Heritage Foundation Leadership for America

Testimony of

Paul Rosenzweig,

Senior Legal Research Fellow

Center for Legal and Judicial Studies

Before the American Bar Association

The United States Sentencing Commission

Regarding:

Sentencing in a Post-Booker World-
It's Deja Vu All Over Again

When I was last here to testify, I paraphrased Abraham Lincoln, suggesting that we are “engaged in a great civil debate to decide whether the current determinate federal sentencing system or any other system so conceived can long endure.” We now know the answer—a determinate sentencing system of the form contemplated by the Sentencing Reform Act can endure for only 15 years—and no longer. The question before us now is “what next?” And we ask that question against a backdrop of, candidly, surprise. Surely, when last this Commission heard testimony on the subject nobody—and I really mean nobody at all—would have predicted that Booker would have been decided as it was and that we would be here today contemplating the questions we are. At a minimum that surprise suggests caution and humility in a proscriptions for the future—having been surprised once, we may well be surprised again.[1]

Brief Background

At the risk of stating the obvious, I wanted to lay out where we are now: Since 1987, Federal defendants have been sentenced under a determinate sentencing system intended to enhance uniformity and consistency by accounting for the so-called “real offense” conduct underlying a conviction and the relevant individual characteristics of each defendant.[2] Under this system, the guideline ranges to be applied to each defendant were determined by the district court, which was obliged to calculate the appropriate sentencing range after making a series of factual findings about the defendant and the nature of his conduct. In other words, the final sentence to be imposed was based both on facts found by jury as part of its determination of guilt or innocence (under a “beyond a reasonable doubt” standard), and facts later found independently by the court prior to sentencing, using a “preponderance of the evidence” standard.

Under the sentencing regime in effect prior to Booker, district courts could depart from the calculated guideline range only under very limited circumstances. Remaining within the guideline range calculated was required “unless the court [found] that there exists an aggravating or mitigating circumstance of a kind, or to a degree, not adequately taken into consideration by the Sentencing Commission in formulating the guidelines.”[3] The Courts had interpreted this provision as establishing an extremely high burden of justification for any departure from the presumptive guidelines sentence.

In Booker, the Court concluded that the existing federal guideline system violated the Sixth Amendment. The Court began by noting the effectively mandatory nature of federal sentencing: “The guidelines as written … are not advisory; they are mandatory and binding on all judges.”[4] Given the binding nature of the guidelines, the Court found that its analysis in Blakely was equally compelling in the federal context. It therefore determined that the Sixth Amendment prohibits the enhancement of sentences based on facts found independently by the district court under a reduced burden of proof: “Any fact (other than a prior conviction) which is necessary to support a sentence exceeding the maximum authorized by the facts established by a plea of guilty or a jury verdict must be admitted by the defendant or proved to a jury beyond a reasonable doubt.”[5]

The Court then turned to the question of how to remedy the flawed structure of the federal sentencing system. The Court rejected a remedy that would have continued to allow the guidelines to operate in a mandatory manner by imposing a jury-finding requirement on all sentencing enhancements.[6] Rather, the Court found that Congress would have preferred the adoption of an advisory guidelines sentencing system that requires sentencing courts to consider the guidelines but permits then to impose sentences that are within the statutory range authorized by law, subject only to review on appeal for the reasonableness of the sentence.[7] To effectuate that remedial determination, the Court severed and excised as unconstitutional two provisions of the Sentencing Reform Act—that making adherence to the guidelines mandatory and that requiring de novo appellate review.[8]

Thus, to comport with the Sixth Amendment today a defendant must be sentenced under an advisory guidelines scheme—one where the district judges “consult [the] Guidelines and take them into account,” but may thereafter impose a sentence at variance with the guideline range based upon the sentencing factors set forth in section 3553(a),[9] so long as they adequately explain their reasons for doing so.[10]

The consequences of this decision seem to me to be as follows:

For now, the courts will “consult” the Guidelines, treat them as advisory, and impose sentences that are subject only to a “reasonableness” review.

