In their article in this issue, Professors Peter Menell and Ryan Vacca describe a federal court docket that is overloaded and unable to process cases efficiently.[2] As they depict it, justice in the federal courts is either delayed or denied,[3] disparity in legal outcomes among circuits is increasing, and the Supreme Court is falling farther and farther behind in resolving circuit splits.[4] While these problems have been around for a while,[5] Menell and Vacca argue they are getting worse and will only continue to worsen if radical action is not taken.[6] Their article provides enough of a factual record to raise the concern that, because of their workload, the federal courts are not resolving cases as capably as they could.

While Menell and Vacca focus on civil litigation in the federal system, this Article looks instead at criminal cases. It first considers whether the problems Menell and Vacca describe on the civil side afflict criminal litigation to the same extent. On the assumption that the problems in the criminal docket are similarly acute, it then considers whether anything can be done about them.

Part I of this Article assesses the efficiency, uniformity, and quality of criminal justice in the federal system. It starts by noting that, while the federal criminal docket is not as overloaded as the civil docket on which Menell and Vacca focus, the number of criminal and prisoner cases commenced in federal court has far outpaced increases in judgeships.[7] Perhaps as a result, resolution of these cases at the district court level has slowed appreciably over the past several decades, and while the rate at which criminal cases are terminated at the appellate level has not changed substantially, that stability appears to have come at a serious cost. Significant circumstantial evidence suggests that the federal appellate courts are not resolving criminal matters as carefully as they once did, in large part because over three-quarters of federal court cases are now handled through decisional shortcuts such as unpublished opinions and surrogate decision-makers, practices that are particularly prevalent in litigation affecting criminal defendants and prisoners.[8] Making matters worse, the Supreme Court has become increasingly less able, or less willing, to resolve circuit court conflicts over criminal law issues.[9]

In short, the trend lines in criminal litigation, while not as dramatic as they are on the civil side, are not moving in the right direction. More importantly, the same courts that are handling the increased criminal caseload also have to deal with a civil caseload that, in proportion to the number of authorized judgeships, is astronomically greater than it was in the 1970s.[10]

One obvious response to this situation is the authorization of more judges. However, that solution is probably a pipe dream in our current politically charged climate, since neither party wants to give the other party a means of dominating the third branch of government.[11] On the assumption that the current impasse on expanding the judiciary is likely to continue,[12] Part II considers a version of a proposal put forward over twenty years ago: creation of a separate federal court system for criminal cases.[13] Under this proposal, the criminal-civil divide would exist at all three levels (trial, intermediate appellate, and ultimate appellate). Litigants could petition the ultimate criminal court for certiorari review, and from that court seek relief at the U.S. Supreme Court. To minimize potential problems posed by an entrenched specialized judiciary, such as capture or judicial ennui, judges could rotate between civil and criminal court at both the trial and intermediate appellate levels.

Such a court system would significantly enhance the capacity of the federal courts to produce quality opinions in the criminal arena, given the greater expertise of a specialized judiciary and the increase in judicial resources that would be devoted to resolving criminal cases. Furthermore, because there would be a superior appellate court devoted to ensuring uniform nationwide rules, such a system could more easily resolve doctrinal conflict on criminal justice issues than the current system, which relies on a Supreme Court that is failing to address most of the differences among the circuits. Perhaps the most important potential benefit of this division of the civil and criminal systems, however, is that the civil system would function more efficiently once criminal cases, which have docket priority at the trial court level, are diverted.[14]

Continuing the reformist agenda, Part III considers a second significant institutional change that could supplement the creation of a separate federal criminal court system: a return to a more indeterminate sentencing regime that would shift much of the heavy lifting regarding criminal dispositions from judges to parole boards. In earlier work, I put forward this proposal for reasons unrelated to trimming judicial assignments, reasons principally having to do with reducing incarceration rates.[15] Here, the proposal is repurposed as a means of lessening the appellate workload and ensuring that trial judges in a specialized criminal court are not debilitated by the psychologically demanding analysis that currently accompanies sentencing.[16]

One definitional point before proceeding: by “criminal” cases, this Article means not only traditional criminal cases trying defendants for crimes but also habeas cases seeking to overturn convictions or sentences and prisoner rights cases alleging unconstitutional prison conditions or treatment. Even though these last two categories of cases are usually considered “civil” in nature,[17] they are closely associated with criminal justice issues and are included in that category in the following discussion.

I. Evaluating Federal Judicial Decision-Making in Criminal Cases

Has the ability of the federal courts to handle criminal cases in a competent manner declined in the past several decades? That question probably cannot be answered directly. But transposing to the criminal arena the methodology that Menell and Vacca use in their survey of the entire federal system, one can get some sense of the answer to this question by looking at changes in the time federal courts take to process cases (efficiency), their ability to resolve conflicts among courts (consistency), and the manner in which they make decisions (quality). That effort is undertaken here, after documenting the main reason the changes are likely to have occurred—a significant increase in caseloads in the federal courts without a corresponding increase in federal judgeships.

A. The Federal Criminal Docket

Menell and Vacca’s data indicate that the average number of traditional criminal cases commenced in federal criminal court went from around fifty thousand during the decade from 1970 to 1980 to about eighty-nine thousand during the decade from 2007 to 2017, despite a fairly significant decrease in criminal case filings in the last four years of the latter decade.[18] This represents an increase of almost 80 percent over a forty-five year period. Habeas and prisoner rights claims increased at an even faster pace, so when they are added into the analysis, the overall change in the number of cases connected to criminal law is very significant, at over 230 percent.[19]

This increase in criminal-related dockets should not be surprising. From 1970 to 2016, the population grew steadily (by about 58 percent[20]), and incarceration rates skyrocketed.[21] Additionally, the explosion in the number of federal criminal statutes,[22] the Warren Court’s expansion of procedural rights,[23] and the passage of the federal sentencing guidelines[24] created entirely new areas of litigation.

Unfortunately, Congress has not seen fit to create a proportionate number of new federal judgeships. Over the past fifty years, the number of authorized full-time judgeships at the court of appeals increased from 97 to 179, an 85 percent change, and the number of authorized full-time judges at the district court level went from 398 to 667, only a 67 percent increase.[25] Obviously, both percentages are well below the increase in the criminal caseload as defined in this Article. Furthermore, even if equilibrium between judgeships and criminal cases had been maintained,[26] the added judges would have had to have dealt with a significant increase in civil cases as well, which typically outnumber criminal cases by a substantial margin.[27] As Menell and Vacca document, the average number of cases filed per judge has increased by 47 percent since 1971, and the average number of cases terminated per judge has increased by 90 percent, with the bulk of the surge coming on the civil side.[28]

The question then becomes whether this extra workload has affected the efficiency, consistency, or quality of federal judicial decisions in criminal cases. The following analysis strongly suggests that, while the pace at which criminal cases are resolved has not changed substantially, unresolved inter-circuit conflicts on criminal issues have increased significantly, and the quality of criminal jurisprudence has noticeably declined.

B. Efficiency

With respect to efficiency, Menell and Vacca report that processing time at the trial level increased by roughly 79 percent between 1972 and 2017.[29] But they do not separate out criminal cases in their analysis. One might conjecture that processing time at the trial level has not increased substantially on the criminal side, given the fact that criminal trial judges must abide by speedy trial rules, the parameters of which have remained fairly constant since 1974.[30] But in fact, the median time from initiation of a traditional criminal case to its termination at the district court level has skyrocketed by more than 200 percent in the past forty-five years, from around three months to over seven months.[31]

The criminal appellate process is not constrained by speedy-trial rules. Nonetheless, here a much different picture emerges. Although data going back to the 1970s are not available, the information we have for the past two decades indicates that the median time it takes to resolve an appellate case—measured from filing the notice of appeal to final disposition—has not changed significantly for either criminal or civil appeals,[32] a tendency that also appears to be true of habeas cases when viewed in isolation, except for those involving capital offenses.[33]

That is not the end of the story, however. The available evidence strongly suggests that the relative stability in criminal case processing times at the appellate level is primarily due to decision-making shortcuts and other adjustments adopted by the federal courts. Those developments are discussed below, after a look at how increasing caseloads have affected doctrinal consistency.

