Talk of reforming federal sentencing law is much in the air. Increased public awareness of the fact that the United States is the world capital of mass incarceration has prompted public officials in both political parties as well as other prominent figures to call for legislation that would reduce the number of federal prisoners. After remaining silent about criminal justice issues for a long time, President Obama recently became the first president in the country’s history to visit a federal prison and has called for curtailing mandatory minimum sentences. Senators such as Dick Durbin, Cory Booker, Mike Lee, and Rand Paul have also expressed support for fewer mandatory minimums, and it appears increasingly likely that Congress will enact some sort of legislation on the subject.
Eliminating mandatory minimum sentences is enormously important and will undoubtedly lead to some reduction in the federal prison population. How substantial the reduction is, of course, will depend on the scope of the legislation. Given the widespread expressions of concern about the number of federal prisoners, however, it is somewhat surprising that there has been little public discussion of approaches to reducing the federal prison population that don’t require new legislation and don’t involve mandatory minimum sentences. The fact is that many federal offenders are unnecessarily imprisoned, and that many overly long sentences are imposed in cases where there is no mandatory minimum. To address this problem, the conversation about excessive imprisonment needs to be expanded to include the impact of the prison-oriented approach of the federal sentencing guidelines and federal judges’ continued adherence to that approach notwithstanding the guidelines’ nonbinding status.
In this article, we first discuss how, by placing far too much emphasis on prison and far too little on sentences served in the community, the guidelines contribute to mass incarceration. Next, we discuss federal judges’ excessive attachment to the guidelines despite their deep flaws and even after the Supreme Court has made clear that they are advisory only and that judges are free to reject them. Finally, we propose an approach to federal sentencing that is far less deferential to the guidelines and far less oriented towards putting people in prison, and we provide several real world examples. In conclusion, we assert that if we are actually to achieve a significant reduction in the federal prison population, federal judges will have to make major changes in their sentencing practices.1
In 1984, Congress passed and the President signed the Sentencing Reform Act (“SRA”). Among other things, the SRA created an entity known as the United States Sentencing Commission (“the Commission”) and directed it to create guidelines for judges to follow in imposing sentence. Despite the modest name given the Commission’s product — “guidelines” — the SRA afforded them the force of law.2 Judges had to follow the guidelines in all but the most unusual cases, and appellate review was available to police their compliance. The SRA further required the Commission to create guidelines with very narrow ranges so that the sentence imposed could not vary by more than the greater of 25 percent or six months from top to bottom.3
After the guidelines went into effect, the average federal sentence increased from 28 months to nearly 50 months.4 While mandatory minimum sentences played a role, much of the increase can be traced to the Sentencing Commission’s decisions regarding the guidelines.5
First, while the Commission purported to base the guidelines on an empirical approach that used as a starting point preguideline sentencing practice, “it used as its baseline only cases in which offenders had been sentenced to imprisonment.”6 This decision significantly skewed the data relating to past practice, as approximately 50 percent of defendants in the preguideline era received sentences of probation.7 To make matters worse, for certain crimes the Commission decided to depart from past practice entirely.8 For instance, the Commission tied the guidelines for drug offenses (the most prevalent federal crime) to the drug weight thresholds set forth in the Anti- Drug Abuse Act (“ADAA”), providing no explanation for why it extended the ADAA’s quantity-based approach across the entire spectrum of drug-trafficking sentences. The Commission also decided to increase sentences in public corruption, white collar, and immigration cases, among others.
“While rates of guideline ‘compliance’ decreased after the Supreme Court confirmed in 2007 that the guidelines really are advisory, average sentence length has remained stubbornly consistent.”
Second, the Commission discounted congressional directives that may have softened the guidelines. For instance, the SRA states that a probationary sentence would generally be appropriate for a first offender who had not been convicted of a crime of violence or an otherwise serious offense. Remarkably, the Commission viewed this as a “problem,” which it solved by simply reclassifying as “serious” many offenses for which courts had previously imposed probation.9 The Commission also appeared to have little regard for the SRA’s directive that it formulate guidelines to minimize the likelihood that the prison population would exceed the capacity of the federal prisons.10 Finally, the Commission apparently took the SRA’s rejection of prison as a vehicle for rehabilitation to mean that only retribution and crime control matter in determining a sentence.11 While Congress did not want judges to put people in prison to rehabilitate them, nothing in the SRA suggests that probation is inappropriate as a means of fostering rehabilitation.