Any attempt to revive a determinate sentencing structure must account for the Constitutional rule: Facts that are not found by the jury or admitted by the defendant may not be determined by a sentencing judge and used in a manner regulated by law to fix the defendant’s sentence.

The rule in Harris[11] (that a judge may find facts which determine a statutory minimum) remains unchanged, for the time being.

The rule in Williams[12] allowing indeterminate sentencing decisions by a judge remains unchanged.

The rule allowing judges to determine facts relating to prior criminal history and convictions[13] also remains unchanged.

I remain uncertain how long this state of legal affairs will last—both Harris and Almendarez-Torres are certainly subject to further modification by the Court if it hews strictly to the logic it has advanced in Booker, though notably, neither was adversely commented upon in Booker. Of the two, Harris with its approval of an asymmetric sentencing system, seems the less likely to survive in the long run while Almendarez-Torres rests, I think on slightly firmer ground. For now at least, I perceive the Constitutional framework as stable—but one doubts if that stability will long endure. And the prospect of instability is another reason for counseling caution.

Booker Will Not Endure

I realize that the principle purpose of this hearing is to discuss issues derived from Booker—that is, learning to live with the decision. In my view however, that should not be the only area of inquiry for this Commission. For in the end, my judgment is that Booker is, itself, only a temporary part of the sentencing structure. It cannot last. The instability of Booker itself and the uncertainty of the régime it creates derives from theoretical, analytical, and political flaws that cannot be concealed and perhaps before we rush to consideration of how best to implement Booker it would be worthwhile to assess its vitality. To be sure, I perceive myself to be in the minority on this point among academics and policy makers but I candidly do not believe that the push me-pull you solution in Booker can long withstand either political or constitutional scrutiny.

First, as a broad theoretical matter Booker lacks coherence. I realize that many reasonable jurists disagree, and that the Court has disregarded that incoherence. But at a fundamental level, I cannot anticipate the longevity of a decision that is so internally inconsistent. At a macro level, any reasonable observer of the political scene knows that the remedial majority opinion rests on a clearly transparent fiction—for no one truly believes that if Congress had been allowed to express a choice it would have opted for advisory guidelines. Not a Congress that rejected advisory guidelines explicitly in the 1980s and moved less than two years ago to enhance the mandatory nature of the guidelines through the PROTECT Act.

Nor can anyone (save perhaps Justice Ginsburg) readily reconcile the Constitutional violation identified with the remedy chosen. Only in Alice in Wonderland does a Constitutional requirement for greater jury participation in sentencing determinations admit of a remedy that empowers the judiciary and excludes the jury altogether. Neither logic nor the law can withstand that tension over the long haul.

On the narrower analytical front, the decision is equally puzzling. The Booker remedial majority has created an unstable world where the Guidelines must be “consulted” and, I assume, accorded some value yet remain “advisory.” I recognize that some believe that this advisory system is consistent with the fundamental underpinnings of the Booker/Blakely constitutional opinions. With all due respect however, I believe it is not—and where a fundamental logical inconsistency exists we should be especially cautious.

The image advanced is that the advisory guidelines will be sufficiently comprehensive that they will, through their moral force and persuasive power, I guess, achieve the goals of uniformity, proportionality and certainty. In other words, the advisory guidelines are viewed by the Booker remedial majority as a mid-way point between completely indeterminate sentencing, and the current overly rigid Guideline structure.

But that mid-point does not, I submit, exist. To have any substantial success in reducing disparity the advisory guidelines will need to have some rigor, or if you prefer, “bite.” In other words they must have some practical effect on what judges do. If they do not, if they are a purely advisory system with absolutely no practical constraint on judicial decision-making, then they will be little more than a return to completely indeterminate sentencing.

But we are already seeing a wide divergence in how judges treat the persuasiveness of the guidelines. Some have said they will give the guidelines “heavy weight” and depart from their dictates only in extraordinary circumstances.[14] Others have said that the guidelines are now but one factor amongst the many factors enunciated in section 3553(a) to be considered and they won’t be given any decisive weight.[15] Others, perhaps most remarkably, have concluded that the discretion they have been given allows them to adopt the Blakely-ized mandatory guidelines system that was rejected by the Booker majority and championed by its dissent.[16] These early efforts highlight the potential for sentencing disparity that underlies any advisory system.

