal system and in Delaware, such changes shift discretion toward the litigation players at the case-specific level. Indeed, one selling point among Minnesota's trial judges, when the state's guidelines were first proposed, was that judicial sentencing authority would actually be increased in comparison with the prior system of indeterminate sentencing. While trial judges are more constrained than formerly by authorities at the systemic level, new controls on the back-end release of prisoners ensure that judicially pronounced punishments bear predictable relationship to sentences actually served.

The full-sized shape on Figure 7 for "the parties" represents the continued importance of plea bargaining within the Minnesota system. Richard Frase, in a 1993 study, found that the pre-guidelines practices of plea negotiation seemed alive and well in Minnesota's guideline structure, but there is no reason to believe that there has been an unintended exacerbation of the importance of charging decisions and plea bargains, as many perceive to have occurred in federal law. The experience of Minnesota, like Delaware, disproves any notion that guidelines structures inevitably enlarge the sentencing discretion of prosecutors. The track record of both states, in comparison with the federal structure, suggests that the best remedy for an undue bulging of discretion in the prosecutor is a system that preserves meaningful sentencing authority in the hands of trial judges. Where trial courts are marginalized, as in the federal system pictured in Figure 5, or as under mandatory penalties as pictured in Figure 4, charging and plea bargaining discretions tend to gain enormous importance.

Guidelines with parole release

The three guideline systems highlighted above are sufficiently different from one another that they give the reader a fair sense of the diverse permutations of sentencing discretion that are possible under guidelines—depending on how the overall guidelines system is designed. The discussion above has not been comprehensive, however, for the sixteen guidelines systems currently up and running in America are all different from one another, and the new guidelines proposals on the table in several additional states promise still more experimentation.

Seven of the existing guidelines states, for example, have not abrogated the release authority of the parole board. Pennsylvania, which adopted guidelines in 1982, is one such jurisdiction. Much like Delaware, Pennsylvania's guidelines are advisory and are not enforced by the state appellate courts, so sentencing discretion at the systemic level does not have many teeth. At the case-specific level, the release authority of the corrections department has been abolished in Pennsylvania, but the parole board still has discretion to release prisoners after a minimum of one-half of their sentence has been served (Pennsylvania Consolidated Statutes, § 9756). Thus, trial judges and representatives of the sentencing commission in Pennsylvania claim that the most important ultimate decision-maker, for prison cases, is the parole board. Certainly, when significant back-end release discretion exists to modify trial-court prison sentences, the power of trial judges is diminished. Although Pennsylvania's structure is in many ways similar to the Delaware system shown in Figure 6, the major differences between Delaware and Pennsylvania occur in the relative powers of the parole board and the trial judiciary. An overview of these discretionary effects in Pennsylvania is presented in Figure 8.

Discretonary actors elsewhere in the process

There are a number of decision-makers who, in some cases, and in some jurisdictions, hold and exercise discretion to influence punishment outcomes, but who are not represented on the discretion diagrams featured above. A number of these decision-makers will be noted in this section.

First, victims of offenses in the 1980s and 1990s have increasingly become a part of the formal criminal process, including sentencing proceedings. Victims commonly submit "victim impact statements" to sentencing courts, in writing, in the form of oral statements at sentencing hearings, or both. Earlier in the process, victims in many states now consult with prosecutors concerning the charges that will be filed in their cases, and whether or not they approve of plea agreements contemplated by prosecutors. Earlier still, victims might be said to exercise a form of discretion when they decide in the first instance whether or not to report a crime committed against them. A crime that is not reported stands small chance of being punished. We have little empirical knowledge concerning the actual effect upon sentences that has resulted from the growing input from crime victims, but the trend across the country remains one of an increasingly important role for victims.

Second, executive clemency was once an extremely important form of sentencing discretion in the United States—it was a major feature of criminal justice before parole was invented in the late nineteenth century. The president and the governors of most states hold power to pardon offenders, before or after conviction, or to commute sentences (that is, to reduce the severity of sentences). In some jurisdictions this executive power is exercised in conjunction with an agency such as a board of pardons, and in other jurisdictions it remains within the sole discretion of the chief executive. The clemency power continues to exist in all U.S. criminal justice systems, but was probably used less often in the latter twentieth century than in previous times. In the 1980s and 1990s, many elected officials became wary of taking action that could identify them as "soft on crime." Executive clemency has hardly disappeared from the legal horizon, but it is seldom a factor in run-of-the-mill criminal prosecutions.

Third, at least in a handful of states, voters can now exert direct impact upon state sentencing laws through the initiative process. Three-strikes laws, for example, appeared on the ballots of a number of states in the 1990s. The voters in Washington State passed a number of mandatory penalties, including a three-strikes law, that have had real impact on prison growth in the state. Such adjustments in the punishment structure can affect other decision-makers at the systemic level. For instance, if mandatory penalties created through the initiative process result in a significant drain upon prison bed spaces, less flexibility remains in the hands of the legislature and sentencing commission to determine priorities in incarceration policy.

Fourth, a small number of states employ jury sentencing in noncapital cases that go to jury trial, and a majority of jurisdictions with the death penalty rely upon juries to make capital sentencing decisions. In such instances, juries possess the kind of sentencing discretion that is familiarly reposed in trial judges. Where such practices thrive, it is often because jurors are considered to be better representatives of community sentiment than a single member of the judiciary, and thus better positioned to render just punishments. Jury sentencing is often criticized, however, on the ground that jurors lack experience across the full range of criminal prosecutions, and are handicapped in their abilities to compare one case with another for purposes of rendering proportional sanctions. Jury sentencing tends to be politically popular where it exists, but there is no trend to introduce jury sentencing, particularly in noncapital cases, to jurisdictions that have not had such a tradition.