Ultimately, the Commission constructed the sentencing grid so as to preclude probation in most cases. Under the guidelines, straight probation is allowed only when the defendant falls in “Zone A,” which occupies a mere sliver of the grid. Defendants who fall in “Zone B” may receive probation with a condition of home or community confinement. Everyone else goes to prison. Even in cases where the guidelines authorize probation, they do not affirmatively recommend it; every guideline sentence is phrased in terms of months in prison.12 Nor did the Commission provide substantive guidance on the “in/out” decision, despite the fact that Congress in the SRA directed that it do so.13
Based in part on the Commission’s decisions (and the SRA’s abolition of parole), the federal prison population increased from about 50,000 in 1988 to more than 200,000 today. Most people agree that this is unsustainable. Even staunchly conservative presidential candidates and members of Congress have called for a fresh approach. As indicated above, legislative reform of harsh mandatory sentences appears to be gaining steam in Congress, and the Commission itself has taken some modest steps to reduce sentences, such as the slight expansion (by one level each) of Zones B and C in 2010, the 2014 amendment reducing base offense levels in drug trafficking cases by two, and the pending revisions to the fraud guideline.
However, the Commission has resisted taking a systematic look at the guidelines. The original guidelines reflected the “law and order” mentality of the late 1980s,14 and they have generally gotten more punitive since then.15 As one commentator has noted, “The pro-prison bias continues despite staff reports in the meantime warning that the Guidelines’ requirement of prison in nearly every case is unnecessary and counterproductive.”16
Before the guidelines went into effect, judges were used to exercising broad sentencing discretion. As a result, the harsh and rigid guideline system created by the SRA came as a shock. Many judges chafed at the new restrictions. Some even quit. Over time, however, judges became accustomed to sentencing based on a grid. Many judges, particularly those who took the bench after the guidelines went into effect, appreciated a sentencing regime based on “objective” rules. The guidelines certainly made sentencing much easier. Perhaps it should not have come as too much of a surprise, then, that after the Supreme Court changed the status of the guidelines from mandatory to advisory in 2005,17 judges continued to follow them largely as they had before. While rates of guideline “compliance” decreased after the Supreme Court confirmed in 2007 that the guidelines really are advisory,18 average sentence length has remained stubbornly consistent.19
Before United States v. Booker, judges sentenced within the guidelines about 65 to 70 percent of the time.20 To the extent that this statistic might be regarded as an indication that judges continued to have sentencing discretion, we note that the vast majority of nonguideline sentences were sponsored by the government, for instance, to reward the defendant for providing substantial assistance to law enforcement.
In fiscal year 2006 (the first full post-Booker year), the number of guideline sentences dropped slightly, to 61.7 percent. Following the Supreme Court’s decisions in Gall v. United States, Kimbrough v. United States, and Nelson v. United States, the figure steadily dropped to the point where guideline sentences now comprise slightly less than half the total (although the government continues to sponsor most of those sentences).
Contrary to what one might expect, however, average sentence length has barely budged, dropping from 47.9 months in FY 2003 to 44 months in FY 2014. Perhaps more surprising, the number of offenders receiving prison-only sentences has actually increased, from 83.3 percent in FY 2003 to 87 percent in FY 2014.