To be sure, there is some hope or expectation that this early confusion at the district court level will be “smoothed out” by appellate review for sentence “reasonableness.” But I confess I am skeptical. There are so many questions that underlie the “reasonableness” standard that I feel utterly confident in predicting years of litigation, multiple circuit splits, and the frequent necessity for clarifying Supreme Court intervention.

But, there is a problem beyond confusion—namely the possibility of the absence of confusion. In my judgment if an advisory system is to have any practical effect in any rigorous and effective way, it must also have a legal effect. For without any legal constraint of any form the system will trend to a practical nullity and a return to indeterminacy. But, to the extent that the system does have practical effect, that practical constraint must be interpreted as equivalent to a legal constraint—in answering questions concerning the legal limitations of Booker and imposing order on the lower courts the courts of appeals must apply rules that have reasonable predictability and certainty. And that, I fear, brings us right back to the lurking Booker/Blakely problem.

Appellate review, by its very nature, necessarily entails consideration of some reasoned decision made by the lower court—some application of law to the facts under a defined standard of proof and a defined standard of review. If any of those constraints operate as a legal requirement—if the lower court must “find” brandishing, for example, by some quantum of proof and presume that a brandishing enhancement is appropriate absent unusual circumstances—then both the necessity of making such a finding and the assessment of whether the sentence thereafter imposed is “reasonable” under the standard required can be subject to appellate review. But the necessity of a finding or the imposition of a standard are necessarily in conflict with the core Booker/Blakely rule—they make either that finding or that standard an “essential” component of the punishment, and if that finding or standard is determined by a judge, rather than a jury, then, the jury guarantee is vitiated. However, without either a required finding or an enforceable standard, meaningful appellate review is impossible. It seems to me that the only person who doesn’t see this logical necessity is Justice Ginsburg—and I fear that it is so palpable a problem that the foundations of our new sentencing system rest on legal quicksand.

In short, the circle cannot be squared. To pass constitutional review after Booker the guidelines must, as a matter of law, leave the judge free to impose any sentence between some statutory minimum (which may, of course, in appropriate circumstances be a sentence with no term of incarceration) that arises from the fact of conviction of the base offense and the statutory maximum. If the advisory guidelines act as a legal constraint on the judge’s decision then, in my view, it logically fails the Booker/Blakely test. So even if we wish advisory guidelines, and even if we hope judges will follow them, we are nonetheless left with a system that, if it is to consider non-jury-determined facts must necessarily be completely indeterminate as a matter of law.[17]

Finally, and perhaps most importantly, the advisory guideline system will not, in my view, withstand political scrutiny. The fundamental goal of the Sentencing Reform Act was to reduce judicial discretion and disparity in sentencing. That goal finds even greater expression in recent congressional enactments, such as the Feeney Amendment, that demonstrate Congress’ continuing attention to, and involvement in, questions of punishment. It appears that Congress may not act in the near term, as the heed the counsels of caution. But act, eventually, they will. Advisory guidelines will engender an absolutist response—if considered as a serious permanent solution, I feel confident that Congress will respond with a plethora of mandatory minimums.

For these reasons, I think that the Commission should not be single-minded in its focus on Booker. Five years from now Booker will no longer be controlling law. So, as a long-term project the Commission should look beyond the current advisory system and recommend its vision of the optimal, stable system. More importantly, even if I am wrong—even if Booker reflects permanent reality—nothing in Booker is inconsistent with revisions and simplifications to the existing Guidelines.

In my testimony before you last November, I recommended a simplified Guideline structure capable of presentation to a jury.[18] Perhaps that is the right answer, perhaps not. But I would not want the Commission to be distracted by the immediacy of Booker from its longer-term, more difficult task.

Learning to Live With Booker

That having been said, I would be remiss if I did not address the more pressing and immediate questions you face. I certainly do not envy you the task. But it is also an important one, for the only thing we can be certain of is that Booker will have both practical and legal effects in the short-term, even if not in the long run.