Fifth, in some types of cases, police officers or other law enforcement officers are said to have a new kind of sentencing discretion. Of course, the police have always had considerable discretion over whom to arrest, and a decision not to arrest usually forecloses any possibility of later punishment for the crime at hand. In recent years, however, a number of cases have arisen under sentencing guidelines and mandatory penalties in which the law enforcement decision concerning when to make an arrest has acquired new importance. If, for example, a mandatory penalty requires that a drug offender buy or sell a specified quantity of drugs, some defendants have alleged that the police have prolonged their undercover operations until charges entailing the triggering quantity of drugs can be brought. In the case of undercover sales, it is possible for the police to engineer a number of transactions, waiting for the drug amount to exceed the statutory threshold. This behavior has sometimes been called "sentencing entrapment." To date, there is no legal remedy for a defendant who claims to have been the target of such police behavior, nor is there consensus that there is anything wrong with such law enforcement activity, despite its deliberate impact upon punishment consequences.

Finally, the offender himself might be seen as one of the most important discretionary actors in the criminal process. Barring a miscarriage of justice, the offender's provable criminal conduct places boundaries on the potential punishment in every case, simultaneously creating and restricting the discretions of all downstream decision-makers. It is sometimes said that the prosecutor has the power to select punishment in a case involving a mandatory minimum penalty, or in a restrictive guidelines system, through the decision of what charge to file. There is much truth to the claim that prosecutors have sizable discretion not to charge (or to drop charges in return for a guilty plea or other cooperation), thereby lowering the punishment range available to the court and later system actors. But prosecutors have much less power to increase the punishment range; they cannot "select" a more severe punishment by filing more serious or additional charges unless there is a realistic chance of proving those charges beyond a reasonable doubt (or in federal system, at least proving such charges by a preponderance of evidence, as "relevant conduct" for sentencing purposes). Thus the prosecutor's power derives directly from the offender's underlying guilt.

Kevin R. Reitz

See also Amnesty and Pardon; Criminal Justice System; Guilty Plea: Plea Bargaining; Probation and Parole: Procedural Protection; Probation and Parole: Supervision; Prosecution: Prosecutorial Discretion; Punishment; Sentencing: Disparity; Sentencing: Guidelines; Sentencing: Mandatory and Mandatory Minimum Sentences; Sentencing: Presentence Report; Sentencing: Procedural Protection.

BIBLIOGRAPHY

Alschuler, Albert W. "Sentencing Reform and Prosecutorial Power: A Critique of Recent Proposals for 'Fixed' or 'Presumptive' Sentencing." University of Pennsylvania Law Review 126 (1978): 550.

American Bar Association. Standards for Criminal Justice: Sentencing, 3d ed. Chicago: ABA Press, 1994.

Bureau of Justice Assistance. National Assessment of Structured Sentencing. Washington, D.C.: Office of Justice Programs, 1996.

——. National Survey of State Sentencing Structures. Washington, D.C.: Office of Justice Programs, 1998.

Frankel, Marvin E. Criminal Sentences: Law Without Order. New York: Hill and Wang, 1973.

Frase, Richard S. "Implementing Commission-Based Sentencing Guidelines: The Lessons of the First Ten Years in Minnesota." Cornell Journal of Law and Public Policy 2 (1993): 279–337.

——. "Sentencing Guidelines in the States: Lessons for State and Federal Reformers." Federal Sentencing Reporter 6 (1993): 123–126.

Gebelein, Richard. "Sentencing Reform in Delaware." In Sentencing Reform in Overcrowded Times: A Comparative Perspective. Edited by Michael Tonry and Kathleen Hatlestad. New York: Oxford University Press, 1997. Pages 88–91.

Knapp, Kay A. "Allocation of Discretion and Accountability Within Sentencing Structures." University of Colorado Law Review 64 (1993): 679–705.

Nagel, Ilene H., and Schulhofer, Stephen J. "A Tale of Three Cities: An Empirical Study of Charging and Bargaining Practices Under the Federal Sentencing Guidelines." Southern California Law Review 66 (1992): 501–566.

Parent, Dale G. Structuring Criminal Sentences: The Evolution of Minnesota's Sentencing Guidelines. Stoneham, Mass.: Butterworth Legal Publishers, 1988.

Reitz, Kevin R. "Sentencing Guideline Systems and Sentence Appeals: A Comparison of Federal and State Experiences." Northwestern Law Review 91 (1997): 1441–1506.

——. "Modeling Discretion in American Sentencing Systems." Law & Policy 20 (1998): 389–428.

Rothman, David J. Conscience and Convenience: The Asylum and Its Alternatives in Progressive America. Boston: Little, Brown and Company, 1980.

Steiger, John C. "Taking the Law into Our Own Hands: Structured Sentencing, Fear of Violence, and Citizen Initiatives in Washington State." Law & Policy 20 (1998): 333–356.

Stith, Kate, and Cabranes, JosÉ A. Fear of Judging: Sentencing Guidelines in the Federal Courts. Chicago: University of Chicago Press, 1998.

Tonry, Michael. Sentencing Matters. New York: Oxford University Press, 1996.

——. Reconsidering Indeterminate and Structured Sentencing. Washington, D.C.: Office of Justice Programs, 1999.

U.S. Sentencing Commission. Special Report to the Congress: Mandatory Minimum Penalties in the Federal Criminal Justice System. Washington, D.C.: U.S. Sentencing Commission, 1991.

Zimring, Franklin E. "A Consumer's Guide to Sentencing Reform: Making the Punishment Fit the Crime." Hastings Center Report 6 (1976): 13–17.