C. Consistency

A second concern about the federal court system that Menell and Vacca pinpoint is that, due to the combined effect of the increase in caseloads, the rise in decisions by subgroupings of the various circuits, and the shrinkage of the Supreme Court docket,[34] uniformity in legal doctrine has suffered. A number of earlier studies suggested that, over the last quarter century, inter-circuit conflicts have become increasingly less likely to be resolved.[35] Similarly, Menell and Vacca’s up-to-date analysis of the civil and criminal dockets leads them to conclude that “there is reason to believe that nearly half of the certiorari-worthy petitions are being denied review today” by the Supreme Court.[36] They also find that even intra-circuit disparity has increased as the ability or willingness of the circuit courts to sit en banc has declined.[37]

One might argue that uniformity is more important in civil cases than in criminal cases, because civil litigants are more likely to have multi-jurisdictional interests that can benefit from interstate consistency and predictability.[38] Further, at least with respect to criminal procedure, Supreme Court pronouncements about the Federal Constitution set the floor rather than the ceiling in the states,[39] and thus variability among states on constitutional criminal law and procedure issues is probably unavoidable to some extent. But those observations should not obscure the countervailing consideration that the confusion and perceived unfairness caused by disparity among circuits work particular mischief in the criminal context, where the courts make momentous decisions regarding deprivations of liberty and control of government power. Recent examples of disarray in criminal doctrine created by the Court include highly consequential rules governing when judges may depart from the sentencing guidelines,[40] the definition of violent crime for sentence enhancement purposes,[41] and when a warrant is needed to obtain information held by a third party.[42]

More importantly, the Supreme Court’s ability or willingness to resolve these types of conflicts is dwindling. In the criminal procedure area, for instance, the number of cases heard by the Court peaked in 1983 and averaged only twenty-two a Term for the three decades after 1990, compared with roughly thirty-seven criminal procedure cases a Term in the three decades before 1990.[43] That reduction in caseload obviously reduces the ability to resolve conflicts in such cases.[44] In 2012, Wayne Logan, after identifying nearly three dozen criminal law and procedure issues on which circuit courts explicitly disagreed, concluded that “the Court regularly fails to reconcile the conflicts, ensuring that the divergent outcomes endure and multiply with the passage of time.”[45] In short, the federal system is less likely now than it was fifty years ago to forge uniform, nationally applicable law in criminal cases.[46]

D. Quality

Of the three metrics assessed here, the quality of judicial decision-making over time is the hardest to gauge. But rough measures of quality at both the district court and appellate levels are possible.

One measure of the quality of district court opinions is reversal rates. While the rate at which circuit courts of appeals reverse district court decisions in criminal cases has remained fairly steady since 2001 (between 4.9 and 7.6 percent), the overall trend has been in the upward direction.[47] Also relevant to the adequacy of trial level determinations is the fact that, compared to a quarter century ago, full-time U.S. magistrate judges now handle a much wider array of criminal dispositions, including felony guilty pleas, misdemeanor trials, pretrial matters, and habeas petitions.[48] This substitution of magistrates for judges has occasioned considerable controversy, largely because magistrates are not Article III judges, but also because of concern that decision-making quality may suffer.[49]

At the appellate level, there is even greater cause for concern. William Richman and William Reynolds point to evidence of what they assert are three proxies for diminished judiciousness in the federal appellate courts: a significant increase in unpublished and summary opinions (which may reduce judicial accountability, given their limited distribution and precedential impact);[50] a smaller proportion of cases going to oral arguments (hypothesized to lead to less careful trial preparation and reduced perceptions of legitimacy among the bar and the public);[51] and a significant expansion in the number and use of law clerks and staff attorneys (who are not meant to be, and perhaps are not qualified to be, surrogate decision-makers).[52] Today over 85 percent of all circuit court of appeals decisions are unpublished,[53] about 75 percent of those decisions were unaided by oral hearing,[54] and staff attorneys (whose number is almost four times what it was in the early 1980s[55]) are largely responsible both for determining which cases are on this “Track Two” and for writing the initial opinions in those cases.[56] Richman and Reynolds argue that, in effect, these developments make today’s federal intermediate appellate courts “certiorari courts.”[57] The result is the creation of two types of appellate cases, one that receives the “Learned Hand treatment” (about 15 to 20 percent of the docket) and one that is handled with a very truncated procedure.[58]

Most pertinent to the adequacy of decision-making in criminal cases, Richman and Reynolds conclude that this second-tier treatment is most likely to occur with a litigant “who is poor, without counsel, and with a boring, repetitive problem,”[59] and they specifically identify prisoners and criminal defendants as paradigmatic of such litigants.[60] Richman and Reynolds label this disparate treatment an “injustice.”[61] It is difficult to disagree with this conclusion.[62]

Another marker suggesting diminished judicial engagement is the federal courts’ greater reliance over time on the harmless error doctrine, which allows them to avoid reversing convictions by finding that, even had the alleged error not occurred, sufficient evidence to convict remains.[63] A survey of federal cases since the 1960s indicates that judicial discussions of harmless error in criminal cases increased over tenfold from the ten-year period leading up to 1975 to the ten-year period leading up to 2019, an increase that amounts to more than three times the pace of new federal criminal filings.[64] This finding could be due to a number of factors: the application of harmless error analysis to a larger range of claims;[65] a greater willingness of courts to find error; or stronger, more difficult-to-overturn cases.[66] But certainly a contributor is the fact that harmless error doctrine is a convenient way to avoid costly retrials (and possible re-appeals).[67]

In sum, the federal court system appears to be falling short relative to its performance fifty years ago, even when the focus is limited to criminal cases. Efficiency, when measured in terms of the time it takes to process a case, appears to have suffered appreciably at the trial level, undoubtedly due in part to the increased caseload in the federal courts. Processing time at the appellate level does not seem to have increased, but that achievement may have come at the cost of a loss in consistency and quality of justice. Justice may not be significantly delayed, and it may not be completely denied, but the litigants seeking it could well be short-changed. On the assumption that this situation warrants some sort of reform of the federal court system, the rest of this Article explores two proposals that might simultaneously improve the efficiency, consistency, and quality of decision-making in criminal cases: a federal court system devoted to criminal cases, and a return to sentencing that relies on parole boards rather than judges.

II. A Separate Federal Criminal Court System

The idea of a separate federal criminal court at the trial and appellate levels is not new. In 1996, Victor Williams proposed that the federal court system be separated into civil and criminal divisions.[68] Under Williams’s plan, the criminal division would consist of U.S. District Criminal Courts, U.S. Courts of Criminal Appeals (shadowing the current geographic distribution of circuit courts), and a National Court of Criminal Appeals.[69] This court system would be responsible for hearing all criminal cases,[70] criminal justice-related Bivens and § 1983 claims, and habeas claims involving federal issues.[71] The National Court of Criminal Appeals, composed of seven members, would have transfer jurisdiction over cases certified from the U.S. Courts of Criminal Appeals as well as the authority to grant certiorari review of cases decided by those courts.[72] The U.S. Supreme Court would have certiorari review of cases decided by the National Court of Criminal Appeals,[73] which is arguably a constitutional requirement, given Article III’s stipulation that there be “one [S]upreme Court.”[74] Judges in the new system would come from the current district courts and circuit courts of appeals, as well as from state courts and from the criminal prosecution and defense bars.[75] Williams claimed that such an organization would better “ensure the speedy and fair adjudication of federal criminal cases” and “promote the development of a motivated and experienced bench of expert criminal procedure and criminal law jurists.”[76]

This organization of the court system would not only be new at the federal level, but it would also be different from any current state court regime. Although Oklahoma and Texas each have a criminal court of appeals that is the final arbiter of criminal issues, and Alabama, New York, and Tennessee have a trial court or appellate court division devoted entirely to criminal cases, no state has the type of three-level criminal court proposed by Williams.[77] Every other state court mixes civil and criminal jurisdiction at both the trial and appellate levels (albeit with some trial judges assigned solely to the criminal or civil docket).[78]

In contrast, in several European countries, designated criminal courts are common. In France, for instance, criminal cases are heard in tribunaux devoted entirely to criminal cases and are appealed to criminal appeals courts.[79] In Germany, trial courts hear both criminal and civil cases, but specialized criminal courts hear appeals of criminal cases in each of Germany’s judicial divisions.[80]

Ideally, some empirical method of figuring out whether any of these specialized criminal courts are more efficient, consistent, and capable than their generalist counterparts could be devised. But no such studies have been done.[81] The number of variables that would need to be held constant to obtain meaningful results, and the difficulties of defining what those variables should be, are daunting. The following discussion is admittedly somewhat speculative on all three fronts.

A. Advantages

Presumably, a specialized criminal court system would be more expedient than the current generalist system, an advantage that could have particularly significant payoff at the trial level, where processing times have exploded. Judges in such a system would be more familiar with the rules and doctrine governing criminal cases, which in some respects diverge significantly from analogous civil law.[82] Of course, in a generalist system, a judge with no prior experience in criminal cases will over time acclimate to the environment, aided by memoranda and briefs from the parties. But judges in the proposed system will handle two to three times the number of trials, habeas hearings, guilty pleas, and sentencings that judges currently handle,[83] which means that the necessary experience and wisdom will come more quickly and that lessons learned will not need to be relearned because of intervening civil matters.[84]

The proposal’s effect on the efficiency of civil cases is also important. Because of the Speedy Trial Act and local rules, criminal cases have priority at the trial level of the federal court system.[85] Many commentators have noted that the criminal docket is a dominant, or perhaps even the primary, reason for delays in civil cases.[86] With criminal cases removed, that backlog would abate.