On a practical level, the effect will be quite immediate. As I said when I was here last, I think that the defendant’s decision to forgo his right to a jury trial more typically has little to do with repentance and remorse and far more to do with the magnitude of the concessions that will be afforded him. Because the jury trial method is already quite costly, to avoid the necessity of the trial prosecutors are often willing to make concessions to secure a guilty plea. It is inevitable that any time the costs to a prosecutor of a trial go up, or the benefits go down, the cost-benefit calculus will be rebalanced, to the benefit of criminal defendants.

Thus, since as a practical matter we cannot really change the clearance rate for plea bargains (because the Federal system is not able to readily accommodate a substantial increase in the number of jury trials annually), Booker will almost certainly and necessarily affect the “price” that prosecutors are willing to (or more accurately, will be obliged to) pay to obtain a guilty plea. The Booker rule will, inevitably, raise the costs of trial by increasing the uncertainty of the sentence to be imposed and more specifically, by reducing the leverage to be gained by the government through the entry of cooperation agreements and 5K1.1 letters. Even if the plea bargain rate does not change, it is inevitable that, without adjustment, Booker will reduce the period of incarceration experienced by criminal defendants. In short, though the direct effect on trials may be modest, I believe that the indirect effects of Booker are likely to be very significant at a practical level.

They are also, as the Commission’s invitation letter makes clear, quite likely to be very significant effects on the law. Sentencing law at this moment is in a state of flux. In addition to the questions already identified by the Commission, one can imagine many others:

What, for example, will constitute harmless error in pending appeals? Prior to Booker the Supreme Court had held that a sentencing error was harmless if it “did not affect the district court’s selection of the sentenced imposed.”[19] Does that rule still apply? If so, does that mean that a district courts failure to follow procedural requirements is harmless if the eventual sentence is nonetheless the same one it would have imposed had it followed procedure? If so, how does this interact with “reasonableness” review?

Or consider this: What will constitute plain error? There is already a circuit split on the question of whether pre-Booker sentences issued under the determinate sentencing system constitute plain error, even if they don’t involve judicial fact-finding.[20] But I am even more interested in how the plain error rule will interact with a “reasonableness” standard of appellate review. If a defendant fails to object to certain procedural or legal flaws in his sentencing, and his sentence is as a result “unreasonable” may that error be noticed on appeal? And I don’t know the answer.

Perhaps the legal system it will trend towards stability, if the “reasonableness” standard is implemented with real vigor, but we might be skeptical of that. Even it’s the most basic questions are indeterminate. For example, is unreasonableness determined solely by reference to the bottom line length of sentence imposed, measured against the general sentencing factors in section 3553(a)? If not, than what other sorts or errors will lead to a conclusion of unreasonableness? Legal errors? Procedural errors? Factual errors?

At the most general level, courts will need to exercise extreme caution as they begin to define “reasonableness.” We can certainly agree with the Second Circuit that reasonableness is “inherently a concept of flexible meaning, generally lacking precise boundaries.”[21] But there are some dangers in interpretation that lurk even so. Courts might, for example, deem all procedural failings to render a sentence “unreasonable”—a view that would essentially eviscerate the harmless error concept. Appellate courts might also defer substantially, on something like an “abuse of discretion” standard to any variance from a calculated guideline range—a review standard that would foster diversity in sentencing and reduce uniformity. Alternatively, appellate courts might, in effect, review each sentence for reasonableness on something like a de novo standard, assessing for themselves in the first instance what the sentence should be.

The broad outlines of the right answer lie somewhere in between—a general set of rules that encourages adherence to the guidelines as a presumptively adequate sentence; a requirement of reasoned decision-making for any variance; and a reasonably deferential review of variances under something like a “clearly erroneous” standard.

There will also be many narrower questions—procedural, legal, and factual. A sampling would include:

Must a district court calculate the applicable guideline range in all cases? Or can it dispense with finding the requisite facts if they are not necessary to a sentence—as for example if the factual dispute will produce two guideline ranges that overlap and the district court intends to sentence within the overlap? Or is any sentence in which the district court fails to consult the guidelines by fully calculating them, per se unreasonable.

Here I believe that as a general principle the district court must continue to calculate a guideline range fully and completely. But I would think that the pre-Booker doctrine of excusing the resolution of complex factual questions when the court chooses a sentence that is consistent with either factual construction should continue to apply. This would also create an indirect incentive to the district courts to accept the guidelines as a substantial factor.

What is the nature of the process that is required for determining facts on which a district court may now rely in imposing an advisory sentence? Is it “unreasonable” to rely on hearsay? Is it “unreasonable” to presumptively accept the facts adduced by the probation office through its pre-sentence investigation?

Caution again is likely the best answer to these questions. Courts should be reluctant to read into the requirement for review for reasonableness a command that we change sentencing practices that have evolved over the last 15 years. Indeed, because of the greater discretion now permitted judges, they will be free to consider the weight of the evidence presented, I assume, in deciding whether to vary from the presumptive calculated guideline range.

Who has the burden of persuasion at sentencing, and by what quantum of proof? Is it unreasonable to rely on ex parte factual submissions or facts that have not been fully disclosed to the other party to the proceeding?

Again, these are settled practices. The Commission might consider whether or not it should call for reform (and indeed there is much good reason to believe that reform is warranted). But that reform if it occurs at all, should be the product of legislative consideration or rules revisions, not of an ad hoc attempt to define reasonableness in terms of procedural requirements.

Are sentences that fall within the calculated guidelines even appealable at all?

Recall that the Booker remedial majority carefully excised only section 3742(e), concerning de novo review, leaving in place the remainder of section 3742. Notably, section 3742(a) (and its parallel subsection (b)) allow for appeals only in a limited number of circumstances. Reflecting on the text of the statute, even after excision, I am not sure that either party may appeal a sentence imposed within a properly calculated range. And that seems to me a reasonable rule—as it will enhance the presumptively reasonable nature of the guidelines themselves and concomitantly provide for an incentive for district courts not to vary their sentences from the guidelines too readily.

Is a district court’s legal error in interpreting and applying the guidelines per se unreasonable? Or can there be a “reasonable” but erroneous guideline range?

I confess I don’t know the answer to this question. Consider again the district court that has adopted, in its discretion, Blakely-ized guidelines. It is fair to assume that choice will eventually be rejected. But what if the sentences imposed thereby are nonetheless “reasonable” on some other, objective standard. Does the use of the wrong process constitute harmless error? But if not, then what will distinguish harmless procedural and legal errors from ones that are not harmless?

One example that presses the point would be this: What if the legal error arises from a legal question in which there is no controlling legal precedent? Would appellate reversal of the underlying legal question (on which reasonable minds might disagree) render the sentence imposed per se unreasonable?

One would think not—if the ultimate length of sentence was reasonable and the legal interpretation one that a district court could have appropriately, though erroneously, chosen, this would seem the perfect scenario for the application of a harmless error analysis. Yet, the “appropriateness” of the legal choice will always lie in the eyes of the beholder—and some appellate courts will disagree on what they see.

How do we define an unreasonable sentence when the district court is obliged to apply an open-ended guideline, such as those that cover fraud, where they guidelines have widely divergent applications, and even the concept of “loss” has a different meaning in different factual contexts.

And what, in the end, does it mean to “consult” the guidelines:

What weight should they be given to the guidelines by the district court—a problem already alluded to?

Here, at the risk of being called a sycophant, I confess complete agreement with Judge Hinojosa’s testimony last week before the House of Representatives. The guidelines should be given substantial weight by the district courts. This is so precisely because of the Commission’s and because the guidelines as constructed are themselves intended to take into account all of the factors enunciated in section 3553(a). While it is readily conceivable that in specific cases the guidelines will not prove dispositive, they should be accorded a presumption of reasonableness and any variation from their guidance should, in my judgment require justification. I doubt the Commission has the authority to require judges to accord the guidelines substantial weight—but it does have significant moral authority to do so.