What is the explanation for this? The most likely candidate is the psychological phenomena known as “anchoring.” As Judge Gerard Lynch explains:
When you have a question that is very subjective that people are called upon to answer and you give them a number as a kind of baseline, that number is very helpful. Whether people like that number or not, even if they are angry about that number, does not matter; they will still be influenced by that number. That is the psychological fact. I think it is psychologically inevitable that the Guidelines will have a powerful influence on sentences, even if they are purely advisory, because they put a number on a question that is otherwise quite subjective.21
Prof. Douglas Berman has similarly suggested that “status quo biases” — “the natural tendency of people to generally prefer things to stay relatively the same” — are at work.22 Prof. Frank Bowman explains the psychology this way: “Judges are men and women of the law. They naturally look for rules and endeavor to apply them.” While the guidelines no longer carry the force of law, they “still look and feel like ‘law.’” And this particular body of law gives guidance in an area where many judges feel in need of it, given the inevitable subjectivity involved in sentencing.23
Another possible explanation for why district court judges stick so closely to the guidelines is that they are aware that sentences within or close to the guideline range are almost always affirmed.24 And, as Prof. Alison Siegler has found, at the same time that appellate courts post-Booker have under-policed sentences that are within the guidelines, they have attempted to ensure that district courts comply with the guidelines by over-policing sentences outside the range. When the Supreme Court rejected some of the efforts by the courts of appeals to maintain the force of the guidelines — holding, for example, that sentences outside the range need not be justified by extraordinary circumstances, that such sentences may not be presumed unreasonable, and that district judges may not presume a guideline sentence to be reasonable — the courts of appeals created new rules that reinforce the guidelines. For instance, some courts of appeals have held that district judges may ignore “stock” arguments, typically based on the defendant’s personal characteristics (e.g., his employment history or family circumstances), reasoning that such arguments do not distinguish the defendant from many other offenders. This rule validates the guidelines’ distaste for “specific offender characteristics,” while essentially requiring the defendant to demonstrate extraordinary circumstances to obtain a nonguideline sentence, the very rule the Supreme Court rejected in Gall. The rule also encourages district judges to ignore categories of relevant evidence and impose within-guideline sentences without fear of reversal.25
Further, even though the Supreme Court has rejected some of the courts of appeals’ efforts to establish or re-establish a sentencing regime heavily influenced by the guidelines, its decisions nevertheless contain language pushing judges towards the ranges set by the guidelines. In Gall, for instance, the Court stated that the guidelines “should be the starting point and the initial benchmark,”26 and went on to say that:
[A] district judge must give serious consideration to the extent of any departure from the Guidelines and must explain his conclusion that an unusually lenient or an unusually harsh sentence is appropriate in a particular case with sufficient justifications. For even though the Guidelines are advisory rather than mandatory, they are . . . the product of careful study based on extensive empirical evidence derived from the review of thousands of individual sentencing decisions.
. . .
If [the judge] decides that an outside-Guidelines sentence is warranted, he must consider the extent of the deviation and ensure that the justification is sufficiently compelling to support the degree of the variance. We find it uncontroversial that a major departure should be supported by a more significant justification than a minor one.27
In Kimbrough, while holding that district judges could categorically reject the guidelines’ treatment (at that time)28 of 1 gram of crack cocaine as the same as 100 grams of powder cocaine, the Court again trumpeted the Commission’s institutional strengths, stating that:
[A] district court’s decision to vary from the advisory Guidelines may attract greatest respect when the sentencing judge finds a particular case outside the heartland to which the Commission intends individual Guidelines to apply. On the other hand, while the Guidelines are no longer binding, closer review may be in order when the sentencing judge varies from the Guidelines based solely on the judge’s view that the Guidelines range fails properly to reflect § 3553(a) considerations even in a mine-run case.29
And, in Rita v. United States, the justices held that courts of appeals may apply a presumption of reasonableness when reviewing within range sentences. While the Court stressed that this was an appellate presumption only,30 a district judge concerned about an appeal might be likely to hew closely to the range.
Another factor pushing district courts to adhere to the guidelines is the approach of the probation officers on whose reports and recommendations judges rely. In our experience in most cases, probation officers recommend sentences within the guidelines. Under the SRA, probation officers, who were once the social workers of the criminal justice system, became the “guardian” of the guidelines.31 In the preguideline era, the primary purpose of the presentence report was “to focus light on the character and personality of the defendant, to offer insight into his problems and needs, to help understand the world in which he lives, to learn about his relationships with people, and to discover those salient factors that underlie his specific offense and his conduct in general.”32 The probation officer served as the court’s eyes and ears, and the report was a tool with which the judge could fashion an individualized sentence geared towards rehabilitating the offender. “With allegiances to no one but the court, the officer’s only ‘agenda’ was to provide as much information about the defendant to the sentencing judge as possible.”33
With the advent of the determinate, mechanistic sentencing called for by the guidelines, the probation officer’s focus shifted dramatically; the probation officer as social worker paradigm had “no place in the Guidelines regime.”34 Instead of preparing reports geared towards understanding the causes of the offender’s behavior and assessing the possibility of change, the officer’s primary task was to apply a set of legal rules — the guidelines — to the facts of the case. Little else mattered. As one probation officer put it, this essentially cast the officer in the role of “punisher,”35 where the skills of a social worker were irrelevant — unnecessary to the task of calculating the guidelines.