A specialized criminal court would also promote doctrinal consistency. Compared to the current system, which relies on the whims of the U.S. Supreme Court, a National Court of Criminal Appeals would vastly increase the capacity to resolve conflicts within the lower courts. Even if the National Court mimicked the same reluctance to decide cases currently evidenced by the Supreme Court, it would be resolving more than three times as many criminal cases as the Court currently does, since such cases comprise roughly one-third of the Supreme Court’s docket.[87] That additional capacity would also probably substantially reduce the need for the en banc process, which the circuits have introduced in an effort to reduce disparity, but which has been cumbersome and problematic in many other ways.[88] And on the off chance that Congress decided to fund additional judgeships and the influx of new arbiters caused, as some have surmised,[89] increased decisional inconsistency, that problem could more easily be addressed.[90]

For similar reasons, a federal criminal court would probably improve the quality of decisions in cases involving criminal litigation, at both the trial and appellate levels. At the trial level, judges who preside over only criminal adjudications cannot help but be more adept at handling pretrial and evidentiary rulings, making decisions about instructions, and understanding the nuances of sentencing. Reliance on magistrates could also be substantially reduced.

At the appellate level, a specialized criminal court system would minimize the likelihood of poorly reasoned opinions from novices to the field, and maximize the probability of persuasive precedent written by judges who are confident of their grasp of the subject.[91] More concretely, a National Court of Criminal Appeals and specialized intermediate appellate courts would significantly enhance judicial capacity to hear criminal appeals, habeas claims, and prisoner rights cases without resorting to unpublished opinions.[92] It would also reduce reliance on staff attorneys as surrogate decision-makers who, in any event, would presumably be better at their job because of their concentration on criminal cases.[93]

Some have argued against a fourth level of court like the proposed National Court of Criminal Appeals on the ground that it would create an additional layer of review that litigants must negotiate and a new body of caselaw that they must master.[94] But the same certiorari papers aimed at obtaining a hearing at the National Court of Criminal Appeals could, if relief is denied by that court, be used to petition the Supreme Court. Moreover, the increased resolution of conflicts that the new court would occasion should reduce the welter of caselaw that currently confronts those trying to ascertain the merits of their case. At the same time, the existence of the National Court would lighten the Supreme Court’s burden, which could then devote more attention to truly significant cases.

Finally, it is worth noting that a federal court system of this sort, specifically designed to handle criminal cases, would track practice in every other area of the law. Lawyers increasingly tend to specialize, especially in the criminal arena, where public defenders are generally considered to be more qualified than private attorneys who represent indigent clients.[95] State court systems are also moving toward more specialization in the criminal domain.[96] At the federal court level, Edward Cheng’s research suggests that circuit courts already recognize the value of having judges concentrate on particular areas by relying on an informal opinion-assignment process that tends to favor certain judges with certain types of cases.[97] The advantage of the proposed system would be that all panelists would be experts, so that no one judge could dominate, or routinely be assigned to write, the “criminal” opinions.[98]

For the likely readers of this Article, however, perhaps the most persuasive argument for a specialized court pertains to the legal academy: no professor of criminal law or procedure would even consider teaching every subject in the curriculum. Yet we ask federal judges to perform an analogous function. Specialization is key to mastering any subject.

B. Criticisms

The advantages of a federal criminal court system are sizeable. But criticisms of such a regime have also been numerous, perhaps best summarized by Judge Diane Wood. While acknowledging that “[j]udges in most other countries are often staggered by the breadth of the American federal judge’s writ,”[99] Judge Wood still enthusiastically supported the current system, claiming that generalist judges bring advantages that specialists do not. As she put it:

Generalist judges cannot become technocrats; they cannot hide behind specialized vocabulary and “insider” concerns. The need to explain even the most complex area to the generalist judge (and often to a jury as well) forces the bar to demystify legal doctrine and to make the law comprehensible.[100]

Judge Wood also argued that a generalist docket encourages the cross-fertilization of ideas and ensures a fresh look at entrenched procedures by judges who have not been co-opted by the system.[101]

Related to this last point is the fear of capture—the idea that, given the increasingly political nature of the selection process, judgeships will be populated by individuals who adhere to the values of groups the courts are supposed to monitor.[102] This could be a particular problem in criminal cases. Even in our current generalist regime, 43 percent of judges are ex-prosecutors, which is four times higher than the proportion of judges who are ex-public defenders.[103] The possibility that this fraction would increase in a specialized court raises the specter of a system even more draconian than the current one.

A third general worry about specialization, especially at the circuit court level, is that a steady diet of similar cases in a court now perceived to be “secondary” will lead to judicial boredom and loss of judicial prestige. Perhaps neither concern is significant in itself. But some have posited that these consequences are not only likely, but will also lead to lower-quality opinions and dissuasion of the most qualified applicants.[104]

While these objections might have force with respect to some types of specialized courts, they are weak in the criminal context. Criminal law is decidedly less technocratic than expert-infested areas such as antitrust, the primary example Judge Wood gives of legalese.[105] Nor is its content mundane or narrow; it raises a host of important constitutional issues, ranging from the scope of the Fourth Amendment to the definition of cruel and unusual punishment, as well as challenging non-constitutional issues involving interpretations of criminal statutes and common law doctrines. When habeas and class action suits brought by prisoners and detainees are added to the mix, civil as well as criminal procedure rules would be in play.[106] The likelihood of capture is also minimal; indeed, capture might actually be reduced in a specialized system because, in contrast to the current situation, the defense bar will know which judicial positions will focus on the criminal docket and can concentrate their lobbying power there.[107] In any event, the assumption that ex-prosecutors will generally be anti-defendant is false (consider Justice Sonia Sotomayor, for example).[108] And unlike the civil arena, where high-paying law firm jobs beckon, there are literally thousands of well-qualified—but not well-compensated—prosecutors and defense attorneys who would find both the work and the pay associated with a federal judgeship highly attractive.

If, however, concerns about capture, tunnel vision, boredom, and the like persist, they could be alleviated by a rotation system of the type routinely practiced in many states.[109] Judges who migrate from civil to criminal court and back again will have both expertise and diverse experiences. At the same time, to ensure the necessary expertise develops, the rotation should be for a fairly long period of time, perhaps two to three years.[110]

Of course, if the proposed federal criminal court system were seriously understaffed, no amount of experience or expertise could compensate. But as suggested in Part I and confirmed by Menell and Vacca’s data,[111] assuming current criminal and civil caseloads, a federal criminal court system needs only about two-thirds the number of judges that the civil docket requires.[112] This ratio also means that not every civil-side judge would need to rotate through the criminal system, and criminal-side judges could always be assured of a job on the civil side when their rotations end.

III. Sentencing by Parole Board

If the goal is to ensure judicial capacity and expertise adequate to handle criminal cases in the federal court system, a second reform—one that could be integrated into either the proposed federal criminal justice system or the current system—should target sentencing practices. Today’s federal courts labor under the much-maligned federal sentencing guidelines, which were driven by a desire to ensure uniformity in sentencing and, to a lesser extent, a desire to implement a just deserts philosophy.[113] In other work, I have proposed a quite different approach to sentencing which, although not originally conceived as a palliative to the problems addressed here, could have the effect of reducing judicial workload at both the appellate and the trial court level.[114]

In brief, the proposal is to rejuvenate and modernize the indeterminate sentencing regime that once existed throughout the country. Indeterminate sentencing was abolished in the federal system in 1984[115] and has been abandoned in a number of states as well, replaced by sentencing frameworks that rely on narrow dispositional ranges determined primarily by an offender’s perceived desert or blameworthiness.[116] In revised form, however, indeterminate sentencing is worth reconsidering for a number of reasons.