May a district judge disregard a guideline prohibition? The guidelines, for example, exclude from consideration including the defendant’s race, sex, national origin, creed, religion, and socio-economic status.[22] The Commission also prohibited consideration of other factors—such as the defendant’s dependence on alcohol, drugs, or gambling, lack of guidance as a youth, disadvantaged upbringing and others—and discouraged consideration of other factors.[23] Is it unreasonable to consider factors that the Commission has expressly excluded?

One would think so. Yet already anecdotal reports suggest that courts are considering factors like youth and socio-economic status in determining to reduce sentences.[24] It would be well to remember why the Commission has excluded such factors—it is precisely because they are of such variable effect on a jurist. Where one judge may see poverty as a basis for downward variance, another might see standing in the community as equally compelling. I would urge the Commission to strongly caution against consideration of such factors and hope that appellate courts develop rules that accord sentences based upon such prohibited factors less deference than those based upon factors that have generally been deemed permissible considerations.

What about congressional determinations? Is it unreasonable to disregard them? For example, the Commission has repeatedly proposed reducing the crack/powder cocaine disparity and Congress has just as repeatedly rejected the recommendations of the Commission, expressing a clear legislative preference?

At least one court—the district judge in Nellum—has already considered the question (without ruling on it) and opined in dicta that the Commission’s superior expertise warrants greater deference than the clearly expressed Congressional intent. News reports suggest that other jurists have agreed with this conclusion.[25] This, of course, turns the “substantial weight” question sideways. The argument obviously is that if we are to accord weight to the Commission because of its expertise, we must do so uniformly. Indeed, I would expect that the crack/powder disparity will be the first locus of significant judicial resistance to the guidelines and a substantial cause of variance from them.

I suspect, however, that this is wrong—and for precisely the reasons that have led to much of the controversy between the courts and the legislature in the past few years. Congress is, in our system, entitled to the final say in matters within its competence, so long as it doesn’t trench on constitutional limitations. I confess to great sympathy for the Commission’s position on this issue—but I think that systematic disregard of a considered congressional judgment will be a catalyst for a counter-revolutionary reaction. Thus, I think the Commission should strongly encourage the district courts to be no less deferential to guidelines that have a congressional imprimatur even though they may have substantive disagreements with the underlying policy decisions.

My opinions on how these questions should be answered are, of course, just that—opinions. I have no confidence at all that I have been successful in discerning what the courts are likely to answer when asked. Indeed, I strongly suspect that we will have very divergent views between the circuits on most, if not all, of these questions.

Sentencing Data

One final word about the need for sentencing data. I was pleased to see the Commission’s initial report to Congress last week. And the Commission should continue to collect data about rates of departure, and the like. The initial report suggests little change in the degree of variance from the guidelines post-Booker and that is a good thing. But, we should recall that the present level of variance (roughly 1/3rd of cases sentenced outside the guideline range) is what motivated Congressional response in the Feeney Amendment—and so the lack of any additional divergence from the guidelines will not be seen by Congress as a success.

But beyond the rates of departure question, I want to encourage you to undertake a more ambitious project, the idea for which was first brought to my attention by Judge Nancy Gertner of Massachusetts. If the advisory system is to function well, it must be based on more information sharing between judges. As many jurists have told me, over time they develop a feeling for sentencing based upon their own experiences, and use those experiences for comparative valuations amongst defendants. “A is worse than B who I sentenced to 8 years but not as bad as C who got 10” for example.

But today a judge’s experience is wholly his or her own and wholly vertical within his own courtroom. The Commission needs to begin systematically collected even greater amounts of information about sentences—crime characteristics and offender characteristics—to allow judges to share information horizontally from courtroom to courtroom. As it stands now, the district judges in a busy urban court (like, say, New York) have virtually no idea how their peers are treating defendants who come before them. Yet it is absolutely certain that there is a commonality amongst the offenders and their offenses. And if that information were readily accessible to all of the judges in New York (not to mention nationwide) that, by itself, would advance the goals of uniformity and transparency. As a long-term goal, I can think of little that would be more beneficial.