Additionally, the officer became the focus of an adversarial sentencing system with a position to defend; once the neutral agent of the court, the officer morphed into the defender of guidelines protocol. This, in turn, changed the nature of the working relationship between sentencing judges and probation officers,36 and led to the creation of a new relationship between probation officers and the Sentencing Commission, which provided training and a staff hotline for guideline questions. In the post-SRA world, probation officers owed a dual loyalty: to the district court and to the Sentencing Commission.37
Post-Booker, little has changed in this dynamic. In 2007, Federal Rule of Criminal Procedure 32 was amended to make “clear that the court can instruct the probation office to gather and include in the presentence report any information relevant to the factors articulated in § 3553(a).”38 And presentence reports now include, after the “factors that may warrant departure” section, a segment called “factors that may warrant a sentence outside the advisory guideline system.” But the guidelines remain the focus of the report, and probation officers, consistent with their post-SRA role, continue to advocate guideline sentences in their recommendations to the court in most cases. Such recommendations are likely a substantial influence on judges. As Judge Jack Weinstein noted, under the SRA district judges became “passive” in the sentencing process, often doing little more than checking the probation officer’s math.39 For the judge looking to sentence based on all of the relevant factors, however, while probation officers may still have observations about defendants that are helpful, their guideline – centric recommendations rarely are.
The same is frequently true of the government’s sentencing recommendations, with prosecutors generally continuing to advocate sentences tied to the guidelines. Even in those cases where the government recommends a nonguideline sentence, based on, say, substantial assistance or an early disposition/“fast-track” program, the recommendation will be linked to the guideline range. Seldom do prosecutors recommend nonguideline sentences based on the factors specified in 18 U.S.C. § 3553(a), the statute which has governed federal sentencing since the guidelines became advisory. Indeed, for a time, Justice Department policy directed federal prosecutors to urge the court to exercise its discretion to impose a sentence in conformity with the guidelines. And the government’s sentencing recommendations likely have considerable influence.
Whatever the explanation,40 the fact is that notwithstanding their advisory status, the guidelines remain hugely influential in federal sentencing. Even when judges vary from the guidelines, they will begin with the guideline range and then adjust upward or downward. And, Commission data indicate that although judges depart from the guidelines more often than they once did, their departures on average are smaller than they were pre-Booker.41 Inasmuch as the guidelines recommend a prison sentence in virtually every case, it is no cause for wonder that very few defendants receive probation. In the next section, we suggest a different approach.
Of course, district judges must adhere to Supreme Court and circuit precedent regarding sentencing procedure. They must begin by correctly calculating the guideline range, as they did before Booker. Then, after giving both parties an opportunity to argue, judges must consider all the factors in § 3553(a) in order to determine an appropriate sentence. It is at this second step that we suggest a break from guideline-think and advocate an approach that places § 3553(a)’s parsimony provision at the fore.
The parsimony provision requires sentencing judges to “impose a sentence sufficient, but not greater than necessary to comply with the purposes” of sentencing.42 Parsimony refers to the principle that a judge should choose the least restrictive/least punitive sanction necessary to achieve defined social purposes,43 and the provision has an interesting place in the history of the SRA. Originally contained in a competing House version of the bill, the parsimony provision was ultimately proposed as an amendment to the Senate version by Sen. Charles Mathias. Under the House version, which called for advisory rather than mandatory guidelines, the judge was supposed to impose the least severe alternative sufficient to meet the purposes of sentencing, even if that meant departing from the guidelines. However, the Senate version, which ultimately became the SRA, rejected advisory guidelines. In the face of binding guidelines, the Mathias amendment was regarded as “hortatory” or “purely cosmetic.” Federal courts agreed, holding that the parsimony provision did not authorize departures from the guidelines.44 Section 3553(a) meant little so long as § 3553(b) required judges to follow the guidelines even when a lesser sentence would suffice.