In the regime that I have proposed, legislatures or sentencing commissions would set relatively broad dispositional ranges based on desert concerns, but ultimate decisions about sentence length would be left to a parole board.[117] For instance, the legislature might provide that a person convicted of armed robbery should receive a sentence of one to ten years, a range that the judge would impose upon conviction. But the date of release within that range would depend on a parole board’s evaluation of the offender’s risk of reoffending and various other factors, and might vary significantly for people convicted of the same offense. While previous manifestations of parole boards, often staffed by political appointees, have been rightly criticized for their uninformed decision-making and unwillingness to take chances on releasing prisoners early,[118] these concerns can be addressed by ensuring that parole boards are composed of professionals who make use of modern risk assessment and risk management techniques and are guided by pre-set criteria.[119]

The original goal motivating this proposal was to reduce our egregious mass incarceration problem by improving the ability to identify low-risk offenders who do not need to be confined for long periods, increasing alternatives to prison that are better at minimizing recidivism, and reducing prosecutorial power over dispositions.[120] However, another beneficial by-product of the proposed system is that it would significantly decrease the number of appeals to the circuit courts. Since the advent of the federal sentencing guidelines, appeals have made up a large portion of the federal criminal appellate docket.[121] If, instead, parole boards were in charge of sentencing, trial court dispositions would be the subject of challenge much less frequently.[122]

This reform could also benefit trial judges. Because sentencing decisions involve deprivations of liberty, they are arguably the most consequential determinations a trial judge makes. More than one judge has expressed dread about making such decisions.[123] If judges heard only criminal cases, as proposed in this Article, they might become even more inured to familiar-sounding excuses and pleas for leniency or more enervated by the psychological demands of doing justice.[124] However, that would be true only if the federal courts continued to abide by the federal sentencing guidelines which, especially in their current “voluntary” form, leave complex and highly charged decisions about retribution, deterrence, and prevention to the judge.[125]

Under the system I propose, the judge would have a much-diminished role at sentencing. The nature of the offense, not the judge, would automatically place the defendant in one of five or so ranges for felonies (for example, from one to twenty years, one to ten years, one to five years, one to three years, and one to two years).[126] The specific length of sentence would depend, as it once did in virtually every jurisdiction,[127] on the parole board’s assessment of the offender’s progress in rehabilitative programs and similar considerations. Because the sentencing decision in a parole-driven regime involves simply matching the crime of conviction with the legislatively authorized range, it would not burden the judge with difficult and psychologically taxing decisions about how much punishment is “deserved,” how “dangerous” the offender is, and how long a sentence is needed to ensure “deterrence.”

Conclusion

A separate federal criminal judiciary, together with a modernized indeterminate sentencing regime, could be the beginning of a formula for streamlining the criminal justice process, ensuring high-quality and more consistent judicial decision-making, and reducing our reliance on prison. These are not the only promising criminal justice reforms, of course; limiting federal criminal jurisdiction, de-criminalization, and shortening sentences across the board would have similar benefits. But the reforms proposed here have received much less attention.

Are the changes proposed in this Article politically feasible? As Menell and Vacca demonstrate, many federal judges are dead-set against a substantial increase in judgeships, dividing up the federal court system into specialty courts, or any other type of radical change, and Congress has heeded their resistance.[128] And the dominant sentencing philosophy in both Congress and the Sentencing Commission could continue to be retribution,[129] a stance that is antithetical to indeterminate sentencing (which is mostly focused on specific deterrence and rehabilitation).[130] In particular, parole boards may still be anathema to most policy-makers.[131]

At the same time, the federal court system has not been entirely hostile to specialization in criminal-related areas. In the past decade, the federal system has experimented with reentry courts, drug courts, and other problem-solving courts (with over sixty-five established as of 2015).[132] The motivation for creating these types of courts is also more consistent with the individual prevention and treatment goals of indeterminate sentencing than with retribution.[133] On both sides of the aisle, there is significant momentum against the harshness of our criminal justice system and the over-use of imprisonment, an attitude demonstrated by the recently passed First Step Act,[134] which bets heavily on risk assessment and management tools in determining who should be imprisoned, for how long, and under what programs.[135]

These trends point in the right direction. If they don’t inspire the types of changes proposed in this Article in the near future, perhaps they will by 2030, when the reforms proffered at this Symposium are meant to go into effect. Without major structural change in the federal court system of some sort now or in the near future, criminal defendants will continue to pay the price for overworked courts and overflowing prisons.

DOI: https://doi.org/10.15779/Z38H41JN2H.

Copyright © 2020 Christopher Slobogin, Milton Underwood Professor of Law, Vanderbilt University. The author would like to thank his Vanderbilt colleagues, especially Ed Cheng and Nancy King, for their feedback, and Laural Hooper and Kristin Garri of the Federal Judicial Center and Professors Ryan Vacca and Michael Heise for their assistance in tracking down relevant data.

2. Peter S. Menell & Ryan Vacca, Revisiting and Confronting the Federal Judiciary Capacity “Crisis”: Charting a Path for Federal Judiciary Reform, 108 Calif. L. Rev. 789, 795 (2020) (providing data about the federal courts that show “expanding caseloads and growing complexity and fragmentation of federal law”). ↑

3. Id. at 872 (“The bottleneck at the top of the judiciary pyramid has become more constricting. The lack of clarity of the law, in conjunction with the high stakes of litigation and relatively low cost of appeal, fuels spiraling litigation and undermines economic, social, and political decision-making and institutions.” (footnotes omitted)). ↑

4. Id. at 873 (“[T]here is reason to believe that nearly half of the certiorari-worthy petitions are being denied review today.”). ↑

5. See, e.g., id. at 815–19 (recounting the findings of the Hruska Commission, which published one report in 1973 and another in 1975 detailing a dramatic increase in caseloads per district court judge, the shortcuts taken by appellate courts in the 1960s and early 1970s, and the need for resolution of inter-circuit conflicts). ↑

6. Id. at 795, 879–80 (proposing a “2030 Commission” charged with recommending reforms that would not go into effect until 2030, in an effort to avoid recommendations tainted by partisan politics). ↑

7. See infra text accompanying notes 17–27. ↑

8. See infra text accompanying notes 48–56. ↑

9. See infra text accompanying notes 33–45. ↑

10. See Menell & Vacca, supra note 1, at 845 (“The annual number of cases terminated has increased 465 percent from 62,955 in 1970 to 355,706 in 2017. Civil terminations have driven much of this increase . . . .”). ↑

11. See id. at 794 (“Judiciary reform has become a legislative third rail, too dangerous for politicians . . . to discuss.”). ↑

12. See id. at 796 (“Despite substantial hand-wringing and studies recommending reform, . . . Congress last increased the number of judicial slots in the early 1990s.”); see also infra note 127. ↑

13. See infra text accompanying notes 67–75 (explaining Victor Williams’s 1996 proposal that the federal court system be separated into civil and criminal divisions). ↑

14. See infra text accompanying notes 84–85 (contextualizing the precedence that criminal cases take over civil matters in federal district court). ↑

15. See, e.g., Christopher Slobogin, A Defense of Risk-Based Sentencing, in Predictive Sentencing: Normative and Empirical Perspectives 107 (Jan W. De Keijser et al., eds., 2019). ↑

16. See infra text accompanying notes 120–26. ↑

17. Wayne R. LaFave et al., Criminal Procedure 139 (3d ed. 2000) (“Most of today’s post-conviction proceedings are viewed as independent civil actions, but some are considered part of the original criminal case, similar to a post-appeal motion for a new trial.”). ↑

18. See Menell & Vacca, supra note 1, at 844 fig.2. ↑

19. Some approximations are necessary to achieve this figure. There were 77,128 traditional criminal case filings in 2017, see Menell & Vacca, supra note 1, at 844 fig.2, and 73,725 habeas and prisoner rights action filings in that year, see Admin. Office of the U.S. Courts, Judicial Business of the United States Courts: Table C-2 (2017), https://www.uscourts.gov/sites/default/files/data_tables/jb_c2_0930.2017.pdf [https://perma.cc/428V-G64X], for a total of 150,853 criminal filings in 2016. There were 49,416 traditional criminal case filings in 1970, Menell & Vacca, supra note 1, at 844 fig.2, and 15,997 habeas and prisoner rights action filings in that year, see Bureau of Justice Statistics, U.S. Dep’t of Justice, Federal Review of State Prisoner Petitions: Habeas Corpus 2 tbl.1 (1984), https://www.bjs.gov/content/pub/pdf/hc-frspp.pdf [https://perma.cc/3XP9-A9YW], for a total of 65,413 criminal filings in that year. 150,853 is roughly 231 percent of 65,413. ↑

20. See US Population by Year, Multpl.com, https://www.multpl.com/united-states-population/table/by-year [https://perma.cc/4V6C-VAGZ]. ↑

21. A conservative estimate is that the United States’ incarceration rate rose from approximately one hundred people per hundred thousand in the late 1960s to somewhere between five and six hundred people per hundred thousand today. See Franklin E. Zimring, Is There a Remedy for the Irrelevance of Academic Criminal Law?, 64 J. Legal Educ. 5, 7 fig.1 (2014). ↑

22. James A. Strazzella, Am. Bar Ass’n, The Federalization of Criminal Law 5 (1998), https://www.americanbar.org/content/dam/aba/publications/criminaljustice/Federalization_of_Criminal_Law.authcheckdam.pdf [https://perma.cc/4U5Y-NPF4] (“Congressional activity making essentially local conduct a federal crime has accelerated greatly . . . .”). ↑

23. See, e.g., Fed. Courts Study Comm., Report of the Federal Courts Study Committee 5 (1990), https://www.fjc.gov/sites/default/files/2012/RepFCSC.pdf [https://perma.cc/4LRR-VVYP] (referring to “the creation of many new federal rights” as one reason for the uptick in federal cases). ↑

24. See Nancy J. King & Michael O’Neill, Appeal Waivers and the Future of Sentencing Policy, 55 Duke L.J. 209, 227–28 (2005) (finding, based on a survey of 971 cases, that “the rate of appeals per conviction peaked in 1994 at about double the rate prior to 1987 (when the Sentencing Reform Act became effective)” but also noting that the rate of appeals “has consistently declined since then,” a development that “coincides with the increased enforcement of waivers by the courts of appeals”). ↑

25. See Admin. Office of the U.S. Courts, U.S. Courts of Appeals: Additional Authorized Judgeships, https://www.uscourts.gov/sites/default/files/appealsauth.pdf [https://perma.cc/3P9S-48QQ] (tracking the number of authorized federal court of appeals judgeships from year to year, excluding temporary judgeships); Admin. Office of the U.S. Courts, U.S. District Courts: Additional Authorized Judgeships [hereinafter U.S. District Courts: Additional Authorized Judgeships], https://www.uscourts.gov/sites/default/files/districtauth.pdf [https://perma.cc/HC7U-PDWH] (same but for federal district court judgeships). ↑

26. Given an increase of fifty-eight district court judgeships in 1970, see U.S. District Courts: Additional Authorized Judgeships, supra note 24, had 1969 been used as the base date, the increase in judges by 2017 would have been 98 percent. ↑

27. For instance, in 2016, a typical year, there were 215,742 “pure” civil filings (i.e., excluding prisoner and habeas petitions) and 152,747 combined criminal filings and prisoner and habeas petitions, see Admin. Office of the U.S. Courts, Judicial Business of the United States Courts: Table C-2 (2016), https://www.uscourts.gov/sites/default/files/data_tables/stfj_c2_1231.2016.pdf [https://perma.cc/9N3Z-TJ5Y]; Admin. Office, U.S. Courts, Judicial Business of the United States Courts: Table D (2016), https://www.uscourts.gov/sites/default/files/data_tables/stfj_d_1231.2016.pdf [https://perma.cc/TZB9-L8HU], which means that the number of criminal filings as defined in this article amounted to only about two-thirds the number of civil filings, and constituted about 40 percent of the total number of cases filed (368,489). ↑

28. Menell & Vacca, supra note 1, at 848. ↑

29. Id. at 851. ↑

30. The federal Speedy Trial Act, which requires that federal district courts establish a plan for commencing criminal trials within one hundred days of arrest or service of summons, went into effect in 1974. See 18 U.S.C. §§ 3161(b), (c)(1). ↑

31. E-mail from Ryan Vacca to author (Apr. 6, 2020) (on file with author) (including table showing that median processing time for criminal cases was 92 days for the five years from 1972 to 1976 and 219 days for the five years from 2013 to 2017). ↑

32. According to the Administrative Office of the U.S. Courts, the median period from notice of appeal to disposition in both criminal cases and civil (including habeas and prisoner) cases in 1997 was 11.4 months, 12.2 months in 2007, and 8.9 months in 2017. See Admin. Office of the U.S. Courts, Judicial Business of the United States Courts, Table B-4 (1997), https://www.uscourts.gov/sites/default/files/statistics_import_dir/b04sep97.pdf [https://perma.cc/4KWV-YTCT]; Admin. Office of the U.S. Courts, Judicial Business of the United States Courts, Table B-4 (2007), https://www.uscourts.gov/sites/default/files/statistics_import_dir/B04Sep07.pdf [https://perma.cc/994F-EYFZ]; Admin. Office of the U.S. Courts, Judicial Business of the United States Courts, Table B-4 (2017), https://www.uscourts.gov/sites/default/files/data_tables/jb_b4_0930.2017.pdf [https://perma.cc/BM5S-QMTU]. ↑

33. Lisa M. Seghetti et al., Cong. Research Serv., Federal Habeas Corpus Relief: Background, Legislation, and Issues 8 fig.3 (2007), https://www.everycrsreport.com/files/20070112_RL33259_ed02b2aa1c0e936bb1cad872e7532e9486ad28f0.pdf [https://perma.cc/CBS5-US4N]. ↑

34. In the 2016 Term, the Supreme Court released fewer decisions than at any time since the 1860s, and the trend has been downward since the 1970s. See Alan Feldman, Looking Back to Make Sense of the Court’s (Relatively) Light Workload, Empirical SCOTUS (Jan. 9, 2018) https://empiricalscotus.com/2018/01/09/light-workload [https://perma.cc/7WD9-P955]. ↑

35. See Menell & Vacca, supra note 1, at 868–69. ↑

36. Id. at 873. ↑

37. Id. at 860 fig.22 (indicating a roughly 300 percent decrease in en banc merits determinations since the 1970s). ↑

38. Cf. Payne v. Tennessee, 501 U.S. 808, 828 (1991) (“Considerations in favor of stare decisis are at their acme in cases involving property and contract rights, where reliance interests are involved; the opposite is true in cases . . . involving procedural and evidentiary rules.” (citations omitted)). ↑

39. See Lego v. Twomey, 404 U.S. 477, 489 (1972) (confirming the states’ ability to provide more protection to criminal defendants under their own constitutions). State courts have been quite quick to invoke this authority in criminal procedure cases. See Robert M. Howard et al., State Courts, the U.S. Supreme Court, and the Protection of Civil Liberties, 40 Law & Soc’y Rev. 845 (2006) (examining state court willingness to provide more protection to criminal defendants under state constitutions). ↑

40. In 2009, Hessick and Hessick stated, “So long as the Court seeks to promote the two conflicting Booker goals of discretion and uniformity in an ad hoc way, the law of appellate sentencing review is likely to remain muddled and in a state of turmoil.” Carissa Byrne Hessick & F. Andrew Hessick, Five Years of Appellate Problems After Booker, 22 Fed. Sent’g Rep. 85, 88 (2009). That assessment remains true today. ↑

41. Caryn Devins, Lessons Learned from Retroactive Resentencing After Johnson and Amendment 782, 10 Fed. Cts. L. Rev. 39, 107 (2018) (“In the wake of Johnson [v. United States], significant circuit splits and legal uncertainty have arisen over the applicability of the remaining clauses of the [Armed Career Criminal Act, which defines “violent felony” for sentencing purposes].”). ↑

42. Because the Court only recently decided Carpenter v. United States, 585 U.S. __, 138 S. Ct. 2206 (2018), which appears to upend the Fourth Amendment third-party doctrine, it remains to be seen whether, and how quickly, the Court will resolve the questions the case inevitably raises. See Susan Freiwald & Stephen Wm. Smith, The Carpenter Chronicle: A Near-Perfect Surveillance, 132 Harv. L. Rev. 205, 227–31 (2018) (projecting the possible impact of the Carpenter decision in the increasingly complex telecommunications arena). ↑

43. Analysis Specifications - Modern Data (1946-2018), Wash. Univ. Law, Sup. Ct. Database, http://supremecourtdatabase.org/analysis.php [https://perma.cc/H2DB-M5CF] (enter “1960” through “2018” as the “Range of Terms” under the “TIME / ERA” parameter and check “Criminal Procedure” as the “Issues” under the “CASE COMPONENTS” parameter). ↑

44. See, e.g., Arthur D. Hellman, Fed. Jud. Ctr., Unresolved Intercircuit Conflicts 67 (1991), https://www.fjc.gov/sites/default/files/2015/Unresolved%20Intercircuit%20Conflicts%20Phase%20I%20Hellman%201991.pdf [https://perma.cc/KS8T-NNQG] (finding that, of the cases involving conflicts that were denied certiorari by the Supreme Court, one-third involved criminal issues). Many criminal-related petitions are in forma pauperis, and Menell and Vacca suggest that in forma pauperis petitions are not normally considered cert-worthy. Menell & Vacca, supra note 1, at 866 n.420. But Hellman found that even within this category, eleven out of ninety-three, or approximately 10 percent, of the petitions alleged conflicts that had been acknowledged by at least one court or other participant in the system. Hellman, supra, at 53–54; see also Wendy L. Watson, The U.S. Supreme Court’s In Forma Pauperis Docket: A Descriptive Analysis, 2 Just. Sys. J. 47, 47 (2006) (finding that such petitions “are not categorically frivolous and unimportant” and suggesting that they “require further consideration in the literature on the Supreme Court’s agenda setting”). ↑

45. Wayne A. Logan, Constitutional Cacophony: Federal Circuit Splits and the Fourth Amendment, 65 Vand. L. Rev. 1137, 1141 (2012). ↑

46. However, it should be noted that some have argued this state of affairs is desirable. See, e.g., Paul A. Holton, Comment, The “Do Nothing” Court: Why A Reduced Supreme Court Docket Is a Good Thing, 50 U.S.F. L. Rev. 291, 301 (2016) (arguing that because they apply facts to law more frequently than the Supreme Court, circuit courts are better equipped to handle most cases). ↑

47. See Statistical Tables for the Federal Judiciary, U.S. Courts, https://www.uscourts.gov/statistics-reports/analysis-reports/statistical-tables-federal-judiciary [https://perma.cc/N69Q-Q2X9] (showing in various tables that in the four years from 2001 to 2004, the average reversal rate was 5.5 percent, whereas with the exception of 2006, 2010, and 2013, the reversal rate has been over 6 percent every year since 2004 through 2019). ↑

48. See generally Douglas A. Lee & Thomas E. Davis, “Nothing Less than Indispensable”: The Expansion of Federal Magistrate Judge Authority and Utilization in the Past Quarter Century, 16 Nev. L.J. 845 (2016) (examining the expansion of the role of magistrate judges over the preceding twenty-five years). ↑

49. The controversy surfaced in Peretz v. United States, which in essence reversed Gomez v. United States, 490 U.S. 858 (1989), by holding that magistrates may, with the consent of the parties, preside over felony jury selection. See 501 U.S. 923, 941 (1991) (Marshall, J., dissenting) (“By discarding Gomez’s categorical prohibition of magistrate felony jury selection, the majority unnecessarily raises the troubling question whether this practice is consistent with Article III of the Constitution. To compound its error, the majority resolves the constitutional question in a manner entirely inconsistent with our controlling precedents.”); see also Kimberly Anne Huffman, Note, Peretz v. United States: Magistrates Perform Felony Voir Dire, 70 N.C. L. Rev. 1334, 1361 (1992) (“Peretz thus marks the beginning of an expansion of magistrate authority, often at the expense of litigants’ rights and constitutional protections, and without regard to whether consent to this authority is express or implied.”). ↑

50. William M. Richman & William L. Reynolds, Injustice on Appeal: The United States Courts of Appeals in Crisis 42–55 (2012). The authors note, for instance, that Judge Kozinski, “an outspoken supporter of limited publication,” has nonetheless stated that unpublished opinions are “not safe for human consumption.” Id. at 44. ↑

51. Id. at 85–90. ↑

52. Id. at 97–114. ↑

53. Id. at 38; see Admin. Office of the U.S. Courts, Judicial Business of the United States Courts: Table B-12, https://www.uscourts.gov/sites/default/files/data_tables/jb_b12_0930.2018.pdf [https://perma.cc/7T59-G3EU] (showing that 88.2 percent of all circuit court opinions during the twelve-month period ending September 30, 2018, were unpublished and that roughly 82 percent of these were unsigned). ↑

54. Richman & Reynolds, supra note 49, at 92. ↑

55. Id. at 112. ↑

56. Id. at 164. ↑

57. Id. at 118. ↑

58. Id. at 163. The authors use the phrase “Learned Hand treatment” to describe opinions signed by a judge after full briefing and oral arguments. Id. at 119–20. ↑

59. Id. at 120. ↑

60. . Id. ↑

61. Thus, their title, Injustice on Appeal. Richman & Reynolds, supra note 49. ↑

62. See Standards Relating to Appellate Courts § 3.10, at 15, § 3.19 (Am. Bar Ass’n 1977) (stating that “thoughtful consideration of the merits by at least three judges” is a basic element of an “appeal of right” and that overuse of appellate clerks and staff undermine this right); Robert L. Stern, Appellate Practice in the United States 17–18 (2d ed. 1989) (similarly arguing that judges’ expanded reliance on staff “may have the effect of substituting the staff attorney for the judge as the actual decider of the case”); Marin K. Levy, Judicial Attention as a Scarce Resource: A Preliminary Defense of How Judges Allocate Time Across Cases in the Federal Court of Appeals, 81 Geo. Wash. L. Rev. 401, 422, 436 (2013) (calling a two-track system “disquieting,” although also noting second-track treatment may be justified for cases that involve repetitive problems, and specifically mentioning sentencing appeals and bail appeals as examples). Certainly, today’s criminal case resolution process is quite different from the process judges and legislators envisioned fifty years ago. See Catherine T. Struve, The Federal Rules of Inmate Appeals, 50 Ariz. St. L.J. 247, 257 (2018) (recounting judicial and congressional efforts in the 1960s to ensure “equal treatment of poor litigants, especially in the criminal context”). ↑

63. See Chapman v. California, 386 U.S. 18, 23–34 (1967) (allowing courts finding constitutional error to avoid reversal of a conviction if they find “beyond a reasonable doubt that the error complained of did not contribute to the verdict obtained”). ↑

64. See Federal Cases, Westlaw Edge, https://1.next.westlaw.com/Browse/Home/Cases/FederalCases (in search bar, enter “advanced: DA(aft 12-31-2009 & bef 01-02-2019) & “harmless error” & criminal”); id. (in search bar, enter “advanced: DA(aft 12-31-1964 & bef 01-02-1975) & “harmless error” & criminal”). Note that, given these search terms, the query may have picked up mostly “traditional” criminal cases (exclusive of habeas and prisoner cases), which only increased by about 80 percent during the period considered, see supra text accompanying notes 17–18, in comparison to the over 1,000 percent increase in harmless error cases, see Federal Cases, supra. ↑

65. See Amy Knight Burns, Note, Counterfactual Contradictions: Interpretive Error in the Analysis of AEDPA, 65 Stan. L. Rev. 203, 215 (2013) (“[O]ver the last fifty years, the trend has been to expand the class of cases in which harmless error analysis applies . . . .”). ↑

66. One study of appeals to the federal circuits from 2011 through 2016 found that roughly 38 percent of alleged errors by the government and 8 percent of alleged errors by the defense were found to be harmless in those cases with a “reasoned” opinion. Nancy J. King & Michael Heise, Appeals by the Prosecution, 15 J. Empirical Legal Stud. 482, 514 (2018). Presumably, the harmless error rate is even higher in those cases that do not involve reasoned opinions. ↑

67. See Daniel J. Meltzer, Harmless Error and Constitutional Remedies, 61 U. Chi. L. Rev. 1, 4 (1994) (“[T]here is a widespread perception that in the Supreme Court, as well as in state and lower federal courts, errors of some substance are nonetheless found harmless so as to permit the affirmance of convictions.”); see also Brandon L. Garrett, Convicting the Innocent 200–01 (2011) (finding that courts often relied on findings of harmless error in affirming convictions of defendants later found to have been wrongly convicted). ↑

68. Victor Williams, A Constitutional Charge and a Comparative Vision to Substantially Expand and Subject Matter Specialize the Federal Judiciary: A Preliminary Blueprint for Remodeling Our National Houses of Justice and Establishing a Separate System of Federal Criminal Courts, 37 Wm. & Mary L. Rev. 535 (1996). ↑

69. Id. at 658. ↑

70. Some types of cases—such as those involving special commitment of sex offenders, see United States v. Comstock, 560 U.S. 126 (2010)—are quasi-criminal in nature, and others—such as securities cases—could trigger criminal, administrative, or civil sanctions, or all three. While the civil judiciary could conceivably have jurisdiction over these types of cases, the better, and easier, determinant of whether they end up on the criminal side of the federal docket should be whether the action is initiated by the criminal division of the Department of Justice. In other words, the executive branch would decide where the case would be heard. ↑

71. Williams, supra note 67, at 658. ↑

72. Id. at 668. ↑

73. Id. at 668–69. ↑

74. U.S. Const., art. III, § 1. This feature would also help avoid competing rulings on issues that affect both civil and criminal cases, such as rules regarding expert testimony, a problem said to afflict the Texas system, whose Court of Criminal Appeals is the ultimate arbiter of criminal law issues. See Ben L. Mesches, Bifurcated Appellate Review: The Texas Story of Two High Courts, A.B.A. (Nov. 1, 2014), https://www.americanbar.org/groups/judicial/publications/judges_journal/2014/fall/bifurcated_appellate_review_the_texas_story_of_two_high_courts [https://perma.cc/9NHL-NMEA] (commenting on the development of Texas’s bifurcated appellate system and the broad authority of its Court of Criminal Appeals). ↑

75. Williams, supra note 67, at 666. ↑

76. Id. at 658. ↑

77. For a survey of the court structure of all fifty states, see State Court Structure Charts, Nat’l Ctr. for State Courts, Ct. Stat. Project (2018), http://www.courtstatistics.org/Other-Pages/State_Court_Structure_Charts.aspx [https://perma.cc/5C2H-VE3F]. ↑

78. See id. ↑

79. See, e.g., Nicolas Marie Kublicki, An Overview of the French Legal System from an American Perspective, 12 B.U. Int’l L.J. 57, 60–61 (1994) (“There are 455 tribunaux d’instance and 175 tribunaux de grande instance in France. Whereas the tribunaux d’instance are presided by a single magistrate, cases before the tribunaux de grande instance are heard by three judges. Each of the two courts contain both a civil and a criminal chamber.” (footnotes omitted)) ↑

80. Daniel J. Meador, Appellate Subject Matter Organization: The German Design from an American Perspective, 5 Hastings Int’l & Comp. L. Rev. 27, 44 (1981) (noting that each appellate court has civil and criminal divisions). ↑

81. But see infra note 91 (comparing appellate decision-making in the four states with specialized criminal courts to appellate decision-making in other states). ↑

82. Examples of significant divergence include discovery rules, settlement and plea bargaining conferences, and the jury selection process, as well as the byzantine rules governing habeas corpus review, which do not have any close analogy on the civil side. See generally Ion Meyn, Why Civil and Criminal Procedure Are So Different: A Forgotten History, 86 Fordham L. Rev. 697, 734 (2017) (analyzing the different paths that reform of the civil and criminal procedure rules took and concluding that “the reform of criminal procedure integrated civil rules that increased efficiency, like notice pleading and liberalized joinder, but rejected countermeasures designed to ensure accuracy, like judicial intervention and discovery tools”). ↑

83. This assumes that criminal cases constitute about 40 percent of the average judge’s docket in today’s generalist system. See supra note 26. ↑

84. Richman & Reynolds, supra note 49, at 198 (“Specialized judges . . . work more efficiently and quickly because they do not need to learn the elementary principles of an unfamiliar subject for each new case on the docket.”). ↑

85. See 18 U.S.C. §§ 3161–3174 (2018); e.g., Zukowski v. Howard, Needles, Tammen, and Bergendoff, 115 F.R.D. 53, 55 (D. Colo. 1987) (“The principle of criminal case preference was . . . made a practical necessity by the enactment of the Speedy Trial Act. While some minor adjustments can be made, there is no longer any doubt criminal cases must take precedence over civil cases on a federal court’s docket.” (citation omitted)). ↑

86. See, e.g., Lauren K. Robel, Grass Roots Procedure: Local Advisory Groups and the Civil Justice Reform Act of 1990, 59 Brook. L. Rev. 879, 906 (1993) (reporting a survey of 194 members of district court advisory groups established by the Civil Justice Reform Act of 1990, where one respondent named “the delay in appointing judges and the priority given the criminal justice system under the Speedy Trial Act” as the two “fundamental causes of undue expense and delay” on the civil side (citation omitted)); Diana E. Murphy, The Concerns of Federal Judges, 74 Judicature 112, 114 (1990) (arguing that “no management system for civil litigation in federal trial courts can be effective without adequate numbers of judges, relief from crushing criminal caseloads, and reduction in time-consuming processes”). ↑

87. See Ryan J. Owens & David A. Simon, Explaining the Supreme Court’s Shrinking Docket, 53 Wm. & Mary L. Rev. 1219, 1233 fig.3 (2012). ↑

88. Some of the questions that arise concern how big the en banc panel should be, the manner in which en banc review is triggered, and the precedential impact of en banc decisions. See Samuel Estreicher & John Sexton, Redefining the Supreme Court’s Role: A Theory of Managing the Federal Judicial Process 133 (1986) (noting that appeals courts so rarely grant petitions for rehearing en banc that “this avenue of redress is for all practical purposes nonexistent”). Claims have also been made that the en banc process is heavily tainted by ideology. See Note, The Politics of En Banc Review, 102 Harv. L. Rev. 864 (1989). ↑

89. See Richman & Reynolds, supra note 49, at 183–204 (explaining why this argument is a “great red herring,” as there is “no evidence that increasing the number of judgeships within a circuit reduces the stability of circuit law or increases the rate of appeal”). ↑

90. Consistency would be further enhanced by amending § 2254(d)(1) of the Antiterrorism and Effective Death Penalty Act, which currently provides that state courts need only follow opinions of the U.S. Supreme Court, Pub. L. No. 104-132, 110 Stat. 1214 (1996), to require states to follow opinions of the National Court of Criminal Appeals as well (unless, of course, the opinion is overruled by the U.S. Supreme Court). ↑

91. Cf. Paul D. Carrington, The Obsolescence of the United States Courts of Appeals: Roscoe Pound’s Structural Solution, 15 J.L. & Pol. 515, 518–20 (1999) (noting that circuit court opinions often have little precedential effect because other panels do not respect the expertise of previous “small, randomly selected” panels). ↑

92. A comparison of the four states with specialized criminal appellate courts indicates that all four are more likely to give defendants a merits review, and three of the four (excepting Oklahoma) are more likely to issue full judicial opinions than the average across the other states. Compare Michael Heise et al., State Criminal Appeals Revealed, 70 Vand. L. Rev. 1939, 1953 tbl.3 (2017) (showing that, on average, 55.5 percent of cases on appeal received merits review and 34.6 percent led to a full judicial opinion), with E-mail from Michael Heise to author (May 18, 2019) (on file with author) (showing the analogous data on Texas to be 60.1 and 40.7 percent; on Oklahoma to be 90 and 10 percent; on Alabama to be 61.9 and 42.9 percent; and on Tennessee to be 58.8 and 58.8 percent). However, oral arguments were rare or non-existent in all four states. E-mail from Michael Heise to author, supra. ↑

93. See Levy, supra note 61, at 443 (arguing that since the success of “nonargument review” of repeating claims “depends in large part on the staff attorneys who assess them . . . , circuit courts would do well to structure staff attorney offices to encourage subject-specific expertise where possible”). ↑

94. See Menell & Vacca, supra note 1, at 830–31 (describing these and other criticisms of the proposal for an Inter-Circuit Tribunal). ↑

95. See Erwin Chemerinsky, Remark, Lessons from Gideon, 122 Yale L.J. 2676, 2681–82 (2013) (“Many studies have been done in capital cases and they are remarkably consistent in documenting that . . . those with government-paid attorneys are much better off with public defenders than with appointed counsel.”). ↑

96. Allegra M. McLeod, Decarceration Courts: Possibilities and Perils of a Shifting Criminal Law, 100 Geo. L.J. 1587, 1590 (2012) (stating that there are approximately three thousand specialized criminal courts in the U.S. and its territories, including drug courts, mental health courts, veterans courts, and reentry courts). ↑

97. See Edward K. Cheng, The Myth of the Generalist Judge, 61 Stan. L. Rev. 519, 526 (2008). ↑

98. See id. at 560 (arguing that “[i]f judges without a criminal law background avoid writing criminal opinions” and “former criminal defense attorneys seldom become judges because of political unpopularity,” then former prosecutors are left to determine the future of federal criminal law). ↑

99. Diane P. Wood, Speech, Generalist Judges in a Specialized World, 50 SMU L. Rev. 1755, 1756 (1997). ↑

100. Id. at 1767. ↑

101. Id.; Paul D. Carrington et al., Justice on Appeal 168 (1976) (arguing that specialized judges “will lose sight of the basic values at stake” and develop “arcane [law] understood only by their own bar”). ↑

102. See Rochelle Cooper Dreyfuss, The Federal Circuit: A Case Study in Specialized Courts, 64 N.Y.U. L. Rev. 1, 3 (1989). ↑

103. Rachel Elise Barkow, Prisoners of Politics 200 (2019) (providing the ex-prosecutor percentage and indicating that the percentage of judges who were previously public defenders is 10.4 percent). ↑

104. See Carrington et al., supra note 100, at 168. ↑

105. Wood, supra note 98, at 1767. ↑

106. See, e.g., 28 U.S.C. § 2255 (2018) (providing a mix of criminal and civil rules for habeas actions); Graves v. Penzone, No. CV-77-00479-PHX-NVW, 2017 WL 782991 (D. Ariz. Mar. 1, 2017) (class action challenging confinement conditions of pretrial detainees in Maricopa County, Arizona). ↑

107. It is also belied by past experience in Texas, where the Court of Criminal Appeals has disagreed with prosecution-oriented decisions by the Fifth Circuit on several occasions. Gary A. Udashen & Barry Sorrels, Criminal Procedure: Confession, Search and Seizure, 45 Sw. L.J. 263, 269–73 (1991); see also Michael J. McCormick et al., Fundamental Defect in Appellate Review of Error in the Texas Jury Charge, 15 St. Mary’s L.J. 827, 827 (1984) (noting the Texas court’s “[c]ontinued expansion” of the automatic reversal standard of review for fundamental error). ↑

108. Perhaps of relevance here is the fact that during the relatively “liberal” Obama administration, 42 percent of district court judges and 37.5 percent of circuit court judges nominated to the federal bench were ex-prosecutors. All. for Justice, Broadening the Bench: Professional Diversity and Judicial Nominations 8–9 (2016), https://www.afj.org/wp-content/uploads/2014/11/Professional-Diversity-Report.pdf [https://web.archive.org/web/20191118050834/https://www.afj.org/wp-content/uploads/2014/11/Professional-Diversity-Report.pdf]. ↑

109. In Denver, for instance, “[a]bout one-third of the Judges change assignments annually[,] with the usual rotation being eighteen months in domestic relations, two years in criminal, and three years in civil.” Denver Dist. Court, Second Judicial District: District and Juvenile Court Overview - 2020, at 1, https://www.courts.state.co.us/userfiles/File/Court_Probation/Supreme_Court/Judicial_Nominating_Commissions/Overviews/02_Overview.pdf [https://perma.cc/6B2Z-8GZA]. ↑

110. See Comm. on the Superior Court Circuit Sys., Bos. Bar Ass’n, Rotation Ripe for Reform 7–8, 9–11 (2003), https://www.bostonbar.org/prs/reports/bbatask122203.pdf [https://perma.cc/SBF6-NG87] (noting that New Jersey, Florida, and Rhode Island courts rotate every two to three years, and criticizing the short-term rotation of judges within the Massachusetts Superior Court Circuit System, measured in months). ↑

111. Menell & Vacca, supra note 1, at 844 fig.2. Recall that “civil cases” include prisoner and habeas cases, and so should be reduced accordingly. ↑

112. See supra note 26 (noting that, in the typical year, the number of criminal cases filed is about two-thirds the number of civil cases filed). ↑

113. See generally Michael M. O’Hear, The Original Intent of Uniformity in Federal Sentencing, 74 U. Cin. L. Rev. 749, 749, 779 (2006) (stating that “[u]niformity has become a dominant objective of the federal sentencing system” and noting that the federal Sentencing Commission “characterized the guidelines as an effort to ‘balance’ certainty with proportionality” (citation omitted)). ↑

114. See Christopher Slobogin, Limiting Retributivism and Individual Prevention, in The Routledge Handbook of the Philosophy and Science of Punishment (Farah Focquaert et al. eds., forthcoming 2020) [hereinafter Slobogin, Limiting Retributivism and Prevention] (on file with author); Christopher Slobogin, Plea Bargaining and the Substantive and Procedural Goals of Criminal Justice: From Retribution and Adversarialism to Preventive Justice and Hybrid-Inquisitorialism, 57 Wm. & Mary L. Rev. 1505 (2016). ↑

115. See Comprehensive Crime Control Act of 1984, Pub. L. No. 98-473, § 218(a)(5), 98 Stat. 1837, 2027. ↑

116. See Edward E. Rhine, The Present Status and Future Prospects of Parole Boards and Parole Supervision, in The Oxford Handbook of Sentencing and Corrections 627, 631–32 (Joan Petersilia & Kevin R. Reitz eds., photo. rprt. 2015) (2012). ↑

117. See Slobogin, Limiting Retributivism and Prevention, supra note 113 (manuscript §§ 2–4). ↑

118. See Rhine, supra note 115, at 630–31; see also S. Rep. No. 98-225, at 40, 56–58, as reprinted in 1984 U.S.C.C.A.N. 3182, 3223, 3239–41 (detailing reasons for abolishing parole in the federal system). ↑

119. See Slobogin, Limiting Retributivism and Prevention, supra note 113 (manuscript § 3). For a more detailed description of professionalized parole boards, see Edward E. Rhine et al., The Future of Parole Release, 46 Crime & Just. 279 (2017). ↑

120. Slobogin, Limiting Retributivism and Prevention, supra note 113 (manuscript § 1) (“[A] system of relatively wide sentence ranges derived from retributive principles, in combination with short minimum sentences that are enhanced under limited circumstances by statistically-driven risk assessment and management, can alleviate many of the inherent tensions between desert and prevention, between deontology and political reality, and between the desire for community input and the allure of expertise. If done properly, it should also significantly reduce prison populations.”). ↑

121. Kevin R. Reitz, Sentencing Guideline Systems and Sentence Appeals: A Comparison of Federal and State Experiences, 91 Nw. U. L. Rev. 1441, 1491 (1997) (“Of 6659 criminal appellate decisions of all kinds in 1994-95 . . . 4314 of these, or 65%, dealt with an appeal from sentence.”). But see King & O’Neill, supra note 23, at 227 (finding that appeal waivers have reduced sentencing appeals since Reitz wrote). ↑

122. Cf. Reitz, supra note 120, at 1443 (“Prior to the guideline innovations of the 1980s, little meaningful appellate review of sentencing decisions had ever occurred in the United States, in federal or state courts.”). Under the proposed regime, however, appeals to the courts would still be necessary to develop a much-needed jurisprudence of risk assessment. Slobogin, Limiting Retributivism and Prevention, supra note 113 (manuscript §§ 5–6). ↑

123. Consider these remarks by a Second Circuit judge, formerly a U.S. District Court judge: “Sentencing is perhaps the most important responsibility of a trial judge, and surely the most difficult. Emotion is one reason it is so difficult. The competing considerations evoke strong sentiments—anger, indignation, shame, sorrow, grief, despondency, and hope. The sentencing judge is not immune from these emotions.” Denny Chin, Essay, Sentencing: A Role for Empathy, 160 U. Pa. L. Rev. 1561, 1579 (2012) (footnote omitted). ↑

124. In researching this Article, I corresponded with two court of appeals judges, two district court judges, and a state court judge, all of whom expressed this view. ↑

125. See Gall v. United States, 552 U.S. 38, 49–50, 50 n.6 (2007) (explaining that, after calculating the appropriate guidelines range, the sentencing judge “should then consider all of the § 3553(a) factors [which include ‘seriousness of the offense,’ ‘deterrence,’ ‘protect[ion of] the public,’ and education and correction treatment] to determine whether they support the sentence requested by a party” and noting that “[i]n so doing, [the judge] may not presume that the Guidelines range is reasonable”). ↑

126. These ranges are taken from the original Model Penal Code, which endorsed indeterminate sentences. Model Penal Code Alternative § 6.06 (Am. Law Inst., Proposed Official Draft 1962). ↑

127. Kevin R. Reitz, The “Traditional” Indeterminate Sentencing Model, in The Oxford Handbook of Sentencing and Corrections 270, 277 (Joan Petersilia & Kevin R. Reitz eds., photo. rprt. 2015) (2012) (“For its policy foundation, the parole release system since late in the nineteenth century has most often been linked with rehabilitation theory. . . . The key function of the paroling authority is to review each prisoner’s progress, and to use its expertise to identify those who have reached the goal.” (citations omitted)). ↑

128. See Menell & Vacca, supra note 1, at 841, 843 (noting, for example, that in the 2000s, “the major governance organizations for the federal judiciary largely shuttered the reform studies and initiatives” and that “[a]fter three decades of concerted efforts toward the end of the last millennium to adapt the federal judiciary to substantial change, the reform machinery ground to a halt as the new millennium began”); see also Richman & Reynolds, supra note 49, at 217–22 (discounting the federal judiciary’s arguments that creating more than one thousand judgeships would dilute quality and prestige, and attributing appellate judges’ resistance to reform to their comfort with their current quasi-administrative role, given its similarity to senior partnership in a large law firm, to their “distaste” for run-of-the-mill cases, and to elitism). ↑

129. See, e.g., Jane L. Froyd, Comment, Safety Valve Failure: Low-Level Drug Offenders and the Federal Sentencing Guidelines, 94 Nw. U. L. Rev. 1471, 1488 (2000) (stating that “[t]his goal of retribution or ‘just desserts’ [sic] was the most commonly voiced reason for instituting mandatory minimum penalties” on drug offenders); James S. Gwin, Juror Sentiment on Just Punishment: Do the Federal Sentencing Guidelines Reflect Community Values?, 4 Harv. L. & Pol’y Rev. 173, 180 (2010) (discussing “the prioritization of retribution” under the Guidelines). ↑

130. See supra text accompanying note 125. ↑

131. See Kenneth Culp Davis, Discretionary Justice 126–33 (Univ. of Ill. Press 1976) (1969) (ultimately calling the federal parole board the most disappointing agency the author had ever encountered). ↑

132. See Drug Courts, Nat’l Inst. of Justice, https://www.nij.gov/topics/courts/drug-courts/Pages/welcome.aspx [https://web.archive.org/web/20190811033756/https://www.nij.gov/topics/courts/drug-courts/Pages/welcome.aspx] (indicating that there are twenty-seven federal district drug courts); Specialized Courts, Nat’l Inst. of Justice, https://www.nij.gov/topics/courts/pages/specialized-courts.aspx [https://web.archive.org/web/20190820231911/https://www.nij.gov/topics/courts/pages/specialized-courts.aspx] (indicating that there are forty-four federal reentry courts and two federal “problem solving” courts). ↑

133. See Richard C. Boldt, Problem-Solving Courts and Pragmatism, 73 Md. L. Rev. 1120, 1127, 1129 (2014) (stating that “[t]he driving force behind the problem-solving courts movement from its inception has been an express commitment to efficacy,” a movement that “offers ‘therapeutic outcomes’ for participants, rather than ‘legal resolutions’ for cases”). ↑

134. Pub. L. No. 115-391, 132 Stat. 5194 (2018). ↑

135. Brandon L. Garrett, The Prison Reform Bill’s Implementation Will Be Tricky, Slate (Dec. 27, 2018), https://slate.com/news-and-politics/2018/12/prison-reform-bill-success.html [https://perma.cc/8RUW-R3QY] (“The final version of the First Step Act, which refers to ‘risk’ 100 times, calls for a ‘risk and needs assessment system’ to be developed in 210 days, and then made public and administered to every federal prison within the following 180 days.”). ↑