But Booker breathed new life into the parsimony provision, excising § 3553(b) while leaving § 3553(a) in place. As Mary Price, of Families Against Mandatory Minimums, has noted, “It seems remarkable (perhaps prescient, given how limited the role of judges at sentencing would become under the guidelines) that Sen. Mathias would insist on the clear command to the courts. Today, judges (and defendants) everywhere owe him a great debt.”45
As the Supreme Court has recognized, the parsimony provision now constitutes the “overarching instruction” of § 3553(a).46 The provision requires judges to select the least restrictive sentence that satisfies the purposes of sentencing. Consequently, the first question a sentencing judge should ask is whether the goals of sentencing can be satisfied by imposing a sentence in which the defendant remains in the community. Put differently, the first question is whether the defendant needs to be imprisoned at all — not, as the guidelines ask, how long should the defendant serve in prison. Based on the parsimony provision, only if the judge concludes that the need for deterrence, desert, or incapacitation requires incarceration should the judge proceed to consider the length of the prison term.
As discussed above, the guidelines do not work that way. The ranges always speak in terms of months in prison. Even when probation is authorized under the grid, the guidelines do not affirmatively recommend it. Nor do the guidelines encourage judges to ask this antecedent question.
We therefore suggest that after calculating the guidelines and receiving the recommendations of the parties, the court first determine whether prison is needed to punish, deter, or protect the public. The guidelines will almost always say that it is, but as we have discussed elsewhere the guidelines for the most commonly prosecuted federal offenses are seriously flawed.47
District judges hesitant to impose more nonprison sentences should realize that probation is the default sentence in the states. “It is the federal sentencing system that is the outlier.”48 Judges also need to overcome the notion that probationary sentences are necessarily lenient. As the Wickersham Commission noted many years ago, probation is not merely “letting an offender off easily.”49 Indeed, given the stringency of the requirements that may be imposed, some offenders actually choose prison over probation. The conditions of probation may serve to punish (by requiring payment of fines, performance of community service, or service of home/community confinement), to incapacitate (by enforcing restrictions on travel, associations, or types of employment), and to deter (by requiring disclosure of financial information, submission to warrantless searches, and drug testing to ensure abstinence).50 Finally, probation can ensure that a defendant receives needed correctional treatment in the most effective manner, whether that be substance abuse, mental or physical health treatment, or assistance with employment training.51
We offer some examples of sentences imposed by Judge Adelman demonstrating how this approach produces more community-based sentences, even in mine-run cases where the guidelines call for multiple-year prison sentences.
Rafael Rodriguez, Miguel Santiago, and Israel Quiroz. Drug offenders comprise nearly half of the federal prison population. As discussed above, the guidelines base drug-trafficking sentences on the weight of the substance involved in the offense, ignoring other relevant considerations. Rodriguez, Santiago, and Quiroz were fairly typical drug-conspiracy defendants — low level street dealers drawn into a trafficking organization through addiction, familial relationships, and/or financial hardship. Despite their lower level involvement and limited prior records, each faced a guideline range of 37–46 months in prison.
After serving a period of time in pretrial detention, each eventually made a positive adjustment on bond, severing ties with negative peers and addressing their substance abuse, educational, and vocational issues. The purposes of sentencing did not require further imprisonment for these defendants. Given their lower level involvement, limited financial gain, the absence of violence, and the role of substance abuse in their conduct, further prison time was not needed to provide just punishment. Nor was it needed to protect the public and deter, given their limited prior records and improved conduct after the case was charged. Their correctional treatment needs could be, and in fact were being, met in the community.
Andre Goodman. If we are to make a serious dent in mass incarceration, we must go beyond focusing narrowly on nonviolent drug offenders. For example, although gun violence is a serious problem in this country, it is not necessary to impose prison sentences on all defendants convicted of firearms offenses. In Andre Goodman’s case, Milwaukee police executed a search warrant at his residence based on information from an informant that he possessed a firearm. Officers found a gun (which had previously been reported stolen) and ammunition in
Goodman’s bedroom, and because he’d previously been convicted of possession with intent to deliver and battery to a police officer, both felonies, the government indicted him under 18 U.S.C. § 922(g)(1). Released on bond, Goodman did well, testing negative for controlled substances, enrolling in an adult education program, and obtaining employment.
Goodman had not used, attempted to use, or threatened to use the firearm, but nevertheless faced a guideline range of 33–41 months in prison. This was so because the guidelines enhance punishment for possessing a firearm if the defendant has a prior felony conviction for either a crime of violence or a controlled substance offense. The guidelines treat all such convictions as indicators of an increased risk of gun violence, regardless of the particulars.52 Goodman had no history of gun violence, and his prior “crime of violence” was a battery which involved struggling with police officers who were trying to arrest him; moreover, his two prior felonies occurred more than a decade prior to the instant offense, and he had no intervening convictions or indications of violence. In addition, the guidelines provided a second enhancement because the gun Goodman possessed was stolen, even though the record contained no evidence that he had stolen the gun or even knew that it had been stolen.53
The purposes of sentencing did not require that Goodman be sentenced to prison; a sentence served in the community was sufficient to punish him for passively possessing a firearm inside his home, and his good performance on pretrial release suggested that supervision would suffice to protect the public and deter. Accordingly, Goodman was sentenced to probation with a home-confinement condition.
Melvin Washington. It is also important that courts avoid imprisoning people based on society’s emotional response to a particular class of offender. Sex offenders are the pariahs of our time. Since the 1990s, they have been subject to many burdensome restraints after serving their sentences including civil confinement, residency restrictions, and registration requirements. Presumably motivated by a wish to protect children from abduction or attack by strangers, these laws ignore the fact that most sex offenses against children are committed by someone close to the victim. They also overlook studies showing that, despite much lore to the contrary, sex offenders actually re-offend at lower rates than many other types of offenders. Moreover, studies show that registration requirements and residency restrictions are often counterproductive; they do not reduce sex crimes, particularly stranger-on-stranger sex crimes, and residency restrictions may make re-offending more likely by driving offenders underground.54
Federal judges grapple with the consequences of these laws in sentencing defendants convicted under the Sex Offender Registration and Notification Act (“SORNA”), which makes it a federal offense for a convicted sex offender to travel from one state to another and then fail to register or update a registration as required by law. In 2000, Melvin Washington was convicted of “aggravated criminal sexual abuse,” arising out of an incident in which he, then 21 years old, had sex with a 14-year-old girl. The state court sentenced him to probation but required lifetime registration as a sex offender. Washington registered with Illinois authorities, but in 2010 he moved to Milwaukee for work and failed to update his registration. He otherwise committed no new offenses in Wisconsin.
Washington faced a guideline range of 15-21 months as a “Tier III offender” — the most serious — despite the fact that his predicate conviction was essentially a statutory rape offense, which the state court judge did not believe required more than probation. Washington’s record contained no subsequent entries suggesting that he was a dangerous sexual predator, and he had moved to Wisconsin for employment, not to evade sex-offender registration requirements. Under these circumstances, prison was not needed to protect the public, and probation with monitoring conditions was sufficient to deter, given the possibility of revocation and imprisonment for any violations of law. A lengthy period of probation with a home-confinement condition ensured a sufficient measure of punishment.55
The approach to sentencing that we advocate here is not new. For much of this country’s history, sentencing judges started with the “in/out” question and imposed sentences of imprisonment only if they were convinced that no less restrictive alternative sufficed to satisfy the purposes of sentencing. Further, as we hope the examples discussed above have shown, noncustodial sentences are not reserved for the exceptional defendant but may be appropriate for many mine-run offenders. Moreover, it is unlikely that we will make much progress in reducing the federal prison population unless federal judges begin paying more attention to the parsimony provision in § 3553(a) and less to the “prison-for-almost-everyone” approach that remains the philosophy of the guidelines. Finally, we note that reducing mass incarceration involves both sending fewer people to prison and doing so for shorter lengths of time. In this article, we have focused on the former solution, but we would be remiss not to also mention the latter.
We conclude with two more statistics. In the year preceding completion of this article, Judge Adelman imposed nonguideline sentences in about 85 percent of all cases, and sentences requiring no further imprisonment (i.e., straight probation, probation with some form of home or community confinement, or time served followed by a period of supervised release) in about 42 